Intraoperative Neurophysiologic Monitoring

Published on 13/06/2015 by admin

Filed under Surgery

Last modified 22/04/2025

Print this page

This article have been viewed 2671 times

Chapter 64 Intraoperative Neurophysiologic Monitoring

Adrien A. Eshraghi, Sarah S. Connell, Ray C. Chang, Fred F. Telischi

Intraoperative neurophysiologic monitoring is a valuable tool for improved patient outcomes because it permits the surgeon to evaluate functional changes in structures at risk. Facial nerve monitoring has reached a level of consistency that makes it a state-of-the-art adjunct to lateral and posterior skull base approaches. Techniques for monitoring auditory function continue to evolve.

MONITORING FACIAL NERVE

History

Krause ¹ first described facial nerve monitoring in 1912 using a faradic stimulation during cochlear nerve section for tinnitus. Twitching of the ipsilateral facial muscles during stimulation helped him preserve the facial nerve, and the patient had transient facial weakness postoperatively. In the 1960s, dedicated facial nerve monitoring systems were developed. The Hilger stimulator² was used principally in the assessment of facial paralysis, but was also used during surgery. Further developments in facial nerve monitoring occurred in the 1970s and 1980s. Delgado and colleagues³ described the use of electromyography (EMG) monitoring in cerebellopontine angle (CPA) surgery. Moller and Jannetta⁴ combined the specificity of EMG recording with the advantage of acoustic feedback to the surgeon.

Indications

The underlying principle for intraoperative monitoring is that some types of injury can be reversed. Although introduction of facial nerve monitoring has improved functional outcomes further, especially in medium and large CPA tumors, facial nerve monitoring should never replace anatomic knowledge, technical skill, and clinical judgment. Facial nerve monitoring can be useful to identify the facial nerve when it is not clearly visible in the surgical field. Other benefits include localizing distal nerve fragments in trauma cases, and identifying sites of nerve compression, as long as wallerian nerve degeneration has not completely occurred. Otologic procedures in which the nerve is at risk include cochlear implantation, revision tympanomastoidectomy, and repair of external auditory canal bony stenosis. Monitoring may augment safety in cases where the anatomy is altered by infection, trauma, or congenital malformation. It may also be beneficial in training centers where some portions of operations are performed by less experienced surgeons.

Physiology

Stimulation of a motor nerve by electric, mechanical, or thermal means results in depolarization of the nerve. A compound muscle action potential is the composite electric activity within the target muscle resulting from synchronous activation of a group of motor neurons within a nerve bundle. EMG monitoring essentially measures this electric activity in the portion of the muscle near the recording electrodes and converts the electric activity to sound via a loudspeaker with or without visual oscilloscope display.

Synchronous activity initiated by electric stimulation produces a biphasic and well-defined waveform. Asynchronous activity, as is generally produced by mechanical stimulation, produces a polyphasic pattern. When these response patterns are converted to sound, electric stimulation results in well-defined pulsed sounds, whereas mechanical stimulation yields a rough, almost scratchy, burst of acoustic energy.

Electromyography

Subdermal needle electrodes are most commonly used for facial nerve monitoring. They have the advantages of ease of use, low impedance, and stability (less likely to be displaced than surface discs or cups). Monitoring more than one muscle provides additional sensitivity and redundancy.⁵ For two-channel bipolar recording, a typical sensing montage includes a pair of electrodes in the orbicularis oculi approximately 1 cm apart and another pair in the orbicularis oris (Fig. 64-1). The ground electrode is placed in the forehead, and the anode for the monopolar nerve stimulator is inserted at the ipsilateral shoulder. The operating room is a noisy place with abundant electric interference. The connections are checked by tapping the electrodes and observing an audible and oscilloscopic response. This mechanical compound muscle action potential results in a characteristic sound from the monitor. A second check, when available, involves testing the impedance of the inserted electrodes.

FIGURE 64-1 Electrode placement for facial nerve intraoperative monitoring. SCM, sternocleidomastoid.

