Intraoperative Monitoring

Stephen W. Bartol

Abstract

Minimally invasive spinal surgery carries with it an increased inherent risk of nerve injury. These risks must be mitigated and this can be accomplished through the use of nerve mapping and monitoring technologies. In this chapter, we review the state-of-the-art techniques for locating, mapping, and monitoring neurological structures during spine procedures. The history of intraoperative neuromonitoring is reviewed and the rationale for each modality is discussed. Step-by-step techniques are outlined for somatosensory evoked potentials, motor evoked potentials, electromyography, and mechanomyography monitoring. Intraoperative anesthesia concerns and complications associated with each technique are reviewed. The advantages and disadvantages of each technology are discussed in a way that is designed to assist the minimally invasive spine surgeon in selecting the most appropriate monitoring options for each individual procedure. The author also provides his own preferences and procedure-specific recommendations.

Keywords: intraoperative neuromonitoring, somatosensory evoked potentials, motor evoked potentials, electromyography, mechanomyography

14.1 Introduction

The application of intraoperative neurophysiological monitoring (IONM) to spinal surgery has expanded rapidly in the last two decades, and the introduction of minimally invasive techniques to the field of spine surgery has led to greater interest in the use of nerve monitoring and mapping. Injury to neurological structures is an inherent risk of minimally invasive spine surgery and can occur through a variety of mechanisms including:

• Direct trauma, usually to nerves that cannot be directly visualized during tissue dissection or during instrumentation.

• Stretch injury, such as during correction of a spinal deformity or through traction (e.g., spondylolisthesis reduction or over-retraction of the femoral plexus).

• Ischemia.

The overall risk of neurological injury during spine surgery is small, but with growing public awareness and increasing emphasis on safety, even a relatively low risk of injury carries a lot of focus and demands attention.¹^,²^,³ Public expectations are that risks will be managed wherever possible, resulting in increasing use of technology to reduce risks whenever such technology is available.⁴^,⁵ The reported incidence of major neurologic complications (such as paraplegia) after scoliosis surgery ranges from 0.55 to 1.6%.⁶^,⁷^,⁸ The placement of pedicle screws has been associated with varying rates of breeched cortex, and while most breeches are not associated with neurological symptoms, nerve symptoms that result in permanent injury or a need for revision surgery are reported in about 1% of patients.⁹^,¹⁰^,¹¹^,¹²^,¹³^,¹⁴^,¹⁵ The growth in popularity of the lateral, transpsoas approach has heightened attention on nerve complications. Up to 30% incidence of transient nerve symptoms has been reported with this approach.¹⁴^,¹⁶^,¹⁷^,¹⁸^,¹⁹

Complex anatomy emphasizes the need for nerve identification and mapping techniques that reduce the risk of nerve complications.²⁰^,²¹ Numerous studies have shown that various forms of intraoperative monitoring may be useful in reducing neurological complications. Modern intraoperative approaches use a combination of techniques tailored to the risks associated with the particular surgery being planned. In this chapter, we will review historical and recent developments within the field of intraoperative monitoring and discuss the clinical applications. We hope that in understanding these various monitoring techniques, surgeons can choose from the menu of options available to them, and make appropriate choices that minimize the risk of nerve injury during surgery.

14.2 Historical Review

The original form of intraoperative monitoring in spine surgery involved the wake-up test during scoliosis surgery: waking patients on the operating room (OR) table, during the procedure, to assess neurologic function. This was generally done immediately following correction of the spinal deformity (usually scoliosis), at which time patients were told to move their feet and toes. Successful movement indicated preservation of spinal cord motor function, and the patient was then reanesthetized for continuation of the surgical procedure. Now known as the Stagnara wake-up test,²² this test is still in use, despite its obvious limitations, which include the risk of uncontrolled patient movement and postoperative anxiety in patients who have recall of the intraoperative experience.⁷^,²³^,²⁴

While comforting when it works, a survey of orthopaedic surgeons reported poor agreement between the results of the wake-up test and the postoperative outcome, reflecting the inherent difficulties associated with administration of this test.⁷

