Effects of vertebroplasty on endplate subsidence in elderly female spines.

OBJECT The aim in this study was to quantify the effects of vertebroplasty on endplate subsidence in treated and adjacent vertebrae and their relationship to endplate thickness and underlying trabecular bone in elderly female spines. METHODS Vertebral compression fractures were created in female cadaveric (age range 51-88 years) thoracolumbar spine segments. Specimens were placed into either the control or vertebroplasty group (n = 9/group) such that bone mineral density, trabecular microarchitecture, and age were statistically similar between groups. For the vertebroplasty group, polymethylmethacrylate bone cement was injected into the fractured vertebral body under fluoroscopy. Cyclic compression (685-1370 N sinusoid) was performed on all spine segments for 115,000 cycles. Micro-CT scans were obtained before and after cyclic loading to quantify endplate subsidence. Maximum subsidence was compared between groups in the caudal endplate of the superior adjacent vertebra (SVcau); cranial (TVcra) and caudal (TVcau) endplates of the treated vertebra; and the cranial endplate of the inferior adjacent vertebra (IVcra). In addition, micro-CT images were used to quantify average endplate thickness and trabecular bone volume fraction. These parameters were then correlated with maximum endplate subsidence for each endplate. RESULTS The maximum subsidence in SVcau endplate for the vertebroplasty group (0.34 ± 0.58 mm) was significantly (p < 0.05) greater than for the control group (-0.13 ± 0.27 mm). Maximum subsidence in the TVcra, TVcau, and IVcra endplates were greater in the vertebroplasty group, but these differences were not significant (p > 0.16). Increased subsidence in the vertebroplasty group manifested locally in the anterior region of the SVcau endplate and in the posterior region of the TVcra and TVcau endplates (p < 0.10). Increased subsidence was observed in thinner endplates with lower trabecular bone volume fraction for both vertebroplasty and control groups (R(2) correlation up to 62%). In the SVcau endplate specifically, these 2 covariates aided in understanding subsidence differences between vertebroplasty and control groups. CONCLUSIONS Bone cement injected during vertebroplasty alters local biomechanics in elderly female spines, resulting in increased endplate disruption in treated and superior adjacent vertebrae. More specifically, bone cement increases subsidence in the posterior regions of the treated endplates and the anterior region of the superior caudal endplate. This increased subsidence may be the initial mechanism leading to subsequent compression fractures after vertebroplasty, particularly in vertebrae superior to the treated level.