Individual impedance values should be less than 5 kilohm (kΩ) to avoid electromagnetic interference from other devices in the operating room. Current instrumentation uses differential amplification techniques that improve signal-to-noise ratio optimally when impedance imbalance is less than 1 kΩ between the electrode pairs. Ideally, the imbalance-to-impedance ratio should be less than 10%. Personnel who perform intraoperative monitoring must be present in the operating room to ensure all equipment is functioning properly.

Many facial nerve monitoring systems are available commercially, including the Nerve Integrity Monitor (Xomed, Inc., Jacksonville, FL) (Fig. 64-2), Neurosign 100 (Smith & Nephew Richards, Inc., Memphis, TN), Brackmann II (WR Medical Electronics Co., Stillwater, MN), and NEI (Grass Instrument Co., Quincy, MA). All of these devices use EMG. The Silverstein Facial Nerve Monitor (WR Medical Electronics Co.,) is an example of a motion detector device. Some systems, such as the Silverstein Monitor, include the ability to electrify instruments to aid with monitoring.

FIGURE 64-2 Facial nerve monitor screen (NIM Response 2.0 Nerve Integrity Monitor; Medtronic ENT, Jacksonville, FL). Two-channel modes are typical for mastoid surgery. A, Burst activity, occurring with mechanical stimulation or brief electric stimulation. B, Pulsatile stimulus. C, Train activity.

Stimulation

Pulsed stimulation may be safer and more efficacious.⁶ The parameters of safe nerve stimulation are 100 to 250 μs pulses with a range of 0.05 to 0.5 mA.⁵ The upper limit of safe nerve stimulation varies along the course of the nerve, but animal studies have shown myelin and axonal injury using 2 mA stimulus for 3 seconds.⁷ A 0.5 mA stimulus applied briefly 50 times was found to manifest no functional or histologic evidence of injury in a mouse model.⁸ Most normal facial nerves should be stimulated with direct contact of the probe using a 100 μs pulse of 0.05 mA. Settings of 0.05 to 0.1 mA are recommended when working close to the nerve. Farther from the nerve, currents of 0.2 to 0.3 mA may be used. Higher settings may be required when the nerve is covered by bone, connective tissue, or granulation tissue. Cerebrospinal fluid or blood may shunt current from a stimulator probe. In these cases, stimulation with a constant voltage may be used.

In an attempt to determine the threshold necessary to detect a surgical dehiscence of the facial nerve electrically, Choung and associates ⁹ prospectively estimated the minimal threshold of electric current needed to change the EMG of facial muscles using the Nerve Integrity Monitor (NIM)-2 in 100 patients. They found that 43% of ears with surgical dehiscence responded to electric stimulation of 0.7 mA or less. The mean threshold of minimal electric stimulation was 0.29 mA for tympanic segments and 0.41 mA for mastoid segments.

It is crucial that patients be adequately grounded to the monopolar electrocautery unit to permit safe current return. Otherwise, the electrocautery current may find a route through the nerve monitoring electrodes, causing severe burns. All patient connections must be optically and/or electrically isolated to prevent patient injury. This isolation separates the patient from power line voltages and currents. Some monitoring devices achieve this using battery power.

Burst Activity

During the course of surgery, many burst potentials may be observed and are usually not associated with significant trauma to the nerve. Bursts are brief discharges in which the stimulus and response are simultaneous (see Fig. 64-2). Burst activity occurring with gentle manipulation suggests a healthy nerve. Very similar responses may be obtained, however, when drilling close to or after complete transection of the nerve. Lack of burst activity during dissection may be associated with minor manipulation of a healthy nerve, significant manipulation of an already injured nerve, or a problem with the monitoring connections and instruments. Electrically stimulating the nerve at this point verifies the integrity of the nerve.

