Introduction
Neurologic injury from surgery results in substantial increased morbidity, mortality, and cost, and, most important, it is devastating to patients and their families. Thus techniques to lessen, reverse, and even avoid neurologic injury are very valuable. Neurologic intraoperative electrophysiologic monitoring (NIOM) can identify impending or ongoing intraoperative injury, thus allowing for interventions. Changes to a patient’s neurologic electrophysiologic baseline values during the procedure alert the operative team that a potential injury may be occurring. The goal of NIOM is to detect dysfunction caused by ischemia, mass effect, stretch, heat, and direct injury in real time before it causes permanent neurologic injury. Monitoring may also be useful in identifying and preserving neurologic structures during a procedure where they are at risk (mapping).
There are several challenges to establishing the efficacy of NIOM. The first is that blind or randomized trials assessing the efficacy of NIOM in humans are lacking. Unfortunately, a substantial trial will likely never examine this issue. The reason behind the lack of high-level evidence is that monitoring is well-established and accepted in clinical practice. Moreover, it is generally extremely low risk to the patient. The general consensus in the surgical community is that monitoring is useful and there would be ethical and medicolegal dilemmas in withholding monitoring in patients who are at potential risk of injury. A second limitation in establishing outcomes for NIOM is that the goal of monitoring is to reverse a significant change if one is seen during a procedure. Thus monitoring may detect an impending injury, which is reversed, but the benefit can never be confirmed because the patient wakes up with a normal examination. The utility of monitoring is based on animal studies and case series with comparisons to historical control subjects. The utility of NIOM may be supported by establishing that monitoring can, in fact, detect injury in cases where injury has occurred (true-positive outcomes), and limiting false-negative outcomes (injury occurred and was not detected) and persistent false-positive outcomes (injury was predicted by NIOM at the end of a procedure but did not occur). Multimodality monitoring is possible, so the ability of different NIOM techniques to predict injury can be compared in the same patient.
Therapies
Various portions of the nervous system can be monitored by using several NIOM techniques. The specific neurologic tissues at risk, as well as the type of potential injury, vary with different surgical procedures. Specific techniques include electroencephalographic (EEG) and evoked potentials, including somatosensory evoked potentials (SSEPs), brainstem auditory evoked potentials (BAEPs), visual evoked potentials (VEPs), electromyography (EMG), nerve conduction studies (NCSs), and transcortical electrical motor evoked potentials (TcMEPs).
EEG is a measure of spontaneous electrical brain activity recorded from electrodes placed in standard patterns on a patient’s scalp or directly on the cortex with sterile electrode strips or grids. The differences in activity between individual electrodes is amplified and then recorded as continuous wavelets that have different frequencies and amplitudes. These data can be displayed as a raw EEG on a display in a series of channels or broken down into the basic components of frequency and amplitude and displayed as a spectral analysis. A change in a patient’s background EEG activity from baseline during a procedure may indicate ischemia of the cerebral cortex either focally or through a generalized loss of activity over the entire cortex. A 50% decrease in EEG amplitude is generally considered a significant change. EEG is routinely used intraoperatively during carotid endarterectomy (CEA), cerebral aneurysm, and arteriovenous malformation surgery or in other procedures that place the cortex at risk.
Evoked potentials are measures of nervous system electrical activity resulting from a specific stimulus that is applied to the patient. Electrodes record responses to repetitive stimuli as averaged wavelets at different locations in the nervous system as this evoked activity propagates along its course.
SSEPs are produced by repetitive electrical stimulation of a peripheral nerve while averaged potentials are recorded as they travel through the afferent sensory system. SSEP waveforms are recorded from peripheral nerve, spinal cord, brainstem, and primary somatosensory cortex. The recording of waveforms at sequential locations along the complete afferent sensory system allows for localization of dysfunction during procedures. This dysfunction could be caused by ischemia, mass effect, or local injury. SSEPs recorded from stimulation of the median nerve are used intraoperatively during CEA and intracranial surgery for anterior circulation vascular lesions. SSEPs recorded from stimulation of the posterior tibial nerve in the leg are used during intracranial surgeries involving vascular lesions in the posterior cerebral circulation. Monitoring both upper and lower extremity SSEPs during procedures that place the spinal cord at risk may be useful in procedures to treat scoliosis, spinal tumors, or descending aortic repairs. The accepted criterion for significant SSEP change, suggesting a potential injury, is a decrease of spinal or cortical amplitudes by 50% or an increase in latency by 10% from baseline.
