PURPOSE To determine the efficacy of percutaneous vertebroplasty in treating severe vertebral body compression fractures, or vertebra plana, in patients with osteoporosis. MATERIALS AND METHODS In 155 patients, 310 percutaneous vertebroplasties were performed during 25 months and 15 days. Of these, 37 patients (27 women, 10 men; mean age, 73.6 years) underwent 48 vertebroplasties for severe osteoporotic vertebral body compression fractures. The fractures were defined as vertebrae that have collapsed to less than one-third of their original height. Imaging and clinical features were analyzed, including the extent of vertebral collapse, location of the involved vertebra, pattern of vertebral compression, volume of polymethylmethacrylate injected, vertebroplasty complications, and clinical outcome. RESULTS Vertebral body collapse averaged 23% (range, 4.5%-33.0%) of the original height. Involved vertebrae were located from levels T5 to L5, with one-half affected at the thoracolumbar junction. Patterns of vertebral compression were divided into gibbus (31 of 48 or 65%), plana (13 of 48 or 27%), and H shape (four of 48 or 8%). The mean volume of the cement injected was 6.0 mL (range, 1.5-12.5 mL). Complications observed on radiographs included cement leakage to the adjacent disc (17 of 48 or 35%) and the paravertebral soft tissues (four of 48 or 8%). There were no major complications. At clinical follow-up (mean duration, 11 months and 3 days; range, 3-24 months), pain relief was complete in 14 (47%) of 30 patients, partial in 15 (50%), and unchanged in one (3%). No patient required surgery. CONCLUSION Percutaneous vertebroplasty for severe osteoporotic vertebral body compression fractures is safe and effective and should not be withheld in this group of patients.