Trains

Episodes of repetitive EMG activity may occur several seconds to minutes after the stimulus, making it difficult to identify the initiating factor or to modify dissection technique. As seen in Figure 64-2, trains are caused by prolonged depolarization of the nerve beyond its threshold for developing an action potential. Subsequent repetitive firing continues until the nerve repolarizes, or can no longer sustain the repetitive activation.⁵ The most common initiating factor is traction on the nerve. Trains may indicate significant trauma has occurred, although this is not always the case. Changes in temperature around the nerve may also precipitate train activity. When caused by cool irrigating fluid, the spontaneous activity usually subsides with warming. When a train pattern develops after laser application or cautery, however, thermal damage should be suspected. Elevated stimulation thresholds after repetitive nerve activity suggest significant injury. Higher frequency train activity (>50 Hz) with an amplitude greater than 250 μV is associated with a more ominous outcome; in a series of 51 patients, Nakao and coworkers¹⁰ found that 86% of patients with train activity having an amplitude greater than 250 μV had severe facial nerve dysfunction. More than 10 seconds of cumulative train time is associated with postoperative facial paresis.¹¹

Anesthetic Issues

It is common practice that when facial nerve monitoring is employed, neuromuscular blocking agents are avoided after induction of anesthesia. In 2003, Kizilay and associates ¹² examined the effects of different levels of neuromuscular blockade (NMB) on electric stimulation thresholds of the facial nerve during otologic surgery. Minimal facial nerve stimulation causing EMG responses in the facial musculature was measured during recovery from the effects of muscular relaxants and with 25%, 50%, 75%, and 100% levels of NMB. All of the patients had detectable EMG responses of the facial musculature at the 50% and 75% levels of NMB in response to the electric stimulation (mean 0.1 mA) of the facial nerve. No responses were measured in 31% of the patients when the level of peripheral NMB was 100%. The investigators concluded that a regulated 50% level of peripheral NMB provides reliable intraoperative EMG monitoring of the facial musculature in response to electric stimulation and adequate anesthesia, with full immobilization of the patient. Chronic injury resulting from compression from a tumor may make the facial nerve more sensitive to the effects of NMB, however.¹³

Spontaneous repetitive responses may be observed when the level of anesthesia becomes inadequate, and the facial muscles contract. This activity is commonly the first indication that the patient is awakening and the prelude to larger movements, such as coughing or bucking. Migration of inadequately secured electrodes may create auditory artifacts that mimic the spontaneous activity of light anesthesia.

Practical Application in Surgery

Locating the Facial Nerve

Monopolar stimulation is especially useful for mapping the nerve throughout its course.⁴ Monitoring becomes more important in cases of large tumors.⁵ A stronger stimulus may need to be used to confirm that the nerve is not in close proximity to an area of tumor to be removed. False-positive responses may be obtained from stimulation of the superior vestibular and cochlear nerves owing to electric current dispersion, especially within the internal auditory canal where the nerves lie in close proximity to one another. Stimulation of the trigeminal nerve may elicit motor activity that can be interpreted as facial nerve stimulation. False-negative responses are usually related to technical errors such as failure to connect the stimulus probe, anesthetic-induced muscle paralysis, or impedance imbalances.

The electric current dispersion is most effectively overcome by using bipolar stimulation or a flush-tip monopolar stimulator on the lowest possible setting. A useful method to check the monitoring system is to stimulate the nerve where it is known to be available and intact. If no response can be obtained at that location, the entire setup should be inspected from the electrode placements in the facial muscles to the recording instrument. When a reliable response is elicited from the nerve, stimulation may be resumed with confidence.

Avoiding Trauma

Continuous EMG monitoring provides real-time feedback, enabling the surgeon to alter technique immediately when a burst response occurs. When performing lateral skull base surgery, it is useful to maintain monitoring during wound obliteration and closure to avoid trauma to the nerve.

Postoperative Prognosis and Prognosis of Acute Injury

The threshold for stimulation has been shown to correlate with postoperative facial nerve function. When using constant current stimulation with a 50 μs pulse, if the threshold for facial nerve stimulation at the brainstem was 0.1 mA or lower, 90% of patients exhibited House-Brackmann grade I or II function at 1 year after CPA surgery.¹⁴ Thresholds of 0.1 to 0.2 mA were associated with grade I or II function in 77% of cases.¹⁵ The amount of energy delivered to the nerve using a 50 μs pulse of 0.1 mA is equal in magnitude to a 100 μs pulse of 0.05 mA as delivered by many commercial facial nerve monitors. When testing the nerve at the root entry zone of the brainstem to determine threshold at the conclusion of the procedure, it is important to contact the nerve long enough to confirm the precisely timed pulses of electric stimulation. A single noise could be a mechanically evoked burst, which may give the surgeon a falsely optimistic estimate of nerve function.

