SPINE Volume 31, Number 4, pp E105–E113
©2006, Lippincott Williams & Wilkins, Inc.
Intraoperative Somatosensory Evoked Potential
Monitoring During Cervical Spine Corpectomy Surgery
Experience With 508 Cases
Mustafa H. Khan, MD,* Patrick N. Smith, MD,* Jeffrey R. Balzer, PhD,†
Donald Crammond, PhD,† William C. Welch, MD,† Peter Gerszten, MD,†
Robert J. Sclabassi, MD, PhD,† James D. Kang, MD,* and William F. Donaldson, MD*
Study Design. Retrospective review.
Objectives. To review consecutive cases of cervical
spine corpectomy surgery performed with intraopera-
tive somatosensory-evoked potential (SSEP) monitoring.
Summary of Background Data. There is controversy
about the utility of SSEP monitoring during anterior cer-
vical spine surgery. There is no study in the literature that
has specifically evaluated the utility of SSEP monitoring
for cervical spine corpectomy surgery.
Methods. Intraoperative SSEP tracings for 508 pa-
tients (average age, 55.7 years; 268 male, 240 female)
who underwent anterior cervical fusion with single-level
or multilevel corpectomies were reviewed. Intraoperative
and postoperative records were analyzed to determine if
any new neurologic deficits developed when the patients
woke up from anesthesia.
Results. The overall incidence of a new postoperative
neurologic deficit in this series of patients was 2.4% (11
with nerve root injury, 1 with quadriplegia). The incidence
of significant SSEP changes was 5.3% (27 of 508 patients).
The most common identifiable cause of SSEP changes
was hypotension, and the most common neurologic def-
icit was deltoid (C5) weakness. One patient had irrevers-
ible SSEP changes, and he woke up with new-onset quad-
riplegia. The calculated sensitivity and specificity of
intraoperative SSEP monitoring for detecting impending
or resultant intraoperative iatrogenic neurologic injury
were 77.1% and 100%, respectively. However, if the iso-
lated nerve root injuries are removed from the analysis,
then both the calculated sensitivity and the negative pre-
dictive values were 100%.
Conclusions. Intraoperative SSEP monitoring can alert
the surgeon to adverse iatrogenic intraoperative events
with potential for neurologic injury. Most SSEP signal
changes are reversible and do not result in a clinical
deficit. Isolated nerve root injury appears to be the most
common iatrogenic intraoperative injury during cervical
spine corpectomy surgery.
Key words: cervical spine surgery, corpectomy, somato-
sensory evoked potentials, neurological injury. Spine 2006;
31:E105–E113
Although a relatively rare occurrence, a neurologic defi-
cit is one of the most feared complications of spine sur-
gery. Therefore, there has been great interest during the
past several decades in developing monitoring methods
that can accurately assess the functional integrity of the
spinal cord during spine surgery. One such widely used
technique is somatosensory-evoked potential (SSEP)
monitoring. The routine use of intraoperative SSEP mon-
itoring has led to a dramatic reduction in new postoper-
ative neurologic deficits during scoliosis and anterior
thoracic spine surgery.1,2 However, the usefulness of
SSEP monitoring for other surgical procedures of the
spine, such as anterior cervical spine surgery, has been
the subject of debate in the literature.3–6
Patients with cervical myelopathy often require ante-
rior fusion and decompression with single-level or mul-
tilevel corpectomies.7 Two risks of this procedure are
iatrogenic spinal cord injury resulting in significant neu-
rologic deficits ranging from limb weakness to quadri-
plegia and isolated nerve root deficits producing single
muscle weakness usually in the deltoid muscle (C5) and
occasionally the biceps (C6). Therefore, detecting such
events in real time during the surgery is of critical impor-
tance since it can enable the implementation of appropri-
ate intraoperative countermeasures. Although SSEP
monitoring is widely performed for this purpose during
anterior cervical spine surgery, to our knowledge there is
no large study that has assessed the usefulness of intra-
operative SSEP monitoring specifically for patients un-
dergoing cervical corpectomy surgery. Unfortunately, re-
ports in the literature cited in support of SSEP monitoring
for anterior cervical spine surgery are hampered by
methodologic issues. For example, most studies have
combined data from patients undergoing cervical spine
corpectomy surgery with patients undergoing anterior
cervical discectomy and fusion.6,8 However, because of
underlying pathophysiologic differences between cervi-
cal myelopathy and radiculopathy, these groups need to
be studied separately. Also, since there are significant
technical differences between cervical spine corpectomy
and discectomy surgeries, the results of these surgical
procedures should be interpreted in the appropriate
context. Therefore, the usefulness of SSEP monitoring
for patients undergoing cervical corpectomy needs to be
the subject of a separate investigation.