BAEPs are wavelets generated by the auditory nerve and brainstem in response to repetitive clicks, delivered to the ear. Typically, five wavelets are recorded from electrodes placed near the ear: the first recorded wavelet represents the response from the peripheral cochlear nerve, and the next four wavelets are generated from ascending structures in the brainstem. Changes in latency and amplitude of these five waves are used to assess the integrity of the auditory pathway during procedures that put them at risk. BAEPs are commonly used in posterior fossa neurosurgical procedures such as acoustic neuroma resections, which place the eighth nerve at risk from either ischemia or stretch injury. BAEPs may also be useful in identifying and preventing injury in procedures such as tumor resections or arteriovenous malformation repairs that place the brainstem itself at risk because of ischemia or mass effect.
VEPs are wavelets generated by the occipital cortex in response to visual stimuli (typically flashing lights delivered with light-emitting diode [LED] goggles in the operative setting). VEPs are recorded from electrodes overlying the occipital cortex and provide information about the integrity of the visual pathway during procedures. VEPs have been have been monitored during neurosurgical procedures involving mass and vascular lesions near the optic nerve and chiasm.
EMG and NCSs can be performed on both peripheral and cranial nerves to assess their integrity and to localize these nerves by recording compound motor action potentials (CMAPs) from the muscles they supply. Monitoring involves placement of pins or electrodes in muscles and identification of the nerve supplying the muscle by stimulating it during the procedure (mapping). NCSs involves determination of whether a specific length of nerve will conduct electrical activity between a stimulating and recording electrode. If a nerve does not conduct the signal, this may indicate that it has been significantly injured along its course. Peripheral nerves are at risk of crush, stretch, ligation, ischemic, and hyperthermic injury during many surgical procedures due to malpositioning, electrocautery, or direct injury. Monitoring is also performed by observation of spontaneous activity from the muscle, which may indicate that a nerve supplying it is suffering unexpected injury. Cranial motor nerves are often monitored in this fashion. Monitoring has been performed on oculomotor, trochlear, abducens, trigeminal, facial, glossopharyngeal, vagus, spinoaccessory, and hypoglossal motor nerves. Cranial nerve VII (facial nerve) is often monitored during posterior fossa procedures where it is at high risk of injury and also during parotid gland procedures or other ear, nose, and throat (ENT) procedures involving the face, ear, or sinuses. The external branch of the superior laryngeal nerve (EBSLN) and recurrent laryngeal nerve (RLN) can be injured during thyroidectomies and other ENT procedures in the anterior neck and have been monitored by detection of movement in the vocal cords. All peripheral nerves in the extremities and trunk can similarly be monitored. Monitoring of peripheral nerves can aid in localizing and protecting nervous tissue during nerve repairs or during spinal surgery for structural repairs or tumor resections.