An alternative (or supplement) to the wake-up test came with the development of the somatosensory evoked potential (SSEP). This electrophysiologic test allows examination of the neuraxis from a peripheral nerve to the central cortex. In normal subjects, electrical stimulation of a peripheral nerve causes a change in the EEG of the patient’s contralateral sensory cortex, which can be studied by averaging many signals over time. This change in EEG can be monitored using EEG electrodes on the cranium. SSEPs were first used in surgery to monitor spinal cord function during scoliosis surgery. Good success was reported early on, and SSEPs were gradually recognized as a viable alternative or adjunct to the use of the wake-up test during spine surgery.²⁵^,²⁶^,²⁷^,²⁸^,²⁹^,³⁰^,³¹^,³²^,³³

It was soon recognized, however, that the SSEP test has significant limitations. Nerve fibers conducting motor command signals from the brain to the spinal cord do not contribute to the SSEP, nor do spinal motor neurons. SSEP is also neither sensitive nor specific for spinal nerve root (lower motor neuron) and peripheral nerve injuries, which are the primary concern with most common lumbar spine procedures. In recognition of these limitations of SSEPs, the field of IONM has been expanded considerably in the last two decades. Motor evoked potentials (MEP) and electromyography (EMG) were introduced to better monitor the motor neuraxis. These offered considerable advances and changed the landscape of neuromonitoring in spine surgery.⁵ Increasingly, however, the demands of surgery are pushing the limits of these technologies. Surgeons need more rapid real-time feedback in order to perform minimally invasive surgery safely and that has led to the push for surgeon-driven, immediate-feedback systems (NuVasive, DePuy Synthes) ( Fig. 14.1).

Fig. 14.1 Surgeon-driven systems provide a user-friendly feedback system that delivers real-time information directly to the surgeon.

A high degree of variability in the interpretation of IONM signals remains a serious issue.³⁴ At the heart of this problem is the tremendous role that electrical noise plays in the signals detected using EEG-type electrodes. Electrical noise picked up by these electrodes causes tremendous variability and complexity of signals.³⁵^,³⁶^,³⁷ Given the need for greater nerve mapping accuracy in minimally invasive surgical procedures, this has led to the push for newer, smart sensor technologies, which offer greater consistency than is possible when using simple electrodes in a complex electrical environment such as the OR. The evolution of mechanomyography (MMG) smart sensor systems (SentioMMG, DePuy Synthes) has occurred largely in response to these needs.

As a result of these various developments, the field of IONM today offers a varied menu of options to the surgeon. Each item in the menu offers advantages and disadvantages, depending on the operative scenario. Choosing the right option(s) for a given procedure is critical if surgeons are to avoid injuring neurological structures, especially during challenging minimally invasive procedures where direct visualization of neurological structures is not possible.

14.3 Somatosensory Evoked Potentials

SSEPs are used to assess dorsal column function within the spinal cord. Specifically, they monitor the integrity of the large fiber sensory system, which is responsible for accurate perception of vibration and joint position sense. SSEPs are obtained by stimulating peripheral nerves (typically tibialis anterior in the lower extremities and ulnar or median in the upper extremities). Recording may be done at various points along the neuraxis (brachial or lumbar plexus, cervical spine, parietal somatosensory cortex) to assess the integrity of the pathway.

The inherent difficulty with SSEP monitoring lies in the low amplitude (0.1–20 μV) of the signals generated and the intense levels of background noise that are created by numerous biologic signals and the ubiquitous electrical equipment found in the OR. Isolation of SSEP signals therefore requires the use of signal-averaging techniques. Signal averaging involves taking the signal of interest, amplifying it, time-locking it to the stimulus, and averaging out random background noise by performing hundreds of successive trials. On average, it takes several hundred stimuli (up to 1,000) to extract a baseline SSEP signal from background noise.

