Case 1. A seventy-four-year-old man with symptoms of
progressive cervical myelopathy underwent an anterior cervical corpectomy and
arthrodesis for decompression of the spinal cord. The spinal cord monitoring
protocol included efferent tceMEP monitoring, afferent subcortical and
cortical SSEP monitoring of the upper and lower extremities, and spontaneous
electromyography5-9.
Muscle sites for tceMEP and spontaneous electromyography recording included
the deltoid, extensor carpi radialis, triceps, first dorsal interosseous,
tibialis anterior, and abductor hallucis muscles. Anesthesia was maintained
through a total intravenous delivery of propofol (125 to 225 µg/kg per
minute), fentanyl (2 µg/kg per hour), and an intermittent low dose (1 to 2
mg) of midazolam. No muscle paralysis was induced except for that required to
effect
intubation7,10,11.
Presented in Figure 1 is a
stack display of tceMEP data from postposition baseline to surgical closure
(the SSEP data are not shown). Following positioning of the patient, baseline
tceMEPs were characterized by well-defined polyphasic responses across all
muscle groups on both the left and right sides. TceMEP amplitudes were
essentially symmetrical for homologous muscles on the left and right sides,
except for the abductor hallucis response, which was smaller on the left.
During the corpectomy stage of decompression, there was a
neurophysiological alert following an acute unilateral loss in amplitude of
the left tceMEP caudad to C7; that is, there were no responses over the first
dorsal interosseous muscle and only fragmentary responses from the tibialis
anterior and abductor hallucis muscles
(Fig. 1). Motor responses on
the right side were unaffected. The ulnar nerve SSEPs, on the other hand, were
essentially unchanged. At that time, the mean arterial blood pressure was
elevated above 85 mm Hg to ensure adequate spinal cord
perfusion4.
Moreover, steroid therapy consisting of high-dose methylprednisolone (NASCIS-2
[National Acute Spinal Cord Injury Study-2]) protocol: intravenous
administration of a 30-mg/kg loading dose followed by 5.4 mg/kg per hour for
the next twenty-three
hours)12 was
initiated for intraoperative treatment of a possible evolving spinal cord
injury.
Despite these interventional measures, no improvement in the motor
responses was seen over the next hour, during which an iliac crest bone graft
and a cervical plate were placed. During closure, there was a second,
precipitous (=95%) decrease in amplitude of the right tceMEP caudad to C7,
progressing to complete response loss. The anesthetic state of the patient was
lightened immediately, and a wake-up
test13 was
performed. As predicted by the neurophysiologic data, the patient was able to
shrug his shoulders but had weak grips and was unable to move either lower
extremity.
Consequently, the wound was reexplored in an attempt to correct any
reversible causes of the neurological deficit. The cervical plate and bone
graft were removed, and the decompression trough was inspected. At that time,
the tceMEPs showed no motor conduction caudad to C7 on the right side or
caudad to C8-T1 on the left (Fig.
1). Moreover, the ulnar nerve SSEPs were absent bilaterally.
An organized epidural hematoma was noted anterior to the spinal canal,
causing compression of the anterior dura. In addition, a single epidural vein
was bleeding briskly along the lateral side of the dura. The hematoma was
evacuated, and the venous bleeder was coagulated with bipolar cautery. On
evacuation, there appeared to be reemergence of motor responses from the
intrinsic muscles of the right hand, although they were >90% reduced from
baseline, but there was no indication of improvement of motor responses from
the recording sites on the right lower extremity. On the left side, there
appeared to be an abnormally delayed and temporally distorted response
fragment from the tibialis anterior muscle with no discernible response from
the corresponding abductor hallucis muscle. There also was reemergence of a
small SSEP on the left side, but not on the right side, following evacuation
of the hematoma.
The graft and the plate were repositioned, the wound was closed, and the
patient was awakened for reexamination. Neurologic examination at that time
revealed improved motor strength, with grade-3 of 5 strength in the intrinsic
muscles of the hands and the tibialis anterior muscle bilaterally. By the
following morning, the patient had improved strength (grade 4 of 5) across all
muscle groups. His strength had returned to normal by the six-week
postoperative office visit.
Case 2. A thirty-seven-year-old woman underwent an anterior
cervical corpectomy of C5, a discectomy at C6-C7, and an arthrodesis of C4 to
C7 for progressive cervical spondylotic myelopathy. The spinal cord monitoring
protocol was the same as that described for the previous patient (Case 1).
Baseline tceMEPs were clearly present bilaterally, as depicted in
Figure 2. Baseline cortical
SSEPs for the ulnar and posterior tibial nerves were within normal limits
(data not shown). The corpectomy and discectomy were performed, and the
arthrodesis was accomplished with use of allograft bone grafts and a cervical
plate. A Penrose drain was placed in the wound prior to closure. There were no
changes in the tceMEPs and the SSEPs for the upper and lower extremities
throughout the entire surgical procedure. The patient emerged from anesthesia
with full muscle strength in all four extremities.
Approximately one hour into the postoperative period, the patient had dense
lower-extremity paralysis. On physical examination, grade-0 of 5 strength was
found in all lower-extremity muscle groups, with no sensation to pinprick up
to the ulnar aspect of the hands. The patient had grade-3 of 5 grip strength
bilaterally and diminished rectal tone. The decision was made to explore the
cervical spine immediately.
