Can We Identify Women at High Risk of Osteoporotic Fractures?

Osteoporosis is a major health problem in the United States; 10 million persons have the disorder, and 18 million others have low bone mass, placing them at increased risk of developing the disease. Skeletal fractures, the consequence and hallmark of osteoporosis, are associated with pain, deformity, functional impairment, and increased mortality. Although previously considered a normal part of aging, this disease can be prevented, and affected subjects can decrease their risk of fractures with a variety of strategies. For most subjects, osteoporosis is a silent disease, until a fracture occurs. Neither bone loss nor the failure to obtain one's potential peak bone mass produces any symptoms. Over the last 20 years, substantial progress has been made in developing noninvasive technologies to assess bone mass. These technologies have been shown to predict the risk of fractures in peri- and postmenopausal women. Two important questions arise when one considers ordering bone mineral density (BMD) tests. First, in an era when controlling healthcare expenditures is increasingly important, it is necessary to determine which women should have BMD evaluations. Second, because BMD does not completely determine fracture risk, it is useful to determine how other risk factors for fractures can be used in conjunction with BMD measurements to guide treatment decisions. In this issue of the Journal of the American Geriatrics Society, Buist et al. describe how three published assessment guidelines (World Health Organization, National Osteoporosis Foundation, and Study of Osteoporotic Fractures–Based Guidelines) perform for assessing those at high risk of fracture in a population of women aged 60 to 79 in a Seattle health maintenance organization.1 In this self-selected group, 22% of the women aged 65 and older not taking hormone replacement therapy had had a clinical fracture after age 50, 31% had a BMD lower than −2.5 standard deviations below the mean BMD for women aged 20 to 24, and 9% had five or more risk factors for osteoporosis. Basing treatment decisions on BMD alone (World Health Organization criteria) missed nearly 60% of women who had already experienced a clinical fracture, women who are well documented as being at high risk for further fractures. Buist et al. point out that any algorithm used to identify women at risk for osteoporosis will involve choices between the sensitivity and specificity of the instrument, although they do not report the specificity of the criteria in their analysis. Although the economic impact of bone density screening programs has not been established, it seems clear that additional factors should be added to BMD measurements when deciding whom to screen and treat. Recently, Canadian investigators conducted a related study in 2,365 women aged 45 and older, comparing the characteristics of four assessment guidelines for identifying women for referral for BMD testing with each other and with the guidelines suggested by the National Osteoporosis Foundation.2 They found that two of the four clinical decision rules (Simple Calculated Osteoporosis Risk Estimation (SCORE)3 and Osteoporosis Risk Assessment Instrument)4 had better sensitivity, specificity, and area under the receiver operating characteristic curve than the other two guidelines or the National Osteoporosis Foundation guidelines. The authors suggest further study of these assessment tools to determine whether they could be used to screen for the use of densitometry services.2 Two other studies assessed the sensitivity, specificity, and area under the receiver operating curve of SCORE in a Canadian population and in the Rancho Bernardo population.5,6 The Rancho Bernardo study of 1,013 community-dwelling women aged 44 and 98 and the Canadian study of 398 postmenopausal women, mean age 64.5, used the same SCORE threshold; both found low specificities (12.5% and 32%, respectively). Both of these studies concluded that the SCORE instrument was not an effective screening tool because of its low specificity. More recently, the results of the National Osteoporosis Risk Assessment Project (NORA) have been published.7 This study was a longitudinal observational study involving 200,160 ambulatory postmenopausal women with a goal of identifying major risk factors for osteoporosis. BMD was assessed by peripheral bone densitometry, and questionnaires assessed risks for fractures and personal fracture history. Age, personal or family history of fracture, Asian or Hispanic heritage, smoking, and glucocorticoid use were associated with a significantly increased likelihood of osteoporosis. Using World Health Organization criteria, 39.6% of the subjects had osteopenia (T score of −1 to −2.49) and 7.2% had osteoporosis (T score of ≤−2.5). The study followed 163,979 subjects for 12 months for evidence of clinical fractures. The fracture rate of subjects with osteoporosis was four times that of women with normal BMD; in subjects with osteopenia the relative risk of fracture was 1.8. Major strengths of this study were the large number of subjects and the ethnically diverse community-based sample of women from 34 states in the United States. Although peripheral BMD measurements were used, the conclusions are sound.8 Other NORA studies suggest comparable predictive ability for fracture risk when peripheral BMD sites are used employing receiver operating characteristic curve analysis.9,10 For practicing geriatricians, what “pearls” for patient management come from these studies? We can incorporate five easily obtainable facts into our decision-making process with our patients when evaluating women at risk for osteoporosis: age, race, smoking history, use of glucocorticoid medication, and a history of a previous fracture after age 45 or 50. These have consistently been shown to be strong risk factors for subsequent fractures and are important in helping patients make informed choices about osteoporosis screening. Nevertheless, as stated by Buist et al, there is still no assessment algorithm available with adequate sensitivity and specificity that can be recommended for screening peri- and postmenopausal women. Further analysis of the large database from NORA may make it possible to create such a screening algorithm.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.