14.3.1 Technique

Typically, baseline signals are recorded at the start of the surgical case. During the procedure, any events that interfere with this sensory nerve conduction will be reflected by diminished amplitude or increased latency of the response at recording electrode(s) proximal to the site of interference.

For operations involving neurologic segments at or above C6, it is common to use the median or ulnar nerves for stimulation; under normal circumstances, median nerve stimulation will give rise to a larger cortical potential compared to the ulnar-mediated SSEP. For operations caudal to the C6 segment, it is common to use the SSEP from stimulation of the posterior tibial nerve at the ankle. If this nerve is not available (e.g., because of fracture, amputation, or severe peripheral neuropathy), it is possible to obtain satisfactory recordings from stimulation of the common peroneal nerve at the knee. The cathode (-ve lead) is placed over the most accessible portion of the nerve, and the anode (+ ve lead) is placed distal to it. The stimulus used is a square wave signal with a pulse duration of 100 to 200 μs, at a frequency that does not divide evenly into 60 (60 Hz being the frequency found most commonly in background electrical noise). Normal nerve recovery time dictates that signal strength diminishes above 5 Hz.³⁸ As a result, we generally use a non–whole number below 5 such as 3.17 Hz. It takes an average of 500 to 1,000 stimulated responses to obtain a strong waveform. It is not uncommon, therefore, for signal averaging to require 2 to 4 minutes in order to develop a satisfactory waveform, meaning that there is almost always a delay associated with starting an SSEP test and obtaining the results. Because of this delay, SSEPs cannot be considered a “real-time” test modality.

Provided that the patient has no neuromuscular block, stimulation should result in a clearly visible twitch in adjacent muscles near the stimulating electrode. In the hand, a thumb twitch is observed, and in the foot the toes will twitch. The appearance of a good motor twitch is a good indication that adequate current levels have been reached. In the upper extremity, this is usually reached at 15 to 20 mA and in the lower extremity at 30 to 60 mA. We rarely increase current up to 75 mA.

Best results are obtained by stimulating left- and right-sided nerves independently. Modern evoked potential machines allow the user to automatically alternate stimulus to the left and right sides and display the two SSEP signals simultaneously. If monitoring of both the upper and lower limbs is being used, then stimulus intervals can be adjusted to allow simultaneous display of all four signals.

If the equipment allows, both upper and lower extremity SSEP monitoring may be used not only for cervical spine surgery but also for thoracic and lumbar spine surgery. Upper extremity SSEP data may help prevent palsy of the brachial plexus and/or other upper extremity palsies resulting from poor positioning or during prolonged procedures.³⁹ Upper extremity SSEP monitoring can also be useful as a comparison check when stimulus malfunction or anesthesia effects are suspected in situations where there is decreased amplitude or loss of responses from the lower extremities.

While SSEP recordings can be done at multiple sites, most commercial systems can accommodate a total of only 16 or 32 channels. Since most of these channels are needed for simultaneous EMG/MEP monitoring, practically speaking, it becomes reasonable to limit the number of channels monitoring SSEPs. When selecting to use SSEPs, such as during spinal manipulation surgery, we therefore routinely use only four sites for SSEP monitoring, leaving plenty of channels for other modalities. We use Cz (for posterior tibial stimulation), C3 and C4 (for upper limb right-sided and left-sided stimulation, respectively), and the cervical spine (except for posterior cervical approaches, in which case we eliminate this lead) ( Fig. 14.2). What we give up with this approach is information about the “traveling wave.” For example, in the lower limb, we lose confirmation of stimulation delivery that would normally be provided by a recording site in the popliteal fossa, and in the upper limb we lose the traveling wave signal in the brachial plexus. However, as long as there is no neuromuscular blockade, stimulation can always be confirmed with either MMG sensors or by feeling under the drapes for muscle twitches at the stimulus site.

