SEARCH

SEARCH BY CITATION

Keywords:

  • bone mineral density testing;
  • osteoporosis;
  • risk assessment;
  • risk factors;
  • screening

Abstract.

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References

Objective.  Although osteoporosis and fragility fracture are common amongst postmenopausal women, the extent of risk varies, and measurement of bone mineral density (BMD) is the standard tool used to diagnose and assess fracture risk. Rates of diagnosis remain relatively low, and several groups have developed instruments to help identify individuals who would most benefit from BMD testing. In this paper, we review and compare the performance of these instruments to identify those most useful in the primary care setting.

Design.  Review of screening instruments comprised osteoporosis clinical risk factors and comparison of the sensitivity and specificity of these algorithms.

Results.  Validated instruments have varying complexity, but similar sensitivity and specificity for identifying individuals who are likely to have low BMD. The area under the receiver operating characteristic curve ranges from 0.75 (SOFSURF) to 0.81 (SCORE). The simplest of the instruments (OST) uses only age and weight and has an AUC of 0.79.

Conclusions.  The Osteoporosis Self-assessment Tool, the simplest of the instruments, performs as well as more complex tools and, because of its simplicity, may be the most useful means for the busy clinician to identify postmenopausal women who would most benefit from BMD testing.


Background

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References

Osteoporotic, or ‘fragility’, fractures are common in postmenopausal women. Lifetime risk of fracture for a 50-year-old White woman is estimated at over 70%; her risk of hip fracture alone is about 14% [1]. Morbidity from fractures is substantial, and mortality is increased by about 20% after hip fracture [1]. Increased mortality has also been reported amongst women with prevalent vertebral fractures [2]. Identification of women who are at high risk of fracture permits implementation of interventions that can reduce this risk. At present, measurement of bone mineral density (BMD) is the standard assessment tool for making the diagnosis of osteoporosis.

Low bone mass (osteopenia) and osteoporosis are widely prevalent in developed countries, particularly amongst postmenopausal women. The World Health Organization (WHO) considers osteoporosis to be second only to cardiovascular disease as a public health concern [3]. Worldwide, approximately one-third of women aged 60–70 years and two-thirds of women aged 80 and older have osteoporosis [3]. In 2002, according to estimates from the National Osteoporosis Foundation (NOF) 21.8 million women in the US had low bone mass (‘osteopenia’) and 7.8 million had osteoporosis; over 14 million men were estimated to have either low bone mass (11.8 million) or osteoporosis (2.3 million) [4]. Because life expectancy is increasing, the NOF estimated that 61.4 million men and women will have low bone mass or osteoporosis by 2020. Fragility fractures are projected to increase as well, because of both larger numbers of persons at risk and age-specific increases in fracture incidence. It has been estimated, for example, that the annual incidence of hip fracture worldwide could triple by 2050, with largest increases to be anticipated in Asia [5]. In Singapore, for example, annual hip fracture incidence amongst women has already increased fivefold during the last 40 years [3]. In the European Union, the incidence is expected to more than double, to nearly 1 million, by 2050 [3].

Until fracture occurs, osteoporosis is asymptomatic: the diagnosis cannot be made by history and physical examination, because the patient may appear to be perfectly healthy. Consequently, BMD testing is used to assess skeletal mass and to diagnose low bone mass and osteoporosis. By convention, based on criteria developed by the WHO, osteoporosis is defined as BMD that is 2.5 or more standard deviations (SD) below the mean of the young adult population [6]. Low bone mass, or osteopenia, is defined as BMD values between 1.0 and 2.5 SD below the young adult mean. The results of BMD testing are reported as T-scores, which correspond to the SD scale, so that, for example, a T-score of −2.5 or below is diagnostic of osteoporosis. Measurements at the hip are currently considered to be the gold standard, because of their precision and their greater ability to predict hip fracture risk than measurements at other sites in the skeleton. The association between BMD and fracture risk is strong and continuous. BMD explains as much as 85% of bone strength, and for each decrease of 1 SD in BMD fracture risk approximately doubles. This association is at least as strong as those between blood pressure and stroke or cholesterol and heart disease [1].

