BACKGROUND: North American women currently have a lifetime risk for osteoporotic fracture of nearly 50%. The mortality rate for hip fracture, the most common site, is nearly 10%. The development of new medications that can prevent loss of bone mineral density (BMD) and reduce the risk of fracture in selected groups of high-risk women has increased interest in the diagnosis of osteoporosis. The dual-energy x-ray absorptiometry (DEXA) scan is the most widely used test and costs between $100 and $200. The purpose of this study was to develop and validate a clinical prediction rule that would identify patients at the greatest risk of osteoporosis.
POPULATION STUDIED: The authors recruited 1376 Canadian women who underwent a DEXA scan of both the lumbar spine and the femoral neck as part of the Ontario Multicentre Osteoporosis Study. The mean age was 63 years, 94.9% were white, and 2.9% were Asian.
STUDY DESIGN AND VALIDITY: Patients were randomly divided into development (n = 926) and validation (n = 450) groups. Logistic regression analysis was used in the development group to identify independent risk factors for low BMD in the femoral neck and lumbar spine. This information was used to create several models with different permutations of the type and number of risk factors included. The most successful combination of variables, designated the Ontario Risk Assessment Instrument (ORAI), was then tested on the validation group. The major threats to validity are the homogenous nature of the population and the fact that these women were part of a clinical trial.
OUTCOMES MEASURED: The primary outcome was the sensitivity and specificity of the clinical prediction rule for the detection of women with a BMD T-Score 2 or more standard deviations below the norm.
RESULTS: An algorithm based on the 3 risk factors of age, weight, and current estrogen was 93.3% sensitive and 46.4% specific. The corresponding positive and negative likelihood ratios were 1.7 and 0.2, respectively. Adding other variables such as race, medical history, and lifestyle risks to the algorithm did not significantly increase the accuracy of the tool. The ORAI selects the following women as candidates for DEXA scanning: all women aged older than 45 years and weighing less than 60 kg (132 lb), all women aged 55 to 64 years weighing less that 70 kg (155 lb) and not taking supplemental estrogen, and all women aged 65 years and older regardless of their weight or current estrogen use status.
The ORAI is helpful for identifying women at risk for low BMD. It is a well-designed and well-validated rule that is easy to apply. However, several important issues must be better understood. For example, we do not presently know the best course of action when osteoporosis is diagnosed. With growing questions about its cardiac effects and a clear association with breast cancer, the role of hormone replacement therapy is in question. The bisphosphonates and drugs like raloxifene are promising treatments for osteoporosis but are expensive and lack long-term efficacy and safety data. In addition, BMD is a disease-oriented outcome. Until we have data from patient-oriented studies of the outcome of population-based screening, interventions aimed at lifestyle modification (eg, diet, exercise, and smoking cessation) and reduction of the risk of falling are cheaper than any of the medications, have no side effects, and are effective in the reduction of fractures.