An intraoperative baseline should be established after the patient is anesthetized, when the patient has adjusted physiologically to anesthesia. The key parameters of interest during intraoperative SSEP recordings are the latency and amplitude of the responses. Latency prolongation and amplitude attenuation should be considered signs of impairment of spinal cord function. Unfortunately, there are no universally accepted warning criteria as to when the monitoring team should inform the surgeon that a change in the electrophysiologic status of the patient has occurred. Different centers may use any of the following to indicate a significant change in the patient’s condition⁸^,⁴⁰^,⁴¹^,⁴²:

• Thirty to fifty percent decrease in the amplitude of the potentials.

• An increase of 2.5 ms in the response latency.

• Five to ten percent increase in the response latency.

• Any combination of the above.

We generally employ the following alarm criteria for SSEP changes: a decrease of greater than 50% in the amplitude of the response or an increase of greater than 10% in the latency of the response, or a sudden loss of evoked potentials. When any of these situations occur, the surgeon is immediately informed of the changes and surgical countermeasures are instituted when appropriate. If the loss is distal to the level of surgery, the surgeon will generally reverse any mechanical event that immediately preceded the change in SSEP. Alternatively, if the loss is unexpected and does not correlate with the level of surgery, alternate explanations are considered. For example, isolated loss in a single upper extremity during thoracic spine surgery may indicate a positioning problem (resulting in circulatory changes).

Loss of signals in all four limbs may occur as a result of anesthesia effects. High levels of inhalational anesthetic agents can depress cortical function and subsequently depress the amplitude and prolong the latency of SSEP waveforms. These changes can mimic the effects of spinal cord compromise. Subcortical responses are, of course, much more resistant to anesthetic changes than cortical responses, and peripheral responses are highly resistant to the effects of anesthesia. Total intravenous anesthesia, with avoidance of inhalational agents altogether, has the least effect on SSEP but requires an experienced anesthesiologist and adds cost. It is critical for the surgical team to make known to the anesthesia team its intention to use SSEP monitoring prior to surgery.

Fig. 14.2 (a,b) International system for electrode placement.

SSEPs are also affected by physiologic variables such as temperature and blood pressure. A decrease in limb or body temperature can result in increased latency and decreased amplitude of SSEP responses. Hypotension with global cerebral hypoperfusion and/or spinal cord ischemia can also result in amplitude attenuation or latency prolongation.

Positioning can influence SSEP readings if there is prolonged pressure on vessels supplying the limb. Resulting vascular compromise will cause a reduction of SSEP signals.³⁹ Brau et al⁴³ demonstrated that retractor-induced iliac vessel compression in anterior lumbar surgery may similarly lead to SSEP alterations. These changes are reversible if vascular occlusion is removed. Failure of SSEP signals to recover may indicate ongoing vascular compromise.

14.3.2 Indications

Surgical procedures that place the spinal cord at risk, such as correction of deformity, resection of a spinal cord mass, thoracic discectomy, and scoliosis correction procedures, are logically good candidates for SSEP monitoring. In any of these cases, a baseline SSEP would be obtained at the beginning of the procedure and then repeated on a regular basis throughout the surgical procedure. If SSEP waveforms change in response to an action by the surgical team (e.g., placement of a graft or hardware; rod distraction), the surgical team could be advised of this change, allowing these steps to be “undone” in the hope that signals recover to baseline levels.

In the absence of changes in SSEP signals, one might hope that whatever surgical interventions had just transpired did not damage the spinal cord. Unfortunately, this is not always borne out.⁴⁴ There are numerous reports of cases in which SSEPs were unchanged between baseline and final measures, yet the patient woke up with paralysis.⁴⁵^,⁴⁶^,⁴⁷^,⁴⁸ Evidence tells us that the command signals for voluntary motor function are carried in corticospinal tract (CST) axons.⁴⁹^,⁵⁰^,⁵¹^,⁵² These axons not only are physically separate from those of the dorsal columns, but also have a much greater dependence on the anterior spinal artery for perfusion than dorsal column axons (which are perfused via the posterior spinal arteries).⁵³^,⁵⁴ Based on results of postmortem studies after traumatic spinal cord injury, the large myelinated axons of the CST are also more susceptible to physical trauma than are axons within the dorsal columns.⁵⁵ Therefore, it should come as no surprise to learn that SSEP monitoring is not always a good indicator of spinal cord damage and specifically motor function. CST (motor) function is best monitored using MEPs (see Motor Evoked Potentials (p.140) section).

