Percutaneous Vertebroplasty for Spinal Metastases: Complications

Radiology. 2006 Jan;238(1):354-62. doi: 10.1148/radiol.2381040841.


Purpose: To retrospectively evaluate complications of percutaneous vertebroplasty (PV) performed with polymethylmethacrylate cement to treat pain in patients with metastases to the spine.

Materials and methods: This study had institutional review board approval; patient informed consent for the review of records and images was not required. In 2 years, 117 patients (38 men [32.5%] and 79 women [67.5%]; mean age, 58.2 years) underwent 159 fluoroscopy-guided PV procedures to treat 304 vertebrae. Spinal metastases included osteolytic, osteoblastic, and mixed lesions. Complications were characterized as local or systemic. Evaluated data included immediate imaging findings (on radiographs and computed tomographic scans) and clinical findings at 30-day follow-up. Chi2 or Fisher exact testing was performed for univariate analysis of variables.

Results: The primary cancers were breast cancers (45.3%), lung cancers (14.5%), myeloma (7.7%), or other cancers (32.5%). Among the 423 cement leakages identified, 332 (78.5%) were vascular and 91 (21.5%) were nonvascular. Vascular leaks were classified as venous epidural leaks, paravertebral and foraminal plexus leaks, and leaks to the vena cava, while nonvascular leaks included puncture trajectory leaks, paravertebral soft tissue leaks, and diskal leaks. Patients with nonvascular leaks were asymptomatic. Eight (6.8%) patients experienced complications, and seven of these complications were symptomatic. Among these eight patients, six (5.1%) had local complications (puncture site hematoma in two patients and radicular pain [successfully treated with nonsteroidal anti-inflammatory drugs or corticosteroids] in four patients), and two (1.7%) had systemic complications (pulmonary embolism resulting from cement migration through the vena cava). One of the latter patients died. Univariate analyses revealed a significant association between cement migration through the vena cava and pulmonary embolism (P = .001) but not between foraminal venous leakage and radicular pain (P = .123).

Conclusion: Despite numerous technical incidents (leaks), PV-induced complications were rare, leading to the hypothesis that systemic complications are a consequence of intravascular leakage while local complications are a consequence of cement-related irritation, compression and/or ischemia, and/or needle-induced trauma.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Bone Cements
  • Chi-Square Distribution
  • Female
  • Fluoroscopy
  • Humans
  • Male
  • Middle Aged
  • Polymethyl Methacrylate
  • Postoperative Complications / epidemiology*
  • Radiography, Interventional
  • Retrospective Studies
  • Spinal Neoplasms / diagnostic imaging
  • Spinal Neoplasms / secondary
  • Spinal Neoplasms / surgery*


  • Bone Cements
  • Polymethyl Methacrylate