In recent years, BMD testing has become more widely available and many countries have developed guidelines regarding testing. Recently, for example, the United States Preventive Services Task Force recommended that all women 65 years of age and older have routine BMD screening and that women 60 and older have testing if they have risk factors for fracture [7]. However, despite increasing awareness of osteoporosis and its consequences, improved availability of testing, and the publication of guidelines for osteoporosis management, rates of diagnosis and treatment are low. More than half of the population at risk in the US, the UK, France, Germany, Italy, Spain and Japan remains unevaluated [3], and fewer than half of women who have been diagnosed as having osteoporosis are being treated. Treatment rates are low, as well, amongst women who have already sustained fragility fractures. Despite the fact that such fractures are associated with a doubling of risk for subsequent fracture, typically only 20–40% of these patients are evaluated or treated [8–10].

Although BMD testing is necessary to diagnose osteoporosis, measuring BMD in the entire population of postmenopausal women is logistically impossible (given the number of eligible women, the number of machines available, and the number of hours in a day), prohibitively expensive, and probably unnecessary, because many postmenopausal women have normal BMD and are at very low risk of fracture. Targeted testing, i.e. testing persons identified as being at high risk of having osteoporosis, is an appealing alternative. Assessment of clinical risk factors to identify women most likely to benefit from BMD testing has great appeal in primary care, if risk factors can correctly identify as high risk a substantial majority of women who have low bone mass or osteoporosis (as well as correctly identifying women whose BMD is likely to be normal), because risk factors can readily and inexpensively be identified in the office. In this paper, we review and compare the performance of several risk factor screening instruments.

Screening instruments

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References

Numerous risk factors for osteoporosis have been identified. Risk assessment questionnaires have been published by the NOF [1] and the International Osteoporosis Foundation, [11] although neither has been subjected to rigorous development and validation processes. In contrast to these risk assessment tools, several screening instruments have been constructed and tested to validate their usefulness in postmenopausal women (Table 1). ‘Usefulness’ of a diagnostic instrument is evaluated in terms of sensitivity, specificity and area under a receiver operating characteristic (ROC) curve (AUC). Sensitivity refers to the ability of the test to correctly identify persons with osteoporosis; specificity to the ability of the test to correctly identify persons without osteoporosis. Both sensitivity and specificity range from 0 to 100%. Because of the inter-relationship between sensitivity and specificity, increasing one generally decreases the other. Threshold values of diagnostic tests are chosen to optimize both sensitivity and specificity. The ROC curve graphs sensitivity and specificity throughout the range of test values, and the AUC quantifies the overall performance of the diagnostic test. The AUC can range from 0 to 1; an AUC of 0.50 means that the diagnostic test has no ability to discriminate between persons with and without the disease of interest. The closer the AUC is to 1, the better the ability of the test to discriminate.

Table 1.  Published instruments for estimation of risk of osteoporosis; the threshold values for recommending BMD testing are shown next to the name of the instrument. The objective of the instrument is to identify for testing groups as specified, using risk factors as shown in elements column. The performance of each instrument is evaluated by sensitivity, specificity, and area under the receiver operating characteristic curve (AUC)
InstrumentObjectiveElementsSensitivity (%)Specificity (%)AUC
SCORE ≥6Identify women likely to have femoral neck T-score ≤−2.0Age, race, weight, history of fracture, history of oestrogen use, history of rheumatoid arthritis89500.81
ORAI ≥9Identify women likely to have femoral neck or lumbar spine T-score ≤−2.0Age, weight, current oestrogen use90450.79
SOFSURF >3Identify women likely to have total hip T-score ≤−2.0Age, weight, history of fracture, current smokingNot reportedNot reported0.75
OST ≤−1Identify Asian women likely to have femoral neck T-score ≤−2.5Age, weight91450.79

Recently, Belgian investigators convened a nationwide expert panel to suggest a set of risk factors for use in both men and women, using questions about multiple medical illnesses, surgery, radiation, chemotherapy, and medication use, in addition to more typical screening questions [12]. This complex and somewhat cumbersome instrument does not include age as a risk factor; however, the authors reported that age 61 best discriminated between persons with and without osteoporosis at the hip or spine. With at least one of the risk factors, this instrument had a diagnostic sensitivity of 75% and a specificity of 37%.

In contrast, four relatively simple to use instruments that have undergone extensive testing and validation are available to the clinician: the Simple Calculated Osteoporosis Risk Estimation (SCORE), the Osteoporosis Risk Assessment Instrument (ORAI), the Study of Osteoporotic Fractures Simple Useful Risk Factors (SOFSURF), and Osteoporosis Self-Assessment Tool (OST). In the following sections, we will describe and compare the published performance of these instruments.