There are also fundamental difficulties associated with SSEP testing within the OR setting. First, the recorded SSEP signal amplitude is very small (often less than 1 μV), resulting in a need for signal averaging. Such averaging introduces delays, measured in minutes. The SSEP is therefore not a “real-time” test. Second, the SSEP is prone to electrical interference from the many sources of electrical noise found in the typical OR environment, including the lights, the OR bed, radios, the anesthesia machine, the microscope, blood warmer, suction, electrocautery, navigation tools, and fluoroscope. This electrical noise makes interpretation of small signals very difficult. Third, SSEP recordings are difficult to obtain in many patients with preexisting sensory conduction abnormalities in peripheral nerves (e.g., diabetes).³²^,⁴¹^,⁵⁶^,⁵⁷^,⁵⁸ Such patients are common in the population of adult patients undergoing most spinal procedures.

An additional limitation of SSEPs is that it is a general test of sensory pathway status and not a specific test. Sensory signals enter the spinal cord via multiple spinal roots. This means that SSEPs are neither specific nor adequately sensitive to detect injury in a single nerve root or an isolated peripheral nerve, which are at risk with most lumbar spine procedures.⁵⁹ For these cases, more appropriate intraoperative monitoring techniques include EMG and MMG. Gundanna et al¹³ assessed the use of SSEPs for placement of 888 pedicle screws in 186 consecutive patients. No patients demonstrated changes in SSEP during surgery, although five patients suffered new post-op radicular changes. Malposition was confirmed on CT imaging and all screws were subsequently removed or revised. The authors concluded that the use of SSEP in evaluating lumbar pedicle screw placement is of limited value.

Finally, it is relatively common for SSEP traces to demonstrate significant worsening from baseline values in amplitude and/or latency, only to have the patient awaken with no obvious change (i.e., deterioration) in sensory status. Such false-positive findings have continued to limit reliance on SSEP monitoring in the OR setting.⁷^,⁵⁸^,⁶⁰^,⁶¹^,⁶²^,⁶³^,⁶⁴^,⁶⁵ All of these limitations of SSEP testing have led to repeated calls for the development of monitors that are specific to both central and peripheral motor conduction.⁶¹^,⁶⁶^,⁶⁷ That has led to the rise in the use of MEPs, EMG, and MMG in spinal procedures.

14.4 Motor Evoked Potentials

The integrity of descending motor pathways can be monitored by depolarizing the corticospinal system through transcranial (Tc) stimulation and then measuring the resulting MEP distal to the level of surgery. The main descending tract responsible for voluntary movement is the CST.⁴⁸^,⁶⁸^,⁶⁹ This tract lies in the lateral portion of the spinal cord and innervates lower motor neuron cells located in the ventral horn (anterior horn cells). The blood supply to both the CST and anterior horn cells comes primarily from the anterior spinal artery. When the cranium is stimulated, all signals reaching muscles distally must travel the length of the motor pathway in the spinal cord. Therefore, transcranially triggered MEPs (TcMEPs) are useful in assessing the integrity of the CST and, indirectly, the integrity of spinal cord oxygenation and blood supply (through the anterior spinal artery). Calancie et al⁴⁵ have pointed out that the large myelinated axons of the CST are very sensitive to mechanical trauma and the anterior horn cells are very sensitive to ischemic changes.⁷⁰ MEP monitoring is especially sensitive to ischemia involving anterior horn cells.⁷¹

