Spine–hip discordance and fracture risk assessment: a physician-friendly FRAX enhancement

The FRAX® tool estimates a 10-year probability of fracture based upon multiple clinical risk factors and an optional bone mineral density (BMD) measurement obtained from the femoral neck. We describe a simple procedure for using lumbar spine BMD to enhance fracture risk assessment under the FRAX system.

Discordance between lumbar spine (LS) and femoral neck (FN) T-scores is common and a source of clinical confusion since the LS measurement is not an input variable for the FRAX algorithm. The purpose of this study is to develop a procedure for adjusting FRAX probability based upon the T-score difference between the LS and FN (termed offset).

The Manitoba BMD database was used to identify baseline LS and FN dual-energy X-ray absorptiometry examinations (33,850 women and 2,518 men age 50 and older) with FRAX estimates for a major osteoporotic fracture categorized as low (<10%), moderate (10–20%), and high (>20%). Fracture outcomes were assessed from population-based administrative data. An approach was developed and internally validated using a split-cohort design.

The offset was found to significantly affect fracture risk [HR, 1.12 (95% CI, 1.06–1.18) per SD LS below FN] independent of the FRAX probability. The following rule was formulated: “Increase/decrease FRAX estimate for a major fracture by one tenth for each rounded T-score difference between LS and FN.” In the validation subgroup, there was a significant improvement in the fracture prediction using FRAX with the proposed offset adjustment for major osteoporotic (P = 0.007) and vertebral fracture prediction (P < 0.001). For those at moderate risk under FRAX, 12.6% showed reclassification using the offset to a risk level that more accurately reflected their observed risk (25.2% reclassification for moderate risk discordant cases).

A simple procedure that incorporates the offset between the LS and FN T-scores can enhance fracture risk prediction under the FRAX system.