The severity of vertebral fractures, defined by qualitative features and vertebral height reductions using visual semiquantitative (SQ) analysis of radiographs, is a strong predictor of vertebral fractures (Delmas et al., Bone 2003). Vertebral fractures occur when the compressive load applied exceeds the bone strength of the vertebra.  Reduced bone strength is a function of decreased bone density and microarchitectural deterioration of trabecular bone.  This study describes the relationship at baseline between SQ grade and trabecular microarchitecture analyses of iliac crest biopsy samples in a subset of patients from the teriparatide Fracture Prevention Trial (Neer et al., N Eng J Med 2001). Bone structure indices measured by 2D histomorphometry and 3D microCT in 74 and 51 patients, respectively, are shown for each fracture grade, classified by vertebral height loss .

	SQ0 (<20%)	SQ1 (20-25%)	SQ2 (25-40%)	SQ3 (>40%)	p*
2D trabecular bone volume	0.25 ± 0.09 (n=6 )	0.17 ± 0.07 (n=23)	0.14± 0.06 (n=35)	0.12 ± 0.05 (n=10)	0.009
2D marrow star volume	16.1 ± 24.0 (n=6)	22.0 ± 30.9 (n=22)	35.0 ± 36.1 (n=31)	41.8 ± 33.6 (n=9)
3D trabecular BV/ TV	0.16 ± 0.05 (n=4)	0.14 ± 0.07 (n=19)	0.13 ± 0.05 (n=19)	0.08 ± 0.03 (n=9)	0.07
3D Structure model index	1.41 ± 0.33 (n=4)	1.79 ± 0.55 (n=19)	1.96 ± 0.46 (n=19)	2.12 ± 0.49 (n=9)

   *p-value from Wilks' Lambda multivariate statistical test; mean ± SD

There were statistically significant correlations observed among the 2D and 3D measurement groups (r = -0.64, p < 0.01 and r = -0.82, p < 0.01, respectively).  Given the small sample size, consistent SQ versus bone architecture relationship, and strong inter-dimensional measurement correlations, a multivariate ANOVA was performed for each set of 2D and 3D measurements.  The results of both the 2D and 3D analyses suggest that a subject's maximum SQ grade is a strong predictor of the bone microarchitecture impairment.  An increase in vertebral fracture severity is characterized by a reduction in trabecular bone volume, impaired trabecular connectivity, and a transformation of trabeculae from plate-like to more rod-like morphology.   These results demonstrate that a deterioration of bone structure, with low trabecular bone volume, is a continuous and progressive process. Consistent with the suggestion from Parkinson and Fazzalari (J Bone Miner Res 2003), microarchitectural deterioration is exponential when bone volume falls below a critical value of  ~15%. In conclusion, this correlation between baseline SQ vertebral fracture severity and histological features of bone quality deterioration explains the accelerated cascade of fracture risk observed in patients with severe vertebral fracture.