In 2005, Grayeli and colleagues ¹⁶ looked at the short-term facial prognostic value of a four-channel facial EMG device in vestibular schwannoma surgery. In 89 patients, EMG detection was performed in frontal, orbicularis oculi, orbicularis oris, and platysma muscles. Postoperative facial function at 6 months was assessed as House-Brackmann grade I or II in 80%, as grade III or IV in 16%, and as grade V or VI in 4% (n = 80). A proximal threshold between 0.01 and 0.04 mA had a positive predictive value of 94% for good facial function (grade I or II). The proximal threshold was lower in patients with improving or stable facial function compared with patients with a delayed deterioration between days 8 and 30. The maximal EMG response was detected in the frontal muscle or the platysma in 27% of cases and in orbicularis oris and oculi in 73% of cases.

Also in 2005, Neff and coworkers ¹⁷ sought to evaluate prospectively whether the intraoperative stimulus threshold and response amplitude measurements from facial EMG can predict facial nerve function at 1 year after vestibular schwannoma resection. In 74 consecutive patients, the minimal stimulus intensity and EMG response amplitude were recorded during stimulation applied to the proximal facial nerve after vestibular schwannoma removal. Of the 74 patients, 66 of 74 (89%) had House-Brackmann grade I or II facial nerve function, and 8 of 74 (11%) had House-Brackmann grade III-VI function at 1 year after surgery. With intraoperative minimal stimulus intensity of 0.05 mA or less and response amplitude of 240 μV or greater, the authors were able to predict a House-Brackmann grade I or II outcome in 56 of 66 (85%) patients at 1 year after surgery. With these same electrophysiologic parameters, only 1 of 8 (12%) patients with House-Brackmann grade III-VI also met this standard and gave a false-positive result. Logistic regression analysis of the data showed that a stimulus threshold of 0.05 mA or less and a response amplitude of 240 μV or greater predicted a House-Brackmann grade I or II outcome with a 98% probability. Stimulus threshold or response amplitude alone had a much lower probability of the same result, however.

A similar study by Isaacson and colleagues ¹⁸ looked at two independent intraoperative monitoring parameters in predicting long-term facial nerve function in 60 patients undergoing resection of vestibular schwannomas. They found that 5 of 60 (8.3%) patients showed significant long-term weakness (i.e., House-Brackmann grade III or worse). Intraoperative monitoring parameters (proximal stimulation threshold, proximal-to-distal response amplitude ratio) were accurate in predicting increased risk of long-term facial nerve dysfunction when used in a logistic regression model. Table 64-1 summarizes a selection of studies using facial nerve monitoring to evaluate postoperative prognosis in acoustic neuroma surgery³⁸^–⁴⁷.

TABLE 64-1 Studies Using Facial Nerve Monitoring to Predict Facial Nerve Function after Acoustic Neuroma Surgery

