Somatosensory-Evoked Potential Monitoring

The use of SEPs in intraoperative monitoring of complex spine surgeries began in the early 1970s.⁸ Although SEP monitoring evaluates primarily the integrity of the posterior columns, it is often used to give an overall assessment of the spinal cord based on the assumption that many intraoperative mechanisms of injury would affect the spinal cord diffusely. An example of such an injury is spine distraction during scoliosis surgery. In addition, ischemic injury may initially result in a more diffuse dysfunction of the spinal cord that could be detected by SEPs (Fig. 14–1). SEP responses are thought to pass through large fiber somatosensory pathways of the dorsal column and possibly through the anterior spinothalamic tract, and this may be another reason why anterior spinal artery ischemia could be detected using this technique.

FIGURE 14–1 Significant amplitude change in cortical response owing to ischemic etiology. Stack on the left shows leg cortical response, and stack on the right shows popliteal fossa response. Both were obtained after left posterior tibial stimulation. Responses are represented with baseline responses shown at the top of the stack and end of monitoring shown at the bottom of the stack. Decrease in leg cortical amplitude can be appreciated at point depicted by arrow. Popliteal fossa responses are intact during this time. There is a return of response by the end of surgery seen at the bottom of the stack. This change was attributed to ischemic change to cord with retractor placed over left iliac artery. Responses returned when retractor was adjusted away from artery.

Generators of Somatosensory-Evoked Potential Responses

The cortical response for the lower extremity is called the P37 potential. The generator of this response arises from the primary somatosensory cortex of the leg, which is located in the mesial parietal cortex. The cortical response for the upper extremity, which is generated from the primary somatosensory cortex of the hand, is called the N20 potential (Fig. 14–2). Two important characteristics of these waveforms are (1) amplitude, which is recorded in microvolts and determined by either a baseline-to-peak or a peak-to-trough measure of the waveform, and (2) latency, which is recorded in milliseconds and is the time interval from the stimulus to the occurrence of the potential. An amplitude change from the initial baseline measure to a decrease of more than 50% is often considered a significant change in SEP amplitudes.⁹

FIGURE 14–2 Typical morphology of cortical generators of median nerve and posterior tibial nerve somatosensory-evoked potential (SEP) waveforms are shown. Display time is different between two modalities. Median nerve SEP is shown in 5 msec per division display, and posterior tibial nerve SEP is shown in 10 msec per division display.

Significant latency changes in SEP monitoring consist of a 10% prolongation beyond the baseline latency value (Table 14–1).¹⁰ Although these deviations from the baseline measures are thought to be significant, they should be interpreted with caution, taking into account various factors, including the evolution of the changes (e.g., a trend toward worsening is an ominous sign) and other intraoperative factors such as length of the surgery, type of anesthetic agent, and temperature effects. Also, significant latency and amplitude changes can occur in isolation. It is quite common to see a significant amplitude change without any associated latency changes. The most significant change is a complete loss of the cortical potential.

TABLE 14–1 Significant Changes in Different Monitoring Modalities

Another measurement made in posterior tibial or peroneal nerve SEP monitoring is the popliteal fossa potential. This is a nerve action potential that is recorded as the impulses pass within the popliteal fossa in the peripheral nervous system. The reason for this measurement is to ensure that an adequate stimulus has been applied. If there is an absence of the popliteal fossa response in addition to an absence of the leg cortical (P37) response, it suggests that the changes seen are not a result of a lesion at the level of the spinal cord. In this case, the change may be technical (e.g., the stimulating needles may have dislodged) or may be seen when the leg is ischemic (e.g., with femoral artery catheterization during thoracoabdominal aneurysm surgery or with direct compression of the peripheral nerve) (Fig. 14–3).

FIGURE 14–3 Significant change in left N20 cortical amplitude owing to arm positioning and nonsignificant latency prolongation of all cortical responses owing to anesthetic effect. Top row of stacks shows popliteal fossa (PF) and leg cortical (P37) responses from posterior tibial stimulation first from left-side stimulation and then from right-side stimulation. Bottom row of stacks shows Erb point (EP) and arm cortical (N20) response from median nerve stimulation with left side shown first followed by right side. There is a decrease in left N20 amplitude. This is highlighted with rectangular box in figure. At this point, there is also loss of left EP response. This change was attributed to malpositioning of left arm. When left arm was repositioned, response returned to baseline. Also noticeable in all leg and arm cortical responses from left and right sides is mild prolongation of latencies in the stacks, but these latencies all return to baseline by end of surgery. These changes are likely from anesthetic effect because they are bilateral, affecting arm and leg responses in a spine operation, which was performed at L3-S1 level.

The other posterior tibial stimulation SEP response that can be monitored (besides the P37 response) is the P31/N34 complex, often termed the subcortical response because the generator for this response is at the level of medulla and midbrain. The benefit in recording these responses is that they are relatively more resistant to the effects of anesthesia compared with the cortical P37 response (see Fig. 14–3). The same is true for the subcortical potentials from median nerve stimulation (P14/N18) potential. In pediatric cases, the subcortical potentials may also be better formed and more easily monitored than cortical responses. Some of this effect may be the result of the variation of myelination in the younger age groups and the more significant effects of anesthetics on these patients. These differences from the adult morphology can persist into the early teenage years. Other factors affecting the responses include core body temperature changes. The core body temperature commonly decreases more than 1° C. The cooling affects the limbs more than the core body temperature, which can result in slowing of conduction.

