Prospective controlled cohort study of 27 adult osteoporotic patients who underwent kyphoplasty for fresh osteoporotic spinal fractures. To define the evolution of vertebral bone mineral density (BMD) at kyphoplasty and adjacent levels along with sagittal spinal alignment to contribute to the etiology of adjacent vertebral fractures after augmentation. Osteoporotic compression fractures can be effectively treated with methylmethacrylate vertebral augmentation. However, to the authors' knowledge the effect of vertebral augmentation on the vertebral endplate BMD of the augmented and adjacent nonaugmented levels has not as yet been described. Twenty-seven consecutive selected patients (9 men, 18 women), with an average age of 72+/-9 years underwent 1, 2, or 3-level percutaneous kyphoplasty for painful fresh osteoporotic vertebral fractures at the thoracolumbar spine. All patients were radiologically examined with plain roentgenograms, computed tomography, and magnetic resonance imaging. Lateral dual energy x-ray absorptiometry in the augmented and on the adjacent vertebrae (1 level above and below kyphoplasty) was used to measure BMD preoperatively to the last postoperative observation in the subchondral bone of the vertebral endplates. Anthropometric data, sagittal global balance (plumbline), and segmental spine reconstruction (vertebral body height, Gardner kyphotic angle) were recorded and analyzed. The patients were followed for at least 2 years. Kyphoplasty was performed between T12 and L5. A total of 48 vertebral bodies were augmented. Thirteen patients received 1 level and the remaining 14 received 2 or 3-level kyphoplasty. No significant changes in the sagittal spinal balance were shown postoperatively. Gardner kyphotic angle and posterior vertebral body height improved postoperatively, however, insignificantly. Significant [analysis of variance (ANOVA), P=0.008] increase of anterior vertebral body height in the fractured vertebra was achieved postoperatively without subsequent loss of correction. BMD increased significantly in the lower endplate of the augmented vertebra (ANOVA, P=0.05). In 1-level augmentation, no BMD changes were shown at the adjacent vertebrae above and below kyphoplasty. On the contrary, in the multilevel augmentation, a statistically significant (ANOVA, P=0.05) decrease of the BMD was shown in the upper endplate of the adjacent level above kyphoplasty. During the 2-year follow-up, there were 5 (18%) new fractures at the T11-T12 area above the augmented vertebra. All of the fractures occurred in patients who received 2 and 3-level kyphoplasty. The observed 2-year evolution of vertebral endplate BMD, after kyphoplasty under stable global sagittal spinal balance, might contribute to the pathogenesis of new fractures in adjacent vertebra. However, other studies with control series and longer follow-up are necessary to show if these BMD changes are the result of vertebral augmentation or are merely natural history.