The goal of this study was to retrospectively review
consecutive cases of cervical spine corpectomy surgery
From the Departments of *Orthopedic Surgery and †Neurological Sur-
gery, University of Pittsburgh Medical Center, Pittsburgh, PA.
Acknowledgment date: June 20, 2005. First revision date: July 15,
2005. Acceptance date: July 19, 2005.
The manuscript submitted does not contain information about medical
device(s)/drug(s).
No funds were received in support of this work. No benefits in any
form have been or will be received from a commercial party related
directly or indirectly to the subject of this manuscript.
Address correspondence and reprint requests to William F. Donaldson,
MD, 3741 Fifth Avenue, Suite 1010, Pittsburgh, PA 15213; E-mail:
donaldsonwf@upmc.edu
E105

performed at our institution over a 10-year period with
intraoperative SSEP monitoring. We wanted to identify
cases accompanied by new postoperative neurologic def-
icits or intraoperative SSEP changes, so that the sensitiv-
ity and specificity of intraoperative SSEP monitoring for
detecting impending or resultant neurologic injury dur-
ing cervical spine corpectomy surgery could be deter-
mined. This is the first, largest, and most detailed study
of its kind to address these issues specifically for cervical
spine corpectomy surgery.
Materials and Methods
Study Design. We retrospectively reviewed consecutive cases
of cervical spine corpectomies with fusion performed at our
institution between 1994 and 2004 by four different surgeons
(J.D.K., W.F.D., W.C.W., P.G.) and monitored by the same
group of neurophysiologists (J.R.B., D.C., R.J.S.). Inclusion
criteria for the study included patients who underwent anterior
cervical fusion with one-, two-, or three-level corpectomies us-
ing allograft or autograft bone. Patients who did not have in-
traoperative SSEP monitoring available or those in whom such
monitoring could not be performed due to technical reasons
were excluded from this study. Patients who did not have a
documentation of postoperative neurologic status were also
excluded. Using these criteria, a total of 508 patients were thus
identified (268 male, 240 female), which included 107 patients
with revision surgery. The average patient age was 55.7 years,
and the most common diagnosis was congenital or spondylotic
cervical stenosis and myelopathy.
Neurophysiologic Monitoring. Intraoperative SSEP moni-
toring is standard at our institution for almost all cases of
cervical spine corpectomy surgeries. Baseline SSEP values were
obtained after the induction of anesthesia before patient posi-
tioning in all cases. Throughout the procedure, continuous up-
per and lower extremity responses were obtained. All patients
were monitored in the same fashion.
Upper Extremity SSEPs. Median or ulnar nerve stimulation
was performed bilaterally in an alternating fashion at the wrist
with subdermal needle electrode pairs. Whether median or ul-
nar nerve was stimulated was based on the cervical level(s)
being decompressed. Pz/Fz and P4/F3 scalp electrodes were
used (per the international 10–20 system). A cervical electrode
was localized at the C7 spinous process and referenced to Fz.
Constant voltage stimulators using sufficient intensity to evoke
a consistent response produced evoked sensory potentials.