TcMEPs are measured after an electrical current is delivered to the motor cortex from electrodes on the scalp and a recording is made of either motor evoked potential (MEP) waveforms (D and I waves) from epidural electrodes near the spine itself or myogenic evoked potentials from muscles (CMAPs) in the upper and lower extremities. MEPs may also be recorded by direct electrical stimulation of the motor cortex after craniotomy (as a means of functional mapping of the motor cortex) or via transcortical magnetic stimulation. TcMEPs provide a real-time assessment of the descending motor pathway from the cortex to muscle during procedures that place the corticospinal tracks at risk. TcMEPs are increasingly being used in advanced neurosurgical, aortic, and orthopedic centers for monitoring motor pathways of the brain and spinal cord during procedures. MEPs appear to have a superior temporal resolution for detection of ischemia compared with SSEPs (less than 5 minutes versus 30 minutes). This is likely because TcMEPs measure spinal gray matter, which is very sensitive to ischemia, in addition to spinal motor myelinated tracts. One downside is that no clear criteria exist in the literature to define a critical change warning that injury is occurring. Studies have used different losses in CMAP amplitude (25% versus 50% versus 80%) or threshold changes (i.e., the amount of stimulation current it takes to obtain the CMAP) to signify a critical change. The ability to perform TcMEPs is also limited by its sensitivity to anesthetics, paralytic agents, and temperature. The use of paralytic agents is discouraged and, if used at all, should be extremely limited and kept relatively constant (at less than 40% neuromuscular blockade). This also means that patients are at higher risk of injury due to spontaneous movements or stimulation during their procedures. Another limitation is the major concern that TcMEPs are often difficult to obtain from the leg. Whether this is because of technical limitations of the modality or pre-existing injury in patients is unclear. Complications are of greater concern than in other modalities because of the stimulus intensity required to induce the response; complications may include rare instances of seizures and tongue lacerations. Finally, the establishment of efficacy for TcMEPs has been limited by the lack of approved equipment and experience in performing the technique.
Evidence
Evidence Supporting the Use of Electroencephalography in Carotid Endarterectomy
One of the most common uses of NIOM is EEG during CEA and other intracranial vascular procedures in which the brain is at risk of ischemic injury from hypoperfusion. Although commonly used to monitor CEAs, few data exist to support its use, including a lack of randomized trials. Intraoperative strokes are rare, occurring in approximately 2% to 3% of CEAs, and a large proportion of these strokes are due to embolism. Despite this, it is clear that a small proportion of these strokes are due to hypoperfusion, and it is known from both animal studies and human blood flow studies that loss of EEG activity reflects a reduction of blood flow in the brain. In a large series of 1152 CEAs, a persistent significant change on intraoperative EEG (12 cases) had 100% predictive value for an intraoperative neurologic complication. A critical point during CEA is clamping of the carotid artery so that the endarterectomy can be performed. If ischemia is detected, elevating the blood pressure or placement of a carotid shunt may be used to alleviate the ischemia. Significant EEG changes can occur in up to 25% of cases during carotid clamping; however, strokes do not occur in a majority of these cases even without shunting. In two separate series with a total of 469 patients undergoing CEA with EEG monitoring but without shunting, 44 patients had significant EEG changes and six of these had intraoperative strokes. Although not all patients experiencing EEG changes during CEA in this cohort had strokes, it is possible that the strokes that occurred in this study could have been averted with the use of selective shunting based on EEG. The use of selective shunting based on EEG is further supported by a series of 369 patients in which 73 patients received shunting based on significant EEG changes; no intraoperative strokes occurred. In addition, in another study of 172 patients the use of EEG and selective shunting reduced neurologic complications from 2.3% to 1.1% in 93 patients.
