Intraosseous Venography during Percutaneous Vertebroplasty: Is It Needed?

In the last 15 years, neuroradiologists in Europe and America have pioneered the technique of percutaneous vertebroplasty for the treatment of pain and decreased mobility associated with osteoporotic and pathologic vertebral-body compression fractures. Multiple reports in the medical literature detail the clinical experience and technique of percutaneous vertebroplasty. In this issue of the AJNR, Kaufmann et al (pages 601–604) address a particular conundrum concerning the technique and performance of vertebroplasty by investigating the relevance of antecedent venography in percutaneous vertebroplasty for treatment of osteoporotic compression fractures. Opinions differ, and the utility of antecedent venography in determining improved clinical outcomes or decreased complications during vertebroplasty is controversial. The authors retrospectively reviewed results in consecutive patients treated with percutaneous vertebroplasty for vertebral-body compression fractures. The first group of patients underwent antecedent venography, whereas the second group was treated without venography. The clinical outcomes were assessed by means of quantitative measurements of pain and mobility. The intraosseous venograms and postvertebroplasty radiographs were evaluated for extravertebral polymethylmethacrylate (PMMA) deposition, amount of extravasation at each treated level, and correlation between results with venography and results with vertebroplasty. The authors found that improvements in pain and mobility did not differ between the two treated groups. Similarly, the demonstrated extravasation of PMMA cement was not significantly different between the two groups. Interestingly, in 14 (64%) of 22 patients in the group who underwent antecedent venography, correlative extravasation was shown with venography. The authors concluded that antecedent venography did not significantly augment the effectiveness or safety of percutaneous vertebroplasty procedures performed by this group of experienced interventional neuroradiologists.

When percutaneous vertebroplasty was initially evaluated and performed in North America at the University of Virginia beginning in 1993, antecedent venography was an integral part of this procedure and performed in every patient. In the initial report about the technique and the early clinical outcomes with this technique in the treatment of painful osteoporotic compression fractures, Jensen et al (1) advocated the use of antecedent venography to decrease potential complications associated with incorrect or suboptimal needle placement in the basivertebral venous plexus or in direct connection with a paravertebral vein. The purpose is to delineate a potentially dangerous route by which PMMA cement might escape the confines of the vertebral body. PMMA cement can escape posteriorly into the spinal canal, causing spinal canal stenosis or cord compression; to the intervertebral foramina, causing nerve root compression; or to the vena cava and pulmonary arteries, causing pulmonary embolism. With a right-to-left cardiac shunt, such as that in patent foramen ovale or ductus arteriosus, a potential stroke may occur, although it may never be reported. This event may necessitate needle readjustment or, as the authors suggest, it may require the use of maneuvers to prevent potential cement extravasation. Such maneuvers include 1) the placement of Gelfoam pledgets prior to injection of PMMA or 2) an initial deposition of PMMA and a period of waiting to allow the cement to harden and obliterate the direct venous connection(s) and then the use of a second needle to inject the cement into the vertebral body. However, some authors (2–4) disagree with this approach and state that, due to differences in the viscosity and flow characteristics of contrast material and cement, venographic findings are not predictive of the actual flow of PMMA cement and path of extravasation. Additionally, the persistence of intravertebral opacification could obscure visualization of cement, for example, during an injection into necrotic cavities in cases of vertebral osteonecrosis or Kümmell disease (5) or during an injection through the endplates to the intervertebral discs. Preoperative imaging with demonstration of intravertebral gas or a fluid collection should alert one to the possibility of vertebral osteonecrosis; therefore, the venographic procedure should be modified by injecting contrast material gently and by using the minimal volume of contrast agent (1–2 mL).

Although venography may not augment the safety of vertebroplasty when it is performed by experienced operators, it may guide novice or inexperienced operators and help them to perform vertebroplasty in a safer manner. In addition, although venographic findings may not be absolutely correlated (64% in Kaufmann et al’s series) with the actual extravasation of cement, the additional information that the delineation of the venous anatomy around the vertebral body may potentially be of benefit. Percutaneous vertebroplasty is an intravascular procedure because the vertebral bony trabeculae is a large venous space with eventual connections to the draining veins. PMMA cement should be considered to be a liquid embolic agent, and it should be treated and used with caution and full knowledge of its possible adverse effects. When the venograms are shown on separate monitors, these reference images may provide guidance in the detection of early-appearing and small quantities of extravasated cement. Although cement extravasation does occurs, the volume and amount of extravasation causes potential clinical complications. Therefore, the early detection and knowledge of cement extravasation is the key to the safe performance of vertebroplasty. Other important factors in the safe performance of vertebroplasty include correct patient selection; good knowledge of vertebral bony and vascular anatomy; adequate opacification of PMMA cement; and high-quality fluoroscopy, preferably with biplane types. Although the authors mention the issue of excessive radiation exposure and the cost of contrast material in the performance of venography, these are likely to have no clinical importance. To my knowledge, no report of the adverse effects of venography exists in the literature. The amount of contrast agent used per venogram is approximately 3–5 mL; therefore, the risk to patients with renal failure is minimal. The authors correctly point out that, in those adept at the performance of vertebroplasty, venography may represent a superfluous step. Venography, however, may be extremely beneficial for less experienced physicians. The use of intraosseous venography should still be advocated during training courses, and its clinical value should be pointed out to the trainees.

The results of this study are valuable and thought provoking. As the authors acknowledge, the study is limited by its retrospective nature, and the sample size may result in a lack of statistical power. Prospective randomized trials to evaluate the effectiveness of vertebroplasty compared with that of sham-vertebroplasty and conservative medical treatment are underway. If possible, an evaluation of the value of antecedent venography in these studies may provide interesting results.