The Ability of a Single BMD and Fracture History Assessment to Predict Fracture Over 25 Years in Postmenopausal Women: The Study of Osteoporotic Fractures

Dennis M. Black, Jane A. Cauley, Rachel Wagman, Kristine Ensrud, Howard A. Fink, Teresa A. Hillier, Li Yung Lui, Steven R. Cummings, John T. Schousboe, Nicola Napoli

Research output: Contribution to journalArticlepeer-review

19 Scopus citations


The ability of bone mineral density (BMD) and other risk factors to predict fracture risk is well-established for as long as 5 to 10 years. However, their value to predict risk over a longer term has not been directly studied. We investigated whether a single assessment of femoral neck BMD and fracture history can predict fracture risk over 20 to 25 years. We used data from the Study of Osteoporotic Fractures (SOF) that assessed BMD and risk factors in 7959 women age ≥67 (mean = 73.4) in 1988–1990. Follow-up for fractures continued for 25 years for hip fracture, and for 20 years for any nonvertebral fracture. Using age-adjusted proportional hazards models, we analyzed the relationships between a single baseline assessment of femoral neck BMD, fracture history and age, and 20–25-year fracture incidence. The 25-year cumulative incidence of hip fracture was 17.9%; 20-year incidence of any nonvertebral fracture was 46.2%. The 25-year hip fracture incidence was highest in those ≥80 years old (22.6%) compared to 13.9% in women aged <70 years. A single femoral neck BMD measurement strongly predicted long-term hip fracture risk to 25 years: 29.6% risk in the lowest BMD quartile versus 7.6% with the highest relative hazard (RH) = 4.9 (95% CI, 4.1 to 6.0). Femoral neck BMD predicted hip fracture with little degradation over time from RH/SD = 2.6 (2.2 to 3.0) for 0 to 5 years to RH/SD = 1.8 (1.4 to 2.4) for 20 to 25 years. Lifetime hip fracture risk was similar (∼30%) regardless of age from 67 to >80 years. History of hip fracture predicted hip fractures only slightly better than history of nonvertebral fracture (RH = 1.6 [95% CI, 1.1 to 2.2] versus RH = 1.4 [95% CI, 1.2 to 1.5], respectively). Fracture history remained strongly predictive up to 25 years. We conclude that a single BMD and fracture history assessment can predict fracture risk over 20 to 25 years. Long-term risk of hip fracture remains extremely high in the oldest age groups, supporting risk assessment and consideration of treatment even in the oldest, highest-risk women.

Original languageEnglish (US)
Pages (from-to)389-395
Number of pages7
JournalJournal of Bone and Mineral Research
Issue number3
StatePublished - Mar 2018

Bibliographical note

Funding Information:
1Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA, USA 2Department of Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, PA, USA 3Amgen Inc, Thousand Oaks, CA, USA 4Division of Epidemiology and Community Health and Department of Medicine, University of Minnesota, Minneapolis, MN, USA 5Center for Chronic Diseases Outcomes Research, Minneapolis VA Health Care System, Minneapolis, MN, USA 6Kaiser Permanente Center for Health Research, Northwest/Hawaii, Portland, OR, USA 7California Pacific Medical Center Research Institute, San Francisco, CA, USA 8Division of Health Policy and Management, University of Minnesota, Minneapolis, MN, USA 9Universita Campus Bio-Medico di Roma, Rome, Italy 10Istituto di ricovero e cura a carattere scientifico (IRCCS), Istituto Ortopedico Galeazzi, Milan, Italy

Funding Information:
DMB reports receiving grant support from Alexion and consulting fees from Amgen, Merck, Asahi-Kasei, and Radius. NN reports receiving consulting fees from Amgen, MSD, UCB, and Eli Lilly. RW is an employee of Amgen.

Funding Information:
The Study of Osteoporotic Fractures (SOF) is supported by National Institutes of Health funding. The National Institute on Aging (NIA) provides support under the following grant numbers: R01 AG005407, R01 AR35582, R01 AR35583, R01 AR35584, R01 AG005394, R01 AG027574, and R01 AG027576. A research grant for part of this analysis was provided to UCSF by Amgen Inc. Amgen had no involvement in the design of SOF or the recruitment of patients and did not participate in the analysis or have access to data. L. Wu (University of California, San Francisco) gave administrative and technical support, and Dr. Eric Vittinghoff and Lisa Palermo provided statistical methodology support.


  • AGE
  • BMD

Fingerprint Dive into the research topics of 'The Ability of a Single BMD and Fracture History Assessment to Predict Fracture Over 25 Years in Postmenopausal Women: The Study of Osteoporotic Fractures'. Together they form a unique fingerprint.

Cite this