Study	N	Result
Lin et al, 2006	38	With 0.3 mA stimulus, proximal-distal amplitude >50% had PPV 93% for HB grade II or better immediately postoperative
Grayeli et al, 2005¹⁶	89	Stimulation threshold of 0.01-0.04 mA had PPV 94% for HB grade II or better 180 days postoperative using four-channel EMG
Neff et al, 2005¹⁷	74	For <0.05 mA stimulus and response amplitude of >0.240 V, authors could predict 85% of patients with HB grade II or better 1 yr postoperative
Isaacson et al, 2003	229	Using proximal-distal amplitude and stimulation threshold, authors developed regression function with sensitivity 89%, specificity 83%, PPV 94% in predicting HB grade III or worse in immediate postoperative period
Fenton et al, 2002	67	Using stimulation current and tumor size, authors developed regression function correctly describing 93% of patients at 2 yr follow-up
Goldbrunner et al, 2000	137	With proximal-distal amplitude >80%, 98.4% had HB grade I at 6 mo
Nissen et al, 1997	116	Median threshold for HB grade II or better group was 0.100 V versus 0.725 V for HB grade III or worse group
Zeitouni et al, 1997	109	HB grade III or worse long-term was related to higher minimum stimulation current
Selesnick et al, 1996¹⁵	49	At 1 yr after surgery, 90% of patients stimulating at 0.1 mA had HB grade II or better
Taha et al, 1995	20	100% HB grade III or better initial and grade I long-term (≤28 mo) function occurred if proximal-distal amplitude >67%
Lalwani et al, 1994	129	Of patients with stimulus of <0.2 V, 98% had HB grade II or better 1 year postoperative versus 50% if stimulus was >0.2 V
Prasad et al, 1993	77	When threshold after tumor removal was ≤0.2 V, 93% patients had early, and 85% had late postoperative HB grade I or II
Niparko et al, 1989	29	83% HB grade I 1 wk postoperative if proximal-distal amplitude = 1 with 88% of these patients HB grade I 1 yr after surgery

HB, House-Brackmann; PPV, positive predictive value.

Artifactual Responses

Artifacts (i.e., monitor activity that is not due to facial nerve stimulation) are common and may cause confusion. Electrocautery obliterates facial nerve responses by saturating the monitor with electric noise. Consequently, monitors have a muting function while electrocautery is in use. During muting, monitoring is disengaged. Ultrasonic aspirators also cause a large electric artifact, and may trigger a muting circuit.¹⁴ When accurate monitoring cannot be performed, many surgeons engage a member of the team to observe visually or palpate the patient’s face.

Troubleshooting the Facial Nerve Monitor

Although the facial nerve monitor can give the surgeon confidence to work safely and expeditiously by verifying anatomy, when the monitor does not respond as expected, the surgeon should proceed cautiously and initiate a sequence of steps systematically to ensure that the equipment is functioning and providing adequate monitoring sensitivity. Manufacturer manuals serve as a resource for troubleshooting specific to the equipment.

Direct communication with the anesthesiologist to confirm complete neuromuscular reversal is warranted. Checking along the electric circuit should be done, including electrode placement in the skin, electrode connections to the monitor, and probe connections. The volume of the monitor might have been turned down. A typical event threshold setting is 100 μV. Event threshold settings set too high may lead to a lack of a response. Conversely, false-positive results can be reduced by increasing the event threshold. The stimulus intensity should be confirmed and can be incrementally increased by 0.05 to 0.1 mA. In the middle ear, high sensitivity can be achieved with stimulus up to 0.7 mA.⁹ If a continuous current is used, stimulus should be changed to pulsed mode. Selesnick¹⁹ calculated the optimal stimulus duration of 50 μs, although many commercial monitors use a default pulse of 100 μs.

Fixed voltage stimulation is an option when the surgical field is not dry to compensate for current leak from the stimulator. The stimulator probe can be replaced with a stand-alone nerve stimulator. When the face is not covered, visual verification of stimulation is possible and can help in troubleshooting. Alternatively, the face may be manually felt to move under surgical drapes during stimulation. Final options include rebooting the monitor or changing the monitor counsel. Ultimately, surgical judgment should prevail; it behooves the surgeon not to perform irreversible steps, unless these potential sources of error and corrective measures are considered.

Antidromic Facial Nerve Monitoring

When a nerve is stimulated, in addition to conduction distally toward the neuromuscular junction (orthodromic), an action potential is conducted proximally (antidromically). The orthodromic wave produces the M wave on muscle contraction. The antidromic electric activity can be measured and used to assess neural function as the signal is reflected back toward the muscle resulting in a more subtle F wave. Wedekind and Klug ²⁰ prospectively evaluated F wave monitoring comparing it with intraoperative EMG, and found it useful monitoring facial nerve injury. They found that transient loss of F waves portends imminent severe facial dysfunction, and reported a 100% positive predictive value for unfavorable facial outcome with a permanent loss of F waves. A major advantage of antidromic monitoring is that it may be performed under NMB. Arriaga and associates²¹ described another use for antidromic potentials to locate the geniculate ganglion during middle fossa craniotomy.