The SCORE was developed as an instrument to identify postmenopausal women who are likely to have a femoral neck BMD measurement that is 2 or more SD below the young adult mean (T-score ≤−2) [13]. Starting with a pool of over 350 potential risk factors, the final version uses six weighted factors: age, race, weight, history of fracture at the hip, rib, or wrist since age 45, history of oestrogen use, and history of rheumatoid arthritis. A value of 6 or more identifies women who should be referred for bone density testing. The sensitivity of SCORE in the original cohort was 89%, the specificity was 50%, and the AUC was 0.81. SCORE has been validated in other populations, in which SCORE sensitivity ranged from 90 to 98%, with specificity between 12 and 51% [14–17]. Thus, SCORE identifies almost all women whose BMD is low (high sensitivity); however, it also identifies many women who have normal BMD (low to moderate specificity).

The ORAI uses a set of three risk factors (age, weight and current oestrogen use) to identify women likely to have either femoral neck or lumbar spine T-scores ≤−2.0 [18]. Any patient with a score of 9 or higher would be advised to have BMD testing. The development cohort comprised women from the Ontario participants in the Canadian Multicentre Osteoporosis Study (CaMOS), and in this population the sensitivity was 90%, the specificity was 45%, and the AUC was 0.79. The ability of ORAI to identify women with femoral neck T-scores ≤−2.0 was subsequently compared with that of SCORE in CaMOS participants other than those who were part of the ORAI development cohort [19]. SCORE had a sensitivity of 98%, a specificity of 21%, and an AUC of 0.77; values for ORAI were 94%, 32%, and 0.76, respectively.

The SOFSURF, which has been published only in abstract form, was developed within the Study of Osteoporotic Fractures cohort and uses four simple, useful risk factors (age, weight, history of fracture and current smoking) to identify women likely to have total hip T-scores ≤−2.5 [20]. A patient with a SOFSURF score of 3 or higher is advised to have BMD testing. Women in the development cohort were 67–72 years of age, and the instrument yielded an AUC of 0.75. By comparison, the AUC for SCORE in this cohort was 0.79. The performance of SOFSURF has not been evaluated in women under the age of 65.

The simplest of instruments is the OST, which uses only age and weight to identify women likely to have a femoral neck T-score ≤−2.5 [21, 22]. The score is calculated by subtracting age from weight, multiplying by 0.2, and truncating to yield an integer [21]. In contrast to the other instruments, lower OST scores are associated with greater likelihood of low BMD, and a score <2 suggests that BMD be tested. OST was developed and tested in women from eight Asian countries, and yielded a sensitivity of 91%, a specificity of 45%, and an AUC of 0.79. It has been validated in several individual Asian countries. OST has also been validated in five large samples of primarily Caucasian women [22–24]. Alternatively, age and weight can be cross-referenced on an OST chart that is divided into three categories, corresponding to low, medium and high risk for osteoporosis (Fig. 1). Over half (58%) of women identified as high risk had T-scores ≤−2.5; only 4% of low risk women and 18% of moderate risk women were found to have osteoporosis [22]. In all of these populations OST performed at least as well as more complex screening instruments.

image

Figure 1. Risk classification using the OST algorithm, for use in assessing need for BMD testing in postmenopausal women. Risk level (low risk, at risk and high risk) is determined by the intersection of age and weight categories, i.e. by looking across the age group category and down the weight category. Even women whose age and weight fall into low risk category may need BMD testing and treatment if they have previously fractured or have other major risk factors. For women identified as ‘at risk’, BMD should be measured. Women at high risk according to weight and age should have BMD measured and, depending on the measurement, may need treatment.

Download figure to PowerPoint

Clinical usefulness of risk factor screening

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References

Usefulness of risk factor assessment depends not only on the diagnostic accuracy of the instrument, but also on ease of utilisation. Factors should be unambiguous and easily determined – ideally, self-reported by the patient. Age is an excellent example. It has the strongest known association with BMD and is included in all instruments in Table 1. The prevalence of osteoporosis increases substantially with age, as reported in the Third National Health and Nutrition Examination Survey (NHANES III) and shown in Fig. 2 [25]. In a recent study of more than 200 000 postmenopausal women in the US, the odds of having osteoporosis, even after controlling for all other significant risk factors, increased exponentially with age and were approximately 23 times greater for women 80 years of age and older than for 50–55-year-old women (Fig. 3) [26].

image

Figure 2. Prevalence of low bone mass (osteopenia, shown in cross-hatched bars) and osteoporosis (shown in dotted bars) amongst women in the US by age group, based on BMD measurements at the femoral neck in NHANES III [Source: 25] Osteoporosis was defined as a T-score ≤−2.5 and osteopenia as a T-score below −1.0 and above −2.5 in NHANES III.

