Prediction of osteoporotic fragility re-fracture with lumbar spine DXA-based derived bone strain index: a multicenter validation study.

2020 
A new qualitative index of bone strength, based on finite element analysis and named bone strain index, has been recently developed from lumbar DXA scan. This study shows that BSI predicts subsequent re-fracture in osteoporotic patients affected by fragility fractures. Introduction Dual-energy X-ray absorptiometry (DXA) can provide quantitative (bone mineral density, BMD) and qualitative (trabecular bone score, TBS) indexes of bone status, able to predict fragility fractures in most osteoporotic patients. A new qualitative index of bone strength, based on finite element analysis and named bone strain index (BSI), has been recently developed from lumbar DXA scan. This study presents the validation results of BSI prediction for re-fracture in osteoporotic patients with fragility fractures. Methods In three academic hospitals, 234 consecutive fractured patients with primary osteoporosis (209 females) performed a spine X-ray for the calculation of spine deformity index (SDI) and DXA densitometry for BMD, TBS and BSI at the basal time and in the follow-up at each clinical check. A subsequent fracture was considered as one unity increase of SDI. Results For each unit increase of the investigated indexes, the univariate hazard ratio of re-fracture, 95% CI, p value and proportionality test p value are for age 1.040, 1.017-1.064, 0.0007 and 0.2529, respectively, and for BSI 1.372, 1.038-1.813, 0.0261 and 0.5179, respectively. BSI remained in the final multivariate model as a statistically significant independent predictor of a subsequent re-fracture (1.332, 1.013-1.752 and 0.0399) together with age (1.039, 1.016-1.064 and 0.0009); for this multivariate model proportionality test, p value is 0.4604. Conclusions BSI appears to be a valid DXA index of prediction of re-fracture, and it can be used for a more refined risk assessment of osteoporotic patients.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    30
    References
    3
    Citations
    NaN
    KQI
    []