MONITORING HEARING

Hearing depends on the integrity of the peripheral and central auditory structures and their vascular supplies. The goal of intraoperative monitoring of auditory function is to preserve hearing, not just anatomy.

Pathophysiology

Operating in the internal auditory canal and CPA poses two potential mechanisms for hearing loss: interruption of blood flow to the cochlea and the cochlear nerve, and injury to neural structures of the auditory pathway.²² In some cases, preoperative audiometric tests suggest a site of the lesion for the hearing loss. An inappropriately large reduction in speech discrimination score with good pure tone average suggests tumor effect on CN VIII or brainstem, rather than on the cochlea. Intact otoacoustic emissions (OAE) with poor auditory brainstem response (ABR) suggest the same.

Retraction of the cerebellum during the early stages of the operation may stretch and damage CN VIII, which is particularly vulnerable in the CPA. Surgeons need an objective measure of hearing preservation because visual confirmation of intact anatomic structures alone is inadequate to ensure postoperative auditory function.

Preoperative Testing

A standard audiogram displays auditory thresholds at frequencies important to language (500 to 8000 Hz). Various word list tests may be administered to corroborate pure tone threshold averages and to quantify speech discrimination ability. Acoustic reflex tests examine the integrity of the CN VII and VIII reflex arc. Cochlear potentials, ABR, and OAE are useful to evaluate the integrity of the auditory pathways further and to establish a baseline for intraoperative monitoring. It is important to know preoperatively whether the physiologic measures to be monitored are abnormal.

Indications

Selecting appropriate cases for monitoring involves analysis of preoperative hearing, disease prognosis, and surgical approach. The role of monitoring hearing is to decrease the risk to the auditory nerve in CPA operations. Hearing can be monitored directly from the auditory nerve, and auditory evoked potentials (electrocochleography [ECoG] potentials) can be used to check the endolymphatic system in endolymphatic sac decompression operations. Operations in which monitoring of auditory function has been reported include microvascular decompression of cranial nerves, vestibular nerve section,²¹ and removal of vestibular schwannomas and other CPA masses. The most common application for monitoring of auditory function is the removal of vestibular schwannomas.

Hearing preservation rates for surgical extirpation of vestibular schwannomas vary inversely with tumor size. There seems to be little reason to monitor hearing during planned total removal of a tumor 4 cm or larger. Conversely, auditory monitoring is ideal when removing smaller tumors or sectioning the vestibular nerve in the face of serviceable hearing. Many other factors may suggest a better prognosis for hearing preservation, including lack of tumor extension into the lateral internal auditory canal and erosion of bony walls, low behavioral thresholds (i.e., good hearing), normal ABR, and reduced caloric response on electronystagmography (i.e., superior vestibular tumor).

The generally accepted limits of serviceable hearing have been 50 dB HL pure tone average threshold and 50% speech discrimination score. Tumors in only hearing ears and bilateral tumors warrant special consideration. Worse hearing may indicate a more severe insult of auditory structures by disease and poor prognosis for hearing conservation. Exceptions exist, however, for hearing improvement when tumor removal alleviates a CN VIII conduction block.

In 2006, Samii and coworkers ²³ published their outcomes in a retrospective review of 200 consecutive vestibular schwannoma resections. They found that anatomic preservation of the facial nerve was possible in 98.5% of patients. By the last follow-up examination, excellent or good facial nerve function had been achieved in 81% of the cases. In patients with preserved hearing, the rate of anatomic preservation of the cochlear nerve was 84%. The overall rate of functional hearing preservation was 51%.

General Considerations

Many lateral skull base approaches involve temporal bone dissection. Fluid and bone dust may accumulate in the middle ear and external auditory canal, resulting in a conductive hearing loss. ABR, ECoG, and direct CN VIII potentials are delayed in the presence of conductive hearing loss. OAE generally cannot be measured in the presence of significant conductive hearing loss. Whenever possible, the middle ear and external canal should be isolated from manipulation when monitoring auditory function.