Stimulation frequency was 2.45 Hz with duration of 0.2 milli-
seconds. Bandpass filters were set at 33 to 300 Hz with a gain of
20k for cortical recordings and 30 to 1,000 Hz with a gain of
50k for cervical recordings. Averages were computed for either
64 or128 trials, depending on the signal quality.
Lower Extremity SSEPs. For the lower extremities, bilateral
alternating tibial nerve stimulation was used, unless reproduc-
ible responses could not be obtained. In such cases, peroneal
nerve stimulation was employed instead. Tibial nerve stimula-
tion was performed at the ankle with subdermal needle elec-
trode pairs with a proximally placed cathode and an anode
placed approximately 1 cm distally. The peroneal nerve was
stimulated using pairs of subdermal needles located at the head
of the fibula and medially in the popliteal fossa. Recordings
were obtained from the scalp and cervical region with subder-
mal electrodes. Pz/Fz and P4/P3 scalp electrodes were used (per
the international 10–20 system). A cervical electrode was lo-
calized at the C7 spinous process and referenced to Fz. Evoked
sensory potentials were produced by constant voltage stimula-
tors using sufficient intensity to evoke a consistent response.
Stimulation frequency was 2.45 Hz with duration of 0.2 milli-
seconds. Bandpass filters were set at 33 to 300 Hz with a gain of
20k for cortical recordings and 30 to 1,000 Hz with a gain of
50k for cervical recordings. Averages were computed for either
64 or 128 trials depending on the signal quality.
AlarmCriteria. The initial recordings made after induction of
anesthesia and before positioning were used as baseline values.
Continuous SSEP signals were collected (approximately 1 every
40 seconds). We considered a 50% reduction in primary so-
matosensory cortical amplitude or a prolongation of response
latency by
10% to be significant. These criteria are generally
agreed on in the literature as being of optimal sensitivity and
specificity for detecting iatrogenic spinal cord injury.9 –13
Henceforth, these threshold signal changes will be referred to
as “significant.” It should be cautioned that while a 50%
change in amplitude and/or a 10% increase in latency are
widely accepted as being clinically significant, the interpreta-
tion of such a change should be made on a case-by-case basis
with clinical correlation.
Medical Record Review. Medical records for all 508 pa-
tients were reviewed to determine if any new neurologic deficits
developed when the patient woke up from anesthesia. The
medical records were reviewed independently without knowl-
edge of the intraoperative SSEP changes in a blinded fashion.
Any new postoperative motor/sensory deficits or bowel/
bladder changes were considered to be iatrogenic intraopera-
tive injuries. Similarly, the intraoperative SSEPs for all 508
patients were reviewed independently without knowledge of
the postoperative neurologic outcome in a blinded fashion.
Next, the intraoperative records of patients who demonstrated
significant SSEP changes were further scrutinized to determine
if such a change occurred in temporal proximity to specific
intraoperative events (e.g., hypotension, retractor placement,
intervertebral disc space distraction, dislodgement of bone
graft, introduction of instrumentation, accidental durotomy, a
specific operative maneuver). In such cases, intraoperative
records were reviewed to determine if any subsequent intraop-
erative countermeasures (e.g., increasing the mean arterial
blood pressure, retractor repositioning, relieving intervertebral
disc space distraction, bone graft removal, instrumentation re-
moval or readjustment, cessation of operative maneuver) were
undertaken and whether or not the SSEP signals improved fol-
lowing these actions. Such countermeasures will be henceforth
referred to as “interventions.”
Data Analysis. To determine the sensitivity and specificity of
intraoperative SSEP monitoring for detecting impending or re-
sultant iatrogenic neurologic injury, we defined and then clas-
sified each the 508 operative cases as one of the following:
True Positive (TP). Significant SSEP signal changes accompa-
nied by a new postoperative neurologic deficit. Or, a case
where significant SSEP signal deterioration occurred as the re-
sult of a recognized intraoperative cause, event or complica-
E106 Spine • Volume 31 • Number 4 • 2006