Variance in 10-Year Fracture Risk Calculated With and Without T-Scores in Select Subgroups of Normal and Osteoporotic Patients
Document Type
Article
Publication Date
4-1-2009
Description
The World Health Organization fracture risk assessment tool (FRAX) uses clinical risk factors to predict the patient's 10-yr probability of sustaining a hip or other major osteoporosis-related fracture. Inclusion of the femoral neck T-score is optional in the calculation. We evaluated the impact of including the T-score in the calculation of fracture risk and resultant treatment recommendation. We retrospectively reviewed charts of 180 white women scanned on a Hologic dual-energy X-ray absorptiometry (DXA). FRAX scores were calculated with T-scores (FRAX+) and without T-scores (FRAX-). We compared the National Osteoporosis Foundation (NOF) treatment recommendations (≥20% risk of a major osteoporotic fracture or ≥3% risk of hip fracture for osteopenic patients) between FRAX+ and FRAX- scores. Agreement between FRAX+ and FRAX- was 89.4%. Disagreement occurred in 2 distinct subgroups of patients (10.6% of cases), that is, FRAX+ scores exceeded the NOF recommended treatment thresholds and FRAX- scores did not, or vice versa. One subgroup comprised older patients with normal T-scores for whom FRAX- scores exceeded the treatment threshold. The second subgroup comprised younger patients with high body mass index (BMI) and low T-scores for whom FRAX- scores did not exceed the treatment threshold. FRAX scores generated without T-scores may lead to treatment recommendations for patients who have normal bone mineral density and no treatment recommendations for patients who have osteoporosis. T-scores should be used for optimal application of FRAX.
Citation Information
Hamdy, Ronald C.; and Kiebzak, Gary M.. 2009. Variance in 10-Year Fracture Risk Calculated With and Without T-Scores in Select Subgroups of Normal and Osteoporotic Patients. Journal of Clinical Densitometry. Vol.12(2). 158-161. https://doi.org/10.1016/j.jocd.2008.12.003 PMID: 19201635 ISSN: 1094-6950