Following induction of anesthesia, the baseline tceMEPs
(Fig. 2) of the upper
extremities were found to be reduced in amplitude by >90% and the lower
extremity tceMEPs were completely obliterated. Likewise, the SSEPs of both the
ulnar and posterior tibial nerves were characterized by a marked reduction in
amplitude. This reduction was most pronounced in the lower extremities and was
consistent with the findings of the neurologic examination.
The surgical wound was reopened, and the cervical plate and both bone
grafts were removed. Exploration of the corpectomy site revealed a dense layer
of hematoma anterior to the dura with slow venous bleeding along the remaining
vertebral body. The organized hematoma was carefully peeled off the underlying
dura, the decompression sites were irrigated, and hemostasis was obtained. At
that time, the steroid protocol
(NASCIS-2)12 was
started, and the mean arterial pressure was elevated above 85 mm Hg to
facilitate adequate spinal cord
perfusion4.
Over the course of the next hour, the tceMEPs on the left side showed
gradual improvement as evidenced by the reappearance of responses from the
lower extremity and the intrinsic muscles of the hand. Similarly, the motor
response of the right tibialis anterior muscle reemerged; however, motor
responses from the flexor muscle of the right foot and the intrinsic muscles
of the hand remained absent. Although the SSEPs for the ulnar nerve were
detected, there was still no measurable improvement in amplitude; however, the
SSEPs for the posterior tibial nerve did show a minor increase in amplitude
and were less labile.
The bone grafts and cervical plate were repositioned, and the wound was
closed. On emerging from anesthesia, the patient was able to move both the
upper and the lower extremities against gravity. On physical examination on
the first postoperative day, she demonstrated full motor strength in all
extremities.
An expanding hematoma following an extensive anterior cervical
decompression is a potentially reversible cause of postoperative neurologic
compromise. The clinical scenario often involves a gradual or rapid loss of
function in a patient with an otherwise stable neurologic status immediately
following surgery. Prompt diagnosis and urgent reexploration are paramount to
salvaging spinal cord function in this situation.
Intraoperative tceMEP monitoring promotes rapid detection of injury to the
spinal motor tracts and is proving to be considerably more sensitive than
conventional SSEP monitoring, which carries an unacceptably high rate of
false-negative results in cervical spine
surgery7,8,14.
Presumably, the low sensitivity of SSEPs is due to the fact that they reflect
the functional integrity of the dorsal sensory spinal cord tracts. In
contrast, tceMEPs, which are mediated by the corticospinal tracts, facilitate
the direct assessment of ventral spinal motor pathways. When combined,
therefore, tceMEPs and SSEPs provide neurophysiological monitoring coverage of
both the anterior motor and posterior sensory spinal cord tracts,
respectively.
Schwartz et
al.4-6
argued for a standard proactive approach to intervention in the presence of
notable changes in neurophysiological monitoring. If there is a
neurophysiological indication of neurologic catastrophe, the surgeon should
consider all possible reversible causes of iatrogenic injury to the spinal
cord. These may include bone graft or instrumentation impingement,
intervertebral overdistraction, vascular compromise, intraoperative
hypotension, and epidural hematoma. In these situations, the mean arterial
blood pressure should be elevated routinely above 85 mm Hg to increase the
perfusion pressure to the spinal cord. If neurophysiological changes do not
resolve within fifteen to twenty minutes, the initiation of a NASCIS-2 steroid
protocol may be considered to minimize the exacerbating effects of edema,
although the true benefits of this interventional protocol remain open to
question5.
The cases of our two patients demonstrate two scenarios in which
intraoperative neurophysiological monitoring of motor evoked potentials was
helpful in the treatment of an evolving epidural hematoma. The case of the
first patient highlights the importance of continuing neurophysiological
monitoring through the end of the procedure, until the patient emerges from
anesthesia. Early neurophysiological detection of functional motor
deterioration during wound closure led to the rapid evacuation of an
underlying epidural hematoma, which in turn led to a complete recovery of
motor function. The case of our second patient illustrated the utility of
neurophysiological monitoring to evaluate spinal cord recovery following
corrective action. The development of a rapid onset of postoperative weakness
despite stable tceMEP and SSEP signals at wound closure raised the suspicion
for an expanding hematoma and directed the decision to return to the operating
room to perform immediate evacuation of the hematoma. The efficacy of this
intervention was confirmed intraoperatively by a partial return of tceMEPs
from the lower extremities and was verified postoperatively by the improved
clinical function of the patient after she had emerged from anesthesia.
Finally, SSEP monitoring alone may be inadequate, given the complexity of
current corrective surgery to the
spine7,8.
Had SSEP monitoring alone been used in our first patient, the neurologic
sequela from the hematoma would not have been appreciated until after he had
emerged from anesthesia, thereby delaying intervention. ?
Note: The authors acknowledge the contributions of Drs. Anthony
Sestokas, Alex R. Vaccaro, and Todd J. Albert, who provided thoughtful
editorial comments for the manuscript. The authors also acknowledge the
contributions of Craig Matsumoto, Cheryl Wiggins, and Andrew Warrington, who
were the attending surgical neurophysiologists for the cases reported herein,
and whose expert interpretation of the neurophysiological monitoring data
permitted a rapid and appropriate intervention for these two patients.