Download figure to PowerPoint

image

Figure 3. Odds of T-score ≤−2.5 in NORA population, according to age group [Source: 26]. BMD was measured at peripheral skeletal sites in more than 200 000 postmenopausal women in the US. Shown here are multiply adjusted odds of T-score ≤−2.5, adjusted for race/ethnicity; years since menopause; education; self-rated health status; personal or maternal history of fracture of hp, rib, wrist, or spine since age 45; maternal history of osteoporosis; body mass index; current use of cortisone or diuretics; former or current use of oestrogen; former or current cigarette smoking; alcohol use; and exercise.

Download figure to PowerPoint

Conclusions

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References

Bone mineral density testing is the best way to identify persons at risk of fracture, and simple risk assessment instruments facilitate efficient identification of women most likely to have low BMD. Virtually all of the risk assessment tools reviewed in this paper appear to ‘work’, i.e. they identify patients who do have low bone mass. However, they cannot be used to diagnose osteoporosis and are not a substitute for BMD measurement. For clinical practice, the simplest approach to determine who should have BMD testing appears to be reading the patient's risk level from her age and weight values, using the OST chart.

Because patients who have already had a fracture are at markedly increased risk of subsequent fractures [27], early diagnosis and intervention directed at preventing the occurrence of the first fragility fracture are essential. Several pharmacological agents that increase BMD and decrease fracture risk are currently available, so that early identification of women at risk has become increasingly useful, from both individual and public health perspectives. Because the prevalence of osteoporosis and fragility fracture is high and because diagnostic and therapeutic tools are available, all physicians should take an active role in evaluating and treating osteoporosis, with the objective of decreasing the disease burden of osteoporosis and fracture. It must be noted, however, that no clinical trials have been conducted to provide direct evidence that risk evaluation programs actually reduce mortality and morbidity of osteoporosis. Nonetheless, risk tools such as OST can provide the busy clinician with a quick means of determining which patients would be most likely to benefit from BMD testing.

Conflict of interest statement

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References

Subsequent to preparation of this manuscript, LEW has become an employee of Merck & Co., Inc.

