Performance of Risk Indices for Prediction of Osteoporosis in Post- and Perimenopausal women / 대한산부인과학회지

ABSTRACT

OBJECTIVE: To examine the ability of published osteoporosis risk indices for prediction of osteoporosis in post- and perimenopausal women. METHODS: Subjects included 1559 of post- and perimenopausal women, 45 year old and older, recruited from Menopause Clinic from January 2002 to June 2004. Their bone mineral density (BMD) of femur neck was determined by dual energy x-ray absorptiometry (DXA), and their historical and clinical risk factors were assessed. We employed the World Health Organization (WHO) diagnostic criteria of osteoporosis for calculating the T scores, using race/ethnic and gender-specific young adult mean values. We applied 6 published osteoporosis risk indices (OSTA, ORAI, SCORE, SOFSURF, ABONE, NOF practice guideline) to each subject. Also, the predictive abilities of risk indices were assessed using the area under receiver operating characteristic (ROC) curve. RESULTS: The prevalence of osteoporosis and osteopenia was 8.7% and 32.6% for femur neck, 16.0% and 30.1% for lumbar spine, respectively. The prevalence of osteoporosis for femur neck was increased dramatically in the coming decades (1.6% for fifties, 8.7% for sixties, 43.0% for seventies, 64.1% for over eighties). ROC curve for predicting osteoporosis according to femur neck by OSTA showed the area under the curve (AUC) of 0.920. Sensitivity, specificity, and positive predictive value (PPV) were 93.4%, 75.4%, 26.6%, respectively. AUC for ORAI, SCORE, SOFSURF, ABONE, and NOF practice guideline was 0.896, 0.916, 0.871, 0.846, and 0.856, respectively. CONCLUSION: Validated risk indices have varying complexity, but similar sensitivity, specificity, and PPV for identifying individuals who are likely to have osteoporosis. However, OSTA is the most excellent risk index and simple tool.