Motor-Evoked Potential Monitoring

As mentioned earlier, various methods have been used to monitor spinal motor pathways during surgery. Most of these methods involve recording of electromyography (EMG) readings from appropriate muscles in response to stimulation of a motor pathway rostral to the operative site. The difference between the various methods is the nature of the stimulation. There are three basic categories of stimulation: rostral spinal stimulation, transcranial magnetic stimulation, and transcranial electrical stimulation (TCES). Magnetic stimulation is effective in nonanesthetized patients for evaluation of motor pathways, but the suppression of cortical responsiveness under anesthesia (mainly inhalational anesthetics) renders this method less effective for surgical use. In addition, the equipment used for magnetic stimulation is expensive, is bulky, and has a tendency to overheat.

Noninvasive stimulation of the brain using TCES was first reported in 1980.¹¹ Single-pulse TCES was subsequently used in monitoring motor pathways.^12–17 Because of effects of general anesthesia, single-pulse stimulation was found to be less effective in reliably recording MEPs.^18–22 With the introduction of the multipulse technique for motor pathway monitoring, reliable and robust MEP recording can now be obtained in most patients using some specific general anesthesia protocols.^23–27 The use of the multipulse technique requires that neuromuscular blockade not be used during this part of the monitoring. Occasionally, the use of partial neuromuscular blockade may still allow for TCES monitoring.²⁸ This method reportedly achieves more reliable stimulation of the motor cortex intraoperatively and is more resilient to the effects of general anesthesia.

The method of MEP monitoring has been revolutionized by the use of multipulse TCES. Previous methods for MEP recording used a variety of stimulation and recording techniques. Spinally elicited neurogenic responses were used in the past and were putatively stated to be a result of activation of the motor pathways in the spinal cord. More recent evidence has suggested that these spinally elicited neurogenic responses are generated through activation of the sensory pathways and retrograde activation of alpha motor neurons. In a collision experiment using stimulation of the spinal cord followed by stimulation of the posterior tibial nerve at various interstimulus intervals, the neurogenic responses were abolished, suggesting that the potentials were colliding in the spinal cord.⁶

At the beginning of TCES-MEP monitoring, threshold voltages for each side of the body and MEP amplitudes are calculated.²³ The motor cortex on the side of the brain receiving the anodal stimulus is typically the first region to activate at the lowest stimulus threshold. The initial current used is typically 100 V, with a train of stimuli delivered to the cortex. After the stimulation, an MEP response is monitored in the muscles contralateral to the side receiving the anodal stimulus. If no response is seen, the voltage is increased typically by 50-V increments, and the process is repeated until an MEP response is seen in all the muscles contralateral to the anodal stimulus. This voltage is called the threshold voltage for that side.

The highest amplitude of the myogenic response below the level of surgery is also noted. Typically, amplitude measures for myogenic responses are best recorded as the area-under-the-curve measurements or simply by just documenting either presence or absence of the myogenic response. This procedure is repeated after reversing the anodal-cathodal configuration using a switch box. The voltage used for TCES-MEP recordings typically does not exceed 500 V. The anticipated latency of the EMG responses ranges from 20 to 40 msec or more, depending on the patient’s height, owing to the conduction time in the descending motor pathways. Latency values have not always been found to be reliable indicators of significant change in TCES monitoring in clinical practice. Another advantage of the multipulse technique is that it elicits a train of pulses that is of sufficient amplitude that it does not require averaging.

Two different methods of recording MEPs can be used. In myogenic MEPs, responses can be recorded directly from the muscle (either a surface electrode or needle electrodes placed within the muscle). In spinal cord MEPs, responses may be recorded directly from the spinal cord with use of an epidural catheter electrode that records a direct “D” wave and a volley of indirect “I” waves. Using single-pulse TCES, recording “D” and “I” waves is frequently required, meaning a “D” wave could be recorded when a myogenic MEP is not yet seen. This is because a series of “D” and “I” waves is required for alpha motor neurons to generate a myogenic response. The spinal recorded responses can also be recorded with full muscle relaxation, whereas myogenic responses require either no or very little muscle relaxation, even with the multipulse technique.

Determining significant changes during the course of surgery typically is most reliable if there is an absolute loss of myogenic responses to stimulation (Fig. 14–4). Some authors have also suggested that reduction of MEP amplitude to 25% of baseline amplitude values is predictive of motor pathway impairments.²⁹ Significant changes can also be determined by the change of voltage required to obtain MEPs of greater than 50 V beyond baseline thresholds used in obtaining MEPs at the beginning of monitoring (see Table 14–1).³⁰

FIGURE 14–4 Significant change in transcranial electrical motor-evoked potential (MEP) response during spinal instrumentation. Figure shows transcranial MEP responses at four points during surgery. Each set of responses shows response from left (L) and right (R) sides of body. First two rows are responses from muscles of upper extremity (biceps and first dorsal interosseous muscle), and bottom three rows are from muscles of lower extremity (anterior tibialis, medial gastrocnemius, and adductor hallucis longus). A, This set is MEPs at baseline or beginning of monitoring obtained at intensity of 300-V train-of-four with interstimulus interval of 2 msec. B, This set depicts decrease in amplitude of leg responses obtained at 500 V at one point during instrumentation. C, These responses are obtained soon after initial change in B with total loss of amplitude in leg MEPs (in box). MEPs from arms are still intact. This suggests a lesion in cord below level of C8-T1. D, With some release of degree of distraction of spine, responses are shown to return in lower extremity.