Performance of osteoporosis risk assessment tools in postmenopausal women aged 45–64 years

Abstract

Osteoporosis risk factor assessment is of uncertain utility in women under 65 years of age. Previous comparative studies of osteoporosis risk assessment tools were not stratified by age. We compared the discriminatory ability of three previously validated osteoporosis risk assessment tools in a referral population of postmenopausal women aged 45–64 years (n=2539) and aged 65–96 years (n=1496) seen at a university-based outpatient osteoporosis center in Belgium. Risk scores for the Osteoporosis Self-assessment Tool, Osteoporosis Risk Assessment Instrument, and Simple Calculated Osteoporosis Risk Estimation were calculated for each patient. The reference standard was osteoporosis at the femoral neck, defined as a T-score ≤−2.5 based on bone mineral density measured by dual energy X-ray absorptiometry. Osteoporosis was present in 139 of 2539 (5.5%) women aged 45–64 years and 241 of 1496 (16.1%) women aged 65 years or older. The tools had similar overall discriminatory ability to identify women with osteoporosis [area under the ROC curve 0.750–0.768, P=0.23 for women aged 45–64 years; area under the ROC curve 0.745–0.762, P=0.06 for women aged 65 years or older (P>0.05 indicates no difference among tools)]. The likelihood ratios for the high-risk score categories ranged from 3.60 to 6.73 for the younger women and 3.45 to 6.99 for the older women when different score thresholds were set to maximize the performance of each tool in each age group. We conclude that the diagnostic accuracy of three osteoporosis risk assessment tools was similar in postmenopausal women aged 45–64 years and women aged 65 years or older. Use of structured risk assessment tools to identify women at high risk of osteoporosis in the early postmenopausal period warrants further study. Of the three tools evaluated, the OST is the simplest and has the best potential for use in clinical practice.