DXA FRAX® May Differ Significantly and Substantially To Web FRAX®

Introduction: Osteoporosis and the associated fractures are a major global health burden for patients, their social network and societies. Access to quality risk assessment, diagnosis and treatment is heterogeneous and inadequate around the globe. Many algorithms have been developed for deciding who to test, assessing fracture risk, when to intervene, and how to monitor the effects of interventions. FRAX® has become the dominant fracture risk assessment tool worldwide which estimates the average 10 year risk of major osteoporotic fracture risk and hip fracture, and is recommended by our global partners: ISCD and IOF. A particular attraction of the FRAX® algorithm is that on-line calculations can be made with or without DXA testing, which is particularly attractive when access to DXA is limited or non-existent. In our centre public waiting times for non-urgent DXA are up to 10 years, so reducing unnecessary testing is a laudable and overarching service goal. A recent Japanese study comparing 4 FRAX calculation tools suggested little difference exists. However in our unit we have noticed at times very different results when using the Ireland embedded DXA-FRAX® tool to the Web-FRAX® Ireland tool.

Objectives: 1) To Compare DXA-FRAX® to Web-FRAX® 10-year estimates for MOF and HF.
2) To Compare Web-FRAX® 10-year estimates for MOF and HF with and without DXA-measured T-scores.

Methods: We used a sub-group of the DXA-HIP Cohort as previously published to gain initial insights into DXA-FRAX® to Web-FRAX® as part of a larger calibration and validation project, which has been approved by our hospitals Ethics and Research Committee. All scans are performed and reported by staff trained to ISCD standards and recommendations. At the time of reporting DXA-FRAX® scores are calculated as part of the analysis as recommended by ISCD and IOF, and more recently Web-FRAX® for Ireland has been calculated and stored with and without BMD data. All data are cleaned, merged, anonymized and stored for analysis. Summary cross-sectional data, and comparison of differences were completed using absolute and relative scales, and Bland-Altman Plots for agreement between measures. Weights were reduced to 125kg for the on-line tool and age to 90 years when values exceeded these thresholds.

Results: Data on 588 adults aged ≥40 years were available of whom 108 (18%) were men. Mean age was 67 years, BMI: 28kg/m2, 249 had a prior fracture, 64 a parental hip fracture, 40 smokers, 42 taking glucocorticoid therapy, 39 had rheumatoid arthritis, 5 were excess alcohol users while 107 had other FRAX®-listed secondary causes of osteoporosis. Mean WF femoral neck BMD T-score was -1.4. Mean MOF DXA-FRAX® was 15.3 (SD:11.1) compared to 12.6 (SD:8.5) for Web-FRAX®, while HF means were 6.7 (SD:8.4) Vs 3.8 (SD:5.5) respectively. Bland-Altman plots showed there was much greater discrepancy for men, and those with higher risk estimates. Comparisons between DXA-FRAX® and Web-FRAX® revealed similar results.

Conclusions: Significant differences exist in results of DXA-FRAX® and Web-FRAX® for Ireland, particularly for men and those with higher risk estimates so these results should be interpreted cautiously. Reassuringly results were similar for those deemed at lower risk and for women.