References

  1. Top of page
  2. Abstract.
  3. Background
  4. Screening instruments
  5. Clinical usefulness of risk factor screening
  6. Conclusions
  7. Conflict of interest statement
  8. References
  • 1
    National Osteoporosis Foundation. Osteoporosis: review of the evidence for prevention, diagnosis, and treatment and cost-effectiveness analysis. Osteoporos Int 1998; 8(Suppl. 4): S188.
  • 2
    Ensrud KE, Thompson DE, Cauley JA et al. Prevalent vertebral deformities predict mortality and hospitalization in older women with low bone mass. J Am Geriatr Soc 2000; 48: 2419.
  • 3
    International Osteoporosis Foundation. The Facts About Osteoporosis and its Impact. Lyon, France: International Osteoporosis Foundation, 2004. http://www.osteofound.org/press_centre/fact_sheet.html .
  • 4
    National Osteoporosis Foundation. America's Bone Health: the State of Osteoporosis and Low Bone Mass in Our Nation. Washington, DC: National Osteoporosis Foundation, 2002.
  • 5
    Cooper C, Campion G, Melton LJ. Hip fractures in the elderly: a world-wide projection. Osteoporos Int 1992; 2: 2859.
  • 6
    World Health Organization. Assessment of Fracture Risk and Application to Screening for Postmenopausal Osteoporosis. Geneva: WHO Technical Report Series, 1994.
  • 7
    US Preventive Services Task Force. Screening for Postmenopausal Osteoporosis. Rockville, MD, USA: Agency for Healthcare Research and Quality, 2002. http://www.ahcpr.gov/clinic/prev/ostscinv.htm.
  • 8
    Hajcsar EE, Hawker G, Bogoch ER. Investigation and treatment of osteoporosis in patients with fragility fractures. Can Med Assoc J 2000; 163: 81922.
  • 9
    Freedman KB, Kaplan FS, Bilker WB, Strom BL, Lowe RA. Treatment of osteoporosis: are physicians missing an opportunity? J Bone Joint Surg 2000; 82-A: 106370.
  • 10
    Cuddihy M-T, Gabriel SE, Crowson CS et al. Osteoporosis intervention following distal forearm fractures: a missed opportunity? Arch Intern Med 2002; 162: 4216.
  • 11
    International Osteoporosis Foundation. Osteoporosis Risk Test. International Osteoporosis Foundation, 2002. http://www.iofcongress.org/risk_test .
  • 12
    Ben Sedrine W, Broers P, Devogelaer JP et al. Interest of a prescreening questionnaire to reduce the cost of bone densitometry. Osteoporos Int 2002; 13: 43442.
  • 13
    Lydick E, Cook K, Turpin J, Melton M, Stine R, Byrnes C. Development and validation of a simple questionnaire to facilitate identification of women likely to have low bone density. Am J Man Care 1997; 4: 3748.
  • 14
    Von Muhlen D, Lunde AV, Barrett-Connor E, Bettencourt R. Evaluation of the Simple Calculated Osteoporosis Risk Estimation (SCORE) in older Caucasian women: the Rancho Bernardo study. Osteoporos Int 1999; 10: 7984.
  • 15
    Cadarette SM, Jaglal SB, Murray TM. Validation of the simple calculated osteoporosis risk estimation (SCORE) for patient selection for bone densitometry. Osteoporos Int 1999; 10: 8590.
  • 16
    Ungar WJ, Josse R, Lee S et al. The Canadian SCORE questionnaire: optimizing the use of technology for low bone density assessment. J Clin Densitom 2000; 3: 26980.
  • 17
    Ben Sedrine W, Devogelaer JP, Kaufman JM et al. Evaluation of the Simple Calculated Osteoporosis Risk Estimation (SCORE) in a sample of white women from Belgium. Bone 2001; 29: 37480.
  • 18
    Cadarette SM, Jaglal SB, Kreiger N, McIsaac WJ, Darlington GA, Tu JV. Development and validation of the Osteoporosis Risk Assessment Instrument to facilitate selection of women for bone densitometry. Can Med Assoc J 2000; 162: 128994.
  • 19
    Cadarette SM, Jaglal SB, Murray TM, McIsaac WJ, Joseph L, Brown JP. Evaluation of decision rules for referring women for bone densitometry by dual-energy X-ray absorptiometry. JAMA 2001; 286: 5763.
  • 20
    Black DM, Palermo L, Pearson J, Abbott T, Johnell O. SOFSURF: a simple, useful risk factor system can identify the large majority of women with osteoporosis. Bone 1998; 23(Suppl.): S605 (Abstract).
  • 21
    Koh LKH, Ben Sedrine W, Torralba TP et al. A simple tool to identify Asian women at increased risk of osteoporosis. Osteoporos Int 2001; 12: 699705.
  • 22
    Geusens PP, Hochberg MC, Van Der Voort DJM et al. Performance of risk indices for identifying low bone density in postmenopausal women. Mayo Clinic Proc 2002; 77: 62937.
  • 23
    Cadarette SM, McIsaac WJ, Hawker GA et al. The validity of decision rules for selecting women with primary osteoporosis for bone mineral density testing. Osteoporos Int 2004; 15: 3616.
  • 24
    Richy F, Gourlay M, Ross PD et al. Validation and comparative evaluation of the osteoporosis self-assessment tool (OST) in a Caucasian population from Belgium. Q J Med 2004; 97: 3946.
  • 25
    Looker AC, Johnston CCJ, Wahner HW et al. Prevalence of low femoral bone density in older U.S. women from NHANES III. J Bone Miner Res 1995; 10: 796802.
  • 26
    Siris ES, Miller PD, Barrett-Connor E et al. Identification and fracture outcomes of undiagnosed low bone mineral density in postmenopausal women: results from the National Osteoporosis Risk Assessment. JAMA 2001; 286: 281522.
  • 27
    Klotzbuecher C, Ross P, Landsman P, Abbott T, Berger M. Patients with prior fractures have an increased risk of future fractures: a summary of the literature and statistical synthesis. J Bone Miner Res 2000; 15: 72139.