Evidence Supporting the Use of Somatosensory Evoked Potentials to Detect Brain and Spinal Injury
The use of SSEPs to identify early spinal cord injury has become widespread. The risk of spinal injury varies with different surgeries but has been reported to occur in 1% to 2% of scoliosis repairs. Significant changes in SSEPs have been predictive of injury in several small case series in complex cervical and thoracic spine procedures, but false-positives and false-negatives do occur. The risk of injury in cases involving intramedullary spinal lesions, such as tumors, has been reported to be up to 65.4%. In a prospective and retrospective cohort study of 19 patients with adequate baseline SSEP signals undergoing intramedullary tumor resections, SSEPs successfully predicted a postoperative motor deficit in five patients, and there were no false-negative results. In a large survey of 242 experienced surgical groups performing major spinal surgery, neurologic complications occurred twice as often in unmonitored cases as in the monitored cases (51,263 total cases). In the monitored cases, 184 neurologic complications occurred, of which 150 (81%) were predicted by SSEPs, although 34 were not identified, resulting in a false-negative rate of only 0.063%. The authors concluded that SSEP monitoring detected greater than 90% of neurologic injuries with a sensitivity of 92% and a specificity of 98.9%. In a second large series by the same investigators, 33,000 SSEP-monitored spinal cases were retrospectively reviewed. In this survey, 0.75% false-positive, 0.48% true-positive, and 0.07% false-negative rates were reported and yielded a sensitivity of 86.5% and a specificity of 99.2%. Specific data were collected for 77 patients who were injured in this group (30 injuries were severe): 17 false-negative and 60 true-positive outcomes occurred. Of the severe injuries, five were not detected by SSEP monitoring. A retrospective review of 508 patients undergoing cervical spine corpectomies was performed with upper and lower extremity SSEPs. In this series, of 27 significant intraoperative SSEP changes, only one of these was persistent, and this patient awoke with quadriparesis. Of the remaining 26 cases with transient SSEP changes, three patients developed peripheral nerve injuries. Eight additional peripheral nerve injuries went undetected by SSEPs, although no spinal cord injuries went undetected in this series. A similar study of monitoring of upper extremity SSEPs in 182 cervical spine procedures demonstrated the identification eight patients with significant persistent loss of SSEPs. Two of the eight patients developed quadriparesis, and two additional patients developed significant transient motor and sensory symptoms on arousal from anesthesia. No spinal injuries went undetected in this study.
In descending aortic repairs, permanent loss of SSEP signals, indicating spinal ischemia, has accurately predicted paraplegia. Furthermore, good outcomes have been reported when a spinal SSEP change is reversed with maneuvers that improve spinal perfusion in small case series. There is a direct correlation with the time of loss of SSEPs (40 to 60 minutes) and the incidence of paraplegia. However, other data in a nonblinded prospective study of 198 patients undergoing thoracic aortic aneurysm (TAA) and thoracoabdominal aortic aneurysm (TAAA) repairs (99 patients underwent surgery with distal artery bypass and SSEP monitoring versus 99 patients without bypass and monitoring) demonstrated no significant differences in neurologic outcomes between the two groups (8% neurologic complication rate in the SSEP group versus 7% in the unmonitored group). No statistical difference was found after logistic regression analysis between the two groups. In a study of 33 patients undergoing TAAA repair with SSEP monitoring, 16 patients had significant changes in their SSEPs. Five patients developed paraplegia in this group, but no paraplegia occurred in cases without SSEP changes. A majority of the SSEP changes in this cohort were transient, likely because of interventions to reverse these findings; however, five of seven patients who had significant changes to their SSEPs lasting longer than 30 minutes developed paraplegia.
Upper extremity SSEPs have been used for monitoring during CEA. A benefit of using SSEPs over EEG in CEA is that they allow for monitoring of subcortical structures, although EEG does provide neurophysiologic information for a much larger area of cortex. In a meta-analysis of seven large studies assessing the use of SSEPs during CEA in 3028 patients, significant central SSEP changes indicated ischemia in 170 patients (5.6%). Although some of these 170 cases used carotid shunting to reverse significant SSEP changes, 34 patients had an ischemic complication. Eight false-negative results were reported in this analysis, but not every study included in the analysis reported false-negative results. The authors concluded that SSEPs and EEG had similar sensitivities and specificities in detecting ischemia during CEA. Another meta-analysis of 15 studies of 3036 patients identified 10 false-negative cases. Of note, there was some overlap between this analysis and the previous review of seven large studies. This study also looked at the predictive value of significant SSEP changes and concluded that it was poor in predicting outcome and in determining the need for carotid shunting. This was based on comparing similar outcomes in patients undergoing selective shunting with SSEP monitoring and 317 patients who had monitoring but who did not undergo shunting regardless of the changes seen on SSEP.