None of the methods for measuring auditory function during surgery should be affected by the use of routine anesthetics, including NMB. Patient factors, such as core temperature, muscle activity, blood pressure, and oxygenation, may alter electrophysiologic responses, however. Small decreases in core body temperature to 34.5° C result in delays of ABR waves.²³ Similarly, increased muscle activity (e.g., fasciculations in a nonparalyzed patient) or hypotension and hypoxia may interfere with accurate measurements.

Auditory Evoked Brainstem Response

ABR is considered a far-field response because it represents activity measured between scalp electrodes that are placed at relatively large distances from CN VIII and brainstem generator sites. Brainstem auditory evoked potentials consist of five to seven peaks representing electric activity of auditory nerve, nuclei, and fiber tracts of the ascending auditory pathways. Peak I refers to the distal portion of the auditory nerve. Peak II represents the central portion of the auditory nerve. Peak III represents the cochlear nucleus. Peak V represents the termination of the lateral lemniscus in the inferior colliculus. Waves III, IV, and V are created by multiple generators. Intraoperative ABR monitoring setup is similar to the setup for routine office measurement (Fig. 64-3). Subdermal needle electrodes are used for recording brainstem auditory evoked potentials. Insert ear plugs are suitable for delivering sound for the recording of auditory evoked potentials. Recorded potentials should be compared with baseline recording before the beginning of the operation. To reduce the time to obtain interpretable evoked potentials, the following actions are recommended:

1. Reduce electric interference from reaching recording electrodes

2. Filter recoded potentials to attenuate background noise

3. Optimize stimulus repetition rate and strength

4. Optimize electrode placement of recording electrodes to decrease electrode impedance

5. Use methods for quality controls that do not require record replication

FIGURE 64-3 Setup montage for auditory brainstem response, including ear canal inserts, active earlobe electrode, reference electrode on low forehead, and ground electrode on high forehead.

In pathologic ears, ABR morphology is likely to deteriorate and become even more variable. Amplitude is likely to decrease, especially in the presence of peripheral hearing loss, and latencies may increase based on the location and size of the lesion. Consequently, during monitoring of a previously impaired ear, changes from a well-established baseline are generally of more use than comparison with a norm. Because wave I is less likely to be present in an impaired ear, monitoring decisions may often depend on wave V. Enhancement of the wave I response can be achieved through ECoG recordings.

Selection of testing parameters for intraoperative ABR must be made with the goal of optimizing the recording of the desired response, while minimizing the interference inherent in the electrically hostile operating room environment. Click stimuli should be present at sufficiently high rates and intensities. Wherever possible, background noise must be eliminated at the source. Finally, state-of-the-art signal processing techniques should be employed to enhance the signal-to-noise ratio and minimize the number of responses averaged to obtain an adequate waveform.

Interpretation of Results

During posterior fossa surgery, damage is most likely to occur at the cochlea, CN VIII, or its root entry zone at the brainstem. An increase in peak V latency with or without a decrease in peak V amplitude and no change in peak III amplitude indicates the brainstem has been manipulated. Changes in interpeak latency I-III indicate stretching or compression of the auditory nerve. Disruption of the cochlear blood supply produces a quick and complete loss of ABR. Rapid response by the surgeon to restore the blood supply, if possible, is vital to prevent permanent damage. Other types of insults, such as stretching, compression, manipulation, or heat from the drill, can affect CN VIII and its root entry zone. ABR responses to these insults are neither as quick nor as easily interpretable as the responses produced by the loss of blood flow.

Much depends on obtaining an optimal preoperative baseline response before entering the operating room and again after induction of anesthesia before initiation of the procedure. Criteria for judging that a change has occurred include increases in absolute latency of wave V from 0.5 to 2 ms.²³ The strongest change is the loss of a repeatable wave V after all potential sources of interference have been ruled out. Harper and Slavit²⁴ used criteria of greater than a 0.1 ms increase in latency or greater than a 50% decrease in amplitude to identify a change in the response.

Electrocochleography

ECoG is a method of measuring the most peripheral of the auditory evoked responses. ECoG response consists of three primary components: cochlear microphonic, summating potential, and action potential. The action potential is an alternating current potential that is associated with the synchronous discharge of numerous neural fibers located in the basal region of the cochlea. The action potential represents the activity similar to wave I of ABR. The action potential component is most useful for intraoperative monitoring.