14.4.1 Technique

To achieve Tc stimulation of the cortex, very high stimulus amplitudes are needed to overcome skull impedance. For most cases, a single corkscrew electrode is used, offering reliable attachment to the scalp. Initial attempts to utilize TcMEPs focused on single, high-intensity pulses (up to 1,000 V).⁷²^,⁷³^,⁷⁴^,⁷⁵^,⁷⁶^,⁷⁷^,⁷⁸^,⁷⁹^,⁸⁰ Later, it was realized that activating upper motor neurons using a short, very high-frequency pulse train is more likely to be successful than a single pulse.⁸¹^,⁸²^,⁸³ Stimulation is applied to cortical sites roughly comparable to those used for EEG recording electrodes. Stimulation between the anode (positive lead) and cathode (negative lead) at sites just anterior to C3 and C4 ( Fig. 14.2) is a common configuration for activating the hand area on one side; reversal of the stimulating polarity activates the hand area on the contralateral side. This same configuration is also adequate for stimulating the leg area, although stronger stimuli are needed.⁴⁵

MEP monitoring is accomplished by recording EMG tracings associated with muscle contractions. Clearly, this requires avoidance of neuromuscular blockade. Muscles whose innervation lies caudal to the area of surgery serve as target muscles. For example, the tibialis anterior would serve as a target muscle for correction of a thoracic scoliosis deformity. However, in most cases where we use MEPs, we attempt to monitor MEPs from at least one muscle receiving innervation above the surgical level. This muscle serves as a control, telling us that instrumentation is working correctly and that anesthesia is appropriate.

Pairs of electrodes are placed over the muscles of interest. Regardless of the spine level, we routinely use hand and foot muscles, abductor pollicis brevus–abductor digiti minimi (APB-ADM) in the hand and abductor hallucis (AH) in the foot. This is because the motor regions of the hand and foot are well represented in the cortex and thus easier to stimulate. Even though the muscles are small, the triggered responses have high amplitude and are very reliable. We use pairs of 0.5 inch, needle electrodes placed just under the skin over the muscle of interest. The interelectrode distance is typically 2 to 4 cm. Filtering in the range of 50 Hz to 2.5 kHz is used for all EMG recording channels and an initial screen sensitivity of 100 μV per division.

Most technicians who use TcMEPs rely on a very strong (i.e., supramaximal) stimulus to elicit a maximal motor response and then watch for reduction in response amplitude.⁸¹^,⁸⁴^,⁸⁵^,⁸⁶^,⁸⁷^,⁸⁸^,⁸⁹^,⁹⁰^,⁹¹ This technique uses a train of stimuli (typically 4–6 and less than 20) starting at 300 to 600 V. Voltage is increased up to 1,000 V if needed to obtain a strong response. Baseline readings are obtained following induction and positioning and then repeated periodically throughout the case, with emphasis on retesting following risky events such as spinal cord manipulation. Performing MEP tests is time-consuming and disruptive to the flow of surgery and, as a result, most surgeons do not want testing performed more than is necessary.

As with SSEPs, alarm criteria vary widely. Most reports have used a criterion of a 50 to 80% decrease in amplitude, but some reports used complete loss of response or morphological change.⁹³^,⁹⁴^,⁹⁵^,⁹⁶^,⁹⁷^,⁹⁸ With the exception of the complete loss criteria, each of the other criteria is associated with higher numbers of alerts and a lower percentage of true neurological changes. Recently, Kobayashi et al⁹⁹ attempted to resolve this conflict by examining 48 cases of true positive MEP signal changes. They concluded that a 70% decrease in signal amplitude was the ideal alarm point. This criterion resulted in 95% sensitivity and 91% specificity.

Calancie et al²⁴ have described an alternative “threshold-level method” of testing MEPs which avoids maximal stimuli. To achieve maximal motor responses to TcMEP, a stimulus of such magnitude is required that peripheral nerves become stimulated, increasing the risk of¹⁰⁰:

• A shift in patient position.

• Bite-induced injury to the tongue, lips, teeth, or jaw.

• Tissue damage caused by a surgical instrument.