The utility of SSEP monitoring during intracranial aneurysm repair has also been studied. In repairs of intracranial aneurysms, temporary occlusion of a proximal vessel such as the carotid may be necessary to increase the safety of aneurysm clip placement. During these periods, monitoring with SSEPs may enable longer periods of temporary ischemia, identification of inadequate collateral flow, or identification of malpositioning of aneurysm clips. In a series of 67 aneurysm clippings, 24 significant SSEP changes were noted during temporary clipping, yet only one patient awakened with a deficit. In a similar study involving 58 intracranial aneurysm repairs, 13 significant SSEP changes were demonstrated, of which only one was persistent and resulted in a neurologic deficit. All the transient changes in this study resolved with intervention, including temporary clip removal, permanent clip adjustment, increase in systemic pressure, or retractor adjustment.
Evidence Supporting the Use of Brainstem Auditory Evoked Potentials in Posterior Fossa Neurosurgical Procedures
BAEPs may be used to monitor surgical procedures involving the brainstem and posterior fossa that place the eighth cranial nerve and the auditory pathway at risk. In a series of 144 acoustic neuroma resections, the normal presence of wave V at the end of the resection, regardless of whether a transient change occurred during the procedure, was consistent with preservation of useful hearing. In a retrospective study, 70 patients undergoing microvascular decompression of the trigeminal nerve with BAEP monitoring were compared with 150 unmonitored patients. In the monitored group, none of the patients experienced hearing loss, whereas 10 patients developed hearing loss in the unmonitored group. In a retrospective study of 156 patients undergoing posterior fossa procedures, the permanent loss of wave V was significantly associated with hearing loss. Finally, in a study of 90 acoustic neuroma resections with BAEP monitoring compared with 90 matched historical controls without monitoring, hearing loss was significantly less in those patients with tumors smaller than 1.1 cm who were monitored.
Evidence Supporting the Use of Electromyography and Nerve Conduction Studies
Cranial nerve monitoring is used in operations of the posterior fossa and brainstem. In a series of 104 acoustic neuroma resections in which only 29 underwent facial nerve monitoring with EMG, significantly better outcomes were seen in monitored patients at 1 year. In a study that compared 56 patients with facial nerve monitoring with EMG during parotidectomy with 61 patients who did not have monitoring, early facial weakness was significantly lower in the monitored group—43.6% versus 62.3%—although the incidence of permanent facial weakness was not significantly different. A randomized study examined monitoring of the EBSLN and RLN during 201 thyroidectomies in female patients. It compared visual inspection to identify the nerves versus use of surface electrodes placed on an endotrachial tube inserted between the vocal cords followed by identification of the nerves by direct simulation using a monopolar handheld stimulator by the operating surgeon. The results of this study demonstrated that the intraoperative monitoring (IOM) technique was significantly able to identify the EBSNL more often (83.8%) than visual inspection (34.3%). The RLN was identified 100% of the time in both groups. Most importantly, the presence of postoperative paresis of the EBSLN was significantly less in the IOM group (5% versus 1%), and significantly improved voice parameters were also noted in the IOM group postoperatively. No large studies evaluating the utility of monitoring other cranial and peripheral nerves have been published.
Evidence Supporting the Use of Visual Evoked Potential Monitoring
The evidence supporting VEP monitoring has been sparse in part because of difficulty in obtaining signals in the operating room. Recent improvements in stimulating devices and anesthetic techniques have shown promise in obtaining VEPs in the operative setting. In a recent study of VEP monitoring in 100 patients (200 eyes) undergoing operations that placed them at risk of visual dysfunction, the authors were able to obtain reproducible signals in 187 eyes. Of the remaining 13 eyes, 12 had severe preoperative visual loss and one eye was unable to be recorded because of technical factors. The authors used a new device consisting of 16 red LEDs embedded on silicon disks. The monitoring detected no changes in 169 patients; one patient had visual loss in both eyes in this group, and the loss went undetected by monitoring. In the remaining 17 eyes in which significant intraoperative VEP changes occurred (50% decrease in amplitude), 14 of these patients had significant visual loss. A study in which a group of 22 patients undergoing VEP monitoring during macroadenoma resection was compared with 14 patients undergoing the procedure without monitoring demonstrated no significant difference in visual outcome. Other small clinical case series have also reported no clear benefits of VEP monitoring.