ECoG has the advantage of being a near-field recording, and as such requires fewer averages and less time to obtain a response. In addition, the response obtained is larger and easier to interpret. Noninvasive extratympanic recordings require 250 to 1200 samples over a 12 to 60 second time course, whereas transtympanic needle electrodes require only 40 to 100 sweeps over a 2 to 5 second time period. The stimulus should be a broad-band rarefaction click of high intensity (85 to 95 dB normalized HL) with a rate of 21.1/second. An impedance of 100 kΩ may be acceptable in a transtympanic montage compared with the need for extremely low impedance of less than 5 kΩ required for surface electrodes. Recording parameters differ from those of ABR primarily in that the positive recording site is the ipsilateral promontory as opposed to a distant surface placement. Negative and ground electrodes remain the same. Another difference is that the number of samples may be reduced to less than 100 compared with the 1500 suggested for ABR.

Interpretation of Results

As with ABR, an adequate preoperative baseline response must be obtained against which to judge changes observed during the procedure. ECoG has been used to ascertain whether the goal of decompression operation of the endolymphatic shunt has been achieved. It has been thought that summating potential potentials normalize when pressure imbalances of the cochlea have been eliminated. Zappia and colleagues ²⁵ suggested that latency changes of greater than 1 ms and an amplitude decrease of greater than 50% should be considered significant changes in the response, although any measurable deviation should be reported immediately to the surgeon. Significant changes in the action potential indicate either direct or ischemic cochlear injury. After interruption of cochlear blood flow, 20 seconds or more may elapse before changes in ECoG response can be detected. This delay presumably occurs as a result of metabolic reserves that sustain cochlear function until their depletion from prolonged ischemia causes failure of electrophysiologic activity.

It has been suggested that simultaneous recording of ABR and ECoG would result in the most effective monitoring system. ECoG provides an interpretable wave I in cases where it cannot be identified in ABR. Using the responses in combination, it would be possible to monitor changes in interpeak latencies, which could not be done using either alone. ECoG alone is not reflective of activity in the proximal region of CN VIII or the brainstem. It is possible to record a normal-appearing ECoG in the presence of significant CN VIII dysfunction. Monitoring of ABR wave V provides information related to the more central areas of auditory processing.

Direct Eighth Cranial Nerve Recording

It was found that potentials measured directly from the cochlear nerve required significantly fewer averages than those of an ABR (0 to 100 versus hundreds to thousands) to display a recognizable wave pattern. This finding suggested that responses could be updated every few seconds under ideal circumstances. In addition, cochlear nerve action potentials were found to be present when ABR and ECoG recordings had been eroded by tumor or other pathology. The disadvantage of this type of recording in vestibular schwannoma surgery became evident during procedures to treat larger tumors that extended to or into the brainstem. There must be some portion of uninvolved cochlear nerve on which to place the electrode. Practically, many larger tumors are unsuitable for hearing conservation surgery anyway.

Recording electrodes are most commonly placed on or around the intracranial segment of the cochleovestibular nerve. A cotton or fibrous wick sutured to the tip of a polytef (Teflon)-insulated wire secures the electrode to the nerve atraumatically. The locations of the reference and ground electrodes are similar to those described for recording ABR.

The response waveforms obtained from direct recordings of the cochlear nerve typically exhibit triphasic patterns.²⁶ The initial positive deflection (downward) represents nerve activity approaching the recording site. The generally larger negative (upward) deflection occurs as the impulse passes under the electrode. As the depolarization moves away (more centrally) from the recording electrode, another positive wave results and completes the triphasic complex. In the case of CN VIII monitoring, baseline recordings are obtained after craniotomy, but before tumor dissection.

Interpretation of Results

Tumor dissection causing pressure of the cochlear nerve results in immediate loss of compound action potentials (CAP) wave. Stretching the cochlear nerve causes increased latency and broadening of the response waveforms. Partial block of conduction, as may occur with direct pressure by a surgical