For these reasons, Calancie et al⁴⁵ have recommended the use of a threshold-level approach to TcMEP. Beginning with a three- or four-pulse train at 100 V stimulus intensity, the applied voltage is increased in increments of 25 or 50 V until a target muscle responds to the stimulation. This threshold reflects the minimal voltage needed to cause a minimal MEP response. If an event during surgery causes a partial conduction block in the spinal cord, some of the axons that were mediating the initial threshold-level response may no longer be capable of conducting action potentials across the region of the block. In order to induce a response, it then becomes necessary to recruit an additional population of upper motor neurons whose axons are not blocked. This additional recruitment occurs with a higher stimulus level. This increase in threshold value serves as the primary outcome measure of this approach. Calancie et al⁴⁵ have shown that changes of 50 V in threshold during the operation are common and do not reflect underlying compromise of neural conduction. Changes of 100 V or more, in the absence of a decline in systemic blood pressure or an increased delivery of anesthetic, are associated with compromised neural conduction. Therefore, the surgical team should be warned in the event that the threshold of a given muscle increases by 100 V or more.¹⁰¹

14.4.2 Indications

Any surgical procedure in which the spinal cord and/or its blood supply are at risk of injury could benefit from MEP monitoring. Put another way, any surgical procedure in which the SSEP is being used is also a candidate for MEP monitoring. Both SSEP and MEP have value. Each is designed to provide feedback about conduction within separate regions of the spinal cord and brain. Each test is good at predicting changes in either postoperative sensory (via the SSEP) or motor (via the MEP) function. The MEP is more likely to remain viable in cases when SSEP recording is not possible (e.g., polyneuropathy) and MEP is likely more sensitive to ischemic changes. Nevertheless, the two forms of monitoring should be considered complementary rather than competing.

The use of TcMEP is not without its drawbacks, including a sensitivity and specificity that are less than perfect. Some surgeons use a wake-up test to confirm test accuracy if observed MEP changes do not reverse. Others simply abandon corrective procedures and rely on the postoperative examination to determine if further attempts should be made to go back and continue with more corrective procedures.

TcMEP also carries inherent risk of injury, including bite-related injuries. There also remains the remote risk of seizures and twitch-related bumping against a surgical instrument. These risks are minimized through the use of an appropriate bite block, through a rigorous adherence to the use of minimal stimulus energy, and by alerting the surgical team to testing so that hazardous instruments are absent from the field during stimulation. We routinely have the surgical team remove all hands from the surgical field while Tc stimulus is applied.

Finally, TcMEPs are nonspecific. They have no value in testing the function of individual nerve roots or peripheral nerves. When these structures are at risk (such as in most lumbar procedures and many cervical procedures), alternative monitoring techniques such as EMG and MMG become necessary.

14.5 Electromyography

Spontaneous EMG occurs through the mechanical activation of axons distal to their cell body or as a result of ongoing central or peripheral nerve pathology. EMG activity is routinely seen with nerve root retraction, is especially prominent when the nerve roots are inflamed due to chronic compression or irritation, and can aid (indirectly) in the anatomic identification of specific nerve roots. EMG activity is measure by placing electrodes in or directly over the muscle(s) of interest ( Fig. 14.3).

The use of spontaneous EMG to guide surgical procedures around delicate nerve roots was first reported for posterior fossa decompression surgery, in an attempt to preserve facial nerve function.¹⁰² This approach was later adopted for spinal nerves.¹⁰³^,¹⁰⁴^,¹⁰⁵^,¹⁰⁶^,¹⁰⁷ When EMG signals are linked to a surgeon-directed user interface, immediacy of feedback represents a distinct advantage over the delay typically associated with SSEP monitoring, establishing a real-time test modality.

The frequency and amplitude of spontaneous EMG discharge generally increases with greater degrees of nerve root manipulation. Despite this observation, investigators have been unable to demonstrate a quantitative relationship between the amount of spontaneous EMG and postoperative motor deficit for spinal nerves. This is in sharp contrast to facial muscle EMG, wherein excessive activity during the surgery almost always predicts significant facial nerve dysfunction postoperatively.¹⁰²