The technique of percutaneous vertebroplasty was developed
by Galibert et al.1 to manage
refractory vertebral-body compression fractures secondary to osteoporosis.
It involves injection of polymethylmethacrylate into the collapsed
vertebral body in order to augment the deficient bone and to relieve
pain. It also requires placement of a large-bore (10-gauge) spinal
needle under fluoroscopic guidance into each affected vertebral
body through a posterior bilateral transpedicular approach2.
Vertebroplasty has been performed for approximately 200 patients
with osteoporotic vertebral collapse2-10,
but not a single instance of spinal-cord compromise or major extrusion
of polymethylmethacrylate into the spinal canal has been described.
I am aware of only one instance in which a peripheral neuropathy
was attributed to the leakage of cement into a neural foramen, and
this transient complication resolved spontaneously3.
The current case report documents the catastrophic extrusion
of polymethylmethacrylate, both into the spinal canal and into neural
foramina at multiple levels of the thoracolumbar spine, following
vertebroplasty in an osteoporotic patient.
A sixty-six-year-old woman with multilevel vertebral osteopenia
and recent spontaneous compression fractures of the tenth and eleventh
thoracic and first lumbar vertebral bodies presented to a local neurosurgeon
because of intractable pain, which had been unrelieved by nonoperative
care over a three-month period (Fig. 1). The findings on neurological
examination were normal.
In September 1999, a percutaneous vertebroplasty was performed
with the patient under general endotracheal anesthesia and positioned
prone on a Jackson spinal frame. Barium-impregnated polymethylmethacrylate
was injected through a 10-gauge needle, which was initially positioned
in the posterior aspect of each vertebral body over a Kirschner wire
and was inserted through a unilateral transpedicular approach under
fluoroscopic control. Prior to the injection of polymethylmethacrylate, vertebral
phlebography was performed at each level to confirm needle placement
and to ensure that no contrast medium extravasated from the confines
of the vertebral cortices. Although biplanar fluoroscopy was used,
clear visualization of the pedicles and vertebral cortices was not
possible because of osteopenia and vertebral collapse and because
the polymethylmethacrylate had not been enhanced by tantalum or
tungsten.
Some extravasation of the liquid polymethylmethacrylate was noted
fluoroscopically at the first lumbar level, although the extent
of this leakage could not be determined intraoperatively. There was
no record indicating the amount of polymethylmethacrylate that was
injected.
Immediately postoperatively, the patient complained of severe
pain radiating from the low back to the lower abdomen and left groin,
buttocks, and thigh, with associated hypoesthesia in the left tenth thoracic
through the second lumbar dermatome. Thereafter, as she resumed
walking, she complained of progressive bilateral sciatic claudication, which
was more severe on the left than on the right. However, the preoperative
mechanical thoracolumbar pain had been relieved.
Computerized tomography of the lower thoracic and lumbar spine,
performed thirteen days postoperatively, revealed extensive extravasation
of polymethylmethacrylate into the spinal canal from the tenth thoracic
through the second lumbar level, resulting in moderate circumferential
constriction of the thecal sac, particularly at the eleventh thoracic
and first lumbar levels (Figs. 2-A and 2-B). In addition, there was a suggestion
of at least a partial left-sided obstruction of the neural foramina at
the twelfth thoracic and first lumbar levels (Figs. 2-B and 2-C), which corresponded
with the findings on the neurological examination.
When the patient was initially evaluated by me, three months
postoperatively, she had classic symptoms of acquired thoracolumbar
spinal and foraminal stenosis, with cramping thigh pain and progressive
numbness and weakness of both legs after walking only a short distance;
the symptoms were worse on the left than on the right. She required
a walker and needed intermittent urinary catheterization. Bowel
function, however, remained normal, and the severe thoracolumbar
pain that she had had preoperatively also remained relieved.
The neurological symptoms and findings did not worsen during
the twelve months following the complication. The hard polymethylmethacrylate compressed
and encircled the dura, and it may have adhered to it and to the
nerve-root sheaths at multiple levels. Because of the high risk
of producing additional neurological complications by attempting
to remove the cement, I chose not to operate and will only consider
doing so if the neurological deficit worsens in the future.