Evidence Supporting the Use of Transcortical Electrical Motor Evoked Potentials in Spinal and Descending Aortic Surgery
The optimal approach to monitor the spinal cord during high-risk procedures is controversial, and it is unclear whether SSEPs or TcMEPs are superior. Patients who may benefit from spinal cord monitoring include those undergoing orthopedic procedures involving structural or vascular lesions and patients undergoing repairs of the descending aorta, which put the spinal cord at risk of ischemia. SSEP monitoring has been the traditional standard, and it has been used in routine clinical practice for spinal procedures since the 1980s. However, SSEPs theoretically monitor only the sensory white matter tracts of the spinal cord, specifically, the posterior columns. The question that arises is whether SSEPs are adequately sensitive for injury to the corticospinal tracts in the cord, which are of primary importance during these procedures. Multiple studies have reported improved outcomes with SSEP monitoring during aortic and spine surgery. As noted previously, significant challenges are associated with TcMEP use. Thus the question is whether TcMEP monitoring provides greater sensitivity to injury of spinal cord structures that are most meaningful to outcome, thereby justifying its use over SSEPs in procedures placing the spinal cord at risk. The fact that TcMEPs may be too sensitive and may identify a significant number of false-positive results may lead to unnecessary interventions during procedures, which, in and of themselves, may lead to injury.
In a study of 142 patients undergoing complex spinal deformity repairs with TcMEP monitoring, 16 patients had significant changes indicating spinal cord motor tract dysfunction during their procedures. In these 16 cases, 11 of the TcMEP changes were reversed during the procedure and no deficit occurred, whereas the five patients with persistent changes awoke with motor deficits. In a cohort of 100 intramedullary spinal tumor resections, TcMEPs were detectable in all nonparaplegic patients. TcMEPs were 100% sensitive and 91% specific, and no patient with stable MEP signals throughout the case awoke with a deficit. Similarly, a study of 50 patients monitored with TcMEPs and SSEPs during intramedullary tumor resection were compared with a group of 50 matched patients without monitoring from a historical cohort of 301 patients. Neurologic outcomes were evaluated at discharge and at 3 months and demonstrated a strong trend at the time of discharge and significant improvement in outcomes at 3 months in the monitored group. Case series have shown a low rate of paraplegia in TAAA procedures when TcMEPs are employed for monitoring. In a study of 75 TAAA repairs, all patients with normal TcMEPs awoke without paraparesis, whereas eight of nine patients with significant changes consistent with spinal cord injury awoke with deficits. Twenty patients in this study had significant MEP changes that resolved intraoperatively, and none of these patients awoke with deficits. Other investigators have demonstrated that significant changes in intraoperative TcMEPs during aortic surgery can be reversed with techniques that increase spinal perfusion, including reimplantation of intercostals and increasing systemic pressure.
Several series have been performed in which TcMEPs and SSEPs were monitored during the same procedure ( Table 58-1 ). This is a rare instance in which head-to-head comparisons have been performed between two monitoring techniques, although analysis of these data is flawed. In all cases, anesthesia was tailored to optimize TcMEPs. Paralytic agents were not used, which increases the difficulty of optimally monitoring SSEPs because of motor artifacts generated from performing stimulation.
Related posts:
Is a Preoperative Screening Clinic Cost-Effective?
Which Patient Should Have a Preoperative Cardiac Evaluation (Stress Test)?
What Is the Optimal Airway Management in Patients Undergoing Gastrointestinal Endoscopy?
What Works in a Patient with Acute Respiratory Distress Syndrome?
Which Are the Best Techniques for Reducing the Incidence of Postoperative Deep Vein Thrombosis?
Aspiration: Is There an Optimal Management Strategy?
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
