A comparative study of using non-hip bone density inputs with FRAX®.

Area of Science:

• Osteoporosis research
• Bone mineral density assessment
• Fracture risk prediction

Background:

• The FRAX algorithm is designed for femoral neck bone mineral density (BMD) input.
• Using alternative BMD inputs like lumbar spine T-scores is not recommended.
• Limited research exists on the impact of non-femoral neck BMD inputs on FRAX performance.

Purpose of the Study:

• To evaluate the effect of different bone mineral density (BMD) inputs on FRAX algorithm performance.
• To compare fracture risk predictions using femoral neck, lumbar spine, and combined BMD T-scores.
• To determine optimal BMD input for accurate osteoporosis fracture risk assessment.

Main Methods:

• Calculated 10-year major osteoporotic fracture probabilities using FRAX with various BMD inputs (femoral neck, lumbar spine, minimum, weighted mean).
• Included 20,477 adults aged 50+ from Manitoba, Canada.
• Assessed fracture outcomes via a population-based administrative data repository.

Main Results:

• All FRAX models demonstrated good risk stratification.
• Lumbar spine or minimum T-scores did not consistently improve FRAX performance and adversely affected calibration.
• Weighted mean T-score and spine-hip T-score offset adjustment showed equivalent results, slightly improving risk classification without altering calibration.

Conclusions:

• The selection of bone mineral density (BMD) input significantly impacts FRAX model performance.
• Lumbar spine or minimum T-scores are not suitable inputs for the FRAX algorithm.
• Utilizing a weighted mean of lumbar spine and femoral neck T-scores enhances fracture risk classification accuracy.