Osteoporosis Among Estrogen-Deficient Women—United States, 1988-1994

MMWR. 1998;47:969-973

1 table omitted

Each year in the United States, hip fractures result in approximately 300,000 hospital admissions and an estimated $9 billion in direct medical costs.¹ Most of these fractures result from osteoporosis among women who experience accelerated bone loss after natural or surgically induced menopause. Measurement of bone mineral density (BMD) is the best tool available to assess osteoporotic fracture risk for women after menopause²; a reduction of one standard deviation (SD) in femoral BMD is comparable to a 14-year increase in age on the risk for hip fracture.³ A technology that allows highly accurate and precise measurement of BMD is dual energy x-ray absorptiometry (DXA). CDC's Third National Health and Nutrition Examination Survey (NHANES III) was the first nationally representative survey that used DXA to estimate osteoporosis prevalence based on BMD in the U.S. population, providing baseline information for assessing national prevention and intervention needs for this disease. This report compares self-reported health information with BMD measurements from NHANES III conducted during 1988-1994; the findings indicate that most estrogen-deficient women in the United States who had femoral osteoporosis based on BMD were unaware of having this condition, reflecting the evolving nature of research and clinical practice regarding osteoporosis.

NHANES III collected data through household interviews and direct standardized physical examinations using specially equipped mobile examination centers. A total of 14,646 men and nonpregnant women aged ≥20 years (excluding those with histories of fractures on both hips) underwent DXA scanning of the proximal femur. This represented 78% of the eligible interviewed sample, and 88% of the eligible examined sample. The analysis for this study was restricted to women who reported natural or surgically induced (i.e., bilateral oophorectomy) menopause and who had never used exogenous hormones. Women with these characteristics were considered to be at high risk for osteoporosis and thus had been identified previously as appropriate candidates for BMD testing.⁴ Women with BMD results that were unacceptable for technical reasons were excluded from the analysis (2%). The final analytic sample comprised 2314 women.

All estimates were generated using SUDAAN. Estimates were stratified by selected risk factors for osteoporosis (i.e., age, race, body mass index [BMI], and whether menopause had been induced surgically) and possible confounders of self-reporting (i.e., education or income level, urban or rural residence, health-care use, and usual source of care). Prevalence estimates of osteoporosis based on BMD were calculated using the World Health Organization (WHO) diagnostic criteria, which defined osteoporosis as a BMD value >2.5 SD below the mean of a young adult reference group.*^6,7 Prevalence is reported for the total femur region because this skeletal site was chosen for standardization of femur BMD between different DXA densitometers.⁸ Prevalence of self-reported osteoporosis was estimated based on responses to the question, "Has a doctor ever told you that you had osteoporosis, sometimes called thin or brittle bones?" The concordance was the percentage of women who reported a diagnosis of osteoporosis out of women with femoral osteoporosis based on BMD measurements.

Among the study group of 2314 women, 17% (95% confidence interval [CI]=15.5%-19.0%) had a femoral BMD value that met WHO's definition for osteoporosis, and 5% (95% CI=4.3%-6.6%) reported having been told by a doctor they had osteoporosis. Based on BMD results, the prevalence of osteoporosis was significantly higher among women aged ≥65 years (29.5%) than among younger women (5.7%), among non-Hispanic white women (18.7%) than among all women of other racial/ ethnic groups (11.6%), and among women with a BMI <25 (33.3%) than among women with a BMI ≥25 (8.0%). The prevalence also was higher among women with bilateral oophorectomy than among those with natural menopause, but the difference was not statistically significant. Among women who self-reported having had osteoporosis diagnosed, risk for osteoporosis was higher among non-Hispanic whites, but there were no significant differences for women in other risk categories. Self-reported data also suggested the prevalence was lower among women aged ≥65 years than among younger women.

Women's knowledge of their osteoporosis varied more by socioeconomic status (SES) and by health-care factors than by BMD measurements. Self-reported prevalence estimates generally were significantly lower among women with income at or below poverty level† (2.2%) and women not seen by a doctor for ≥6 months (2.6%). Although sample sizes in some of these strata were too small to detect statistical significance, lower SES and fewer health-care resources were associated with lower self-reported prevalence of osteoporosis. A similar association was not found in BMD measurements, which showed no differences between most of these categories or slightly higher prevalence among women with lower education levels or women with income at or below poverty level.

Overall, 7% of women whose osteoporosis was diagnosed by BMD were aware of their condition. The concordance between self-reported and BMD data was low in all population subgroups, particularly for women with lower SES and fewer health-care resources (range: 1%-5%) compared with women with higher SES and more health-care resources (range: 8%-10%).

Div of Health Examination Statistics, National Center for Health Statistics, CDC.

The findings in this report indicate that 93% of estrogen-deficient women with osteoporosis as defined by BMD were unaware of this condition; this finding, coupled with the observed high prevalence of osteoporosis in certain populations of women, underscores the need for greater awareness about the disease, especially among high-risk women and their health-care providers. Routine BMD screening has not been recommended for several reasons, including the high cost of densitometry, insufficient evidence that screening will influence treatment decisions and decrease fracture incidence, and lack of universally accepted criteria for using BMD results to initiate treatment.⁵ However, the level of awareness of osteoporosis among women described in this report who had undergone natural or surgically induced menopause and did not use hormone replacement therapy was low, even though these women have been identified as appropriate candidates for bone density testing.⁴

Several factors probably explain the discrepancy between presence and awareness of low BMD among high-risk women at the time of NHANES III. The evolving nature of research and clinical practice regarding osteoporosis probably is a primary contributor. For example, the definition of osteoporosis, which traditionally required the presence of a fracture, was expanded in 1994 by a WHO expert panel to include diagnostic criteria that are based on low BMD alone.^7,9 Development of accurate and precise noninvasive techniques to measure bone density and the availability of results from prospective studies linking low BMD to fracture risk were the basis of change to BMD-based diagnostic criteria. However, many of these developments occurred at roughly the same time NHANES III was being conducted, and access to and familiarity with bone densitometry among primary-care physicians was low during NHANES III.⁸ Finally, results of the study described in this report suggest that some of the discordance between presence and awareness of low BMD may reflect patterns of health-care access and use in different socioeconomic groups (e.g., those with lower health-care use have less opportunity to be evaluated for osteoporosis).

The NHANES III data in this study are subject to at least two limitations. First, self-reported data may include reporting errors because some respondents could not recall a diagnosis of osteoporosis or did not understand the osteoporosis question correctly. Second, the self-reported data included a physician's diagnosis of osteoporosis at any sites of the skeleton, while the DXA data included only BMD measurements of the proximal femur.

The proportion of women with osteoporosis who know their condition may increase in response to changes in health-care practices, reimbursements, and treatment options. The Medicare Bone Mass Measurement Coverage Standardization Act established national criteria for bone density test reimbursement in the Medicare program as of July 1, 1998; previously, each Medicare carrier made its own coverage decision. The new act authorized standardized coverage for persons at high risk for osteoporosis, such as estrogen-deficient women, persons with vertebral abnormalities, persons receiving long-term glucocorticoid therapy, persons with primary hyperparathroidism, and persons being monitored to assess the response of an osteoporosis drug therapy approved by the Food and Drug Administration.‡ The number of therapeutic options for addressing osteoporosis also has doubled with the approval of alendronate in 1995 and raloxifene in 1997 for treating and/or preventing osteoporosis, providing alternatives to the established therapies of estrogen replacement therapy and calcitonin. Finally, the National Osteoporosis Foundation (NOF) has released more detailed guidelines on bone density testing and treatment for osteoporosis.¹⁰ Additional information is available from NOF, telephone (800) 400-1079 or World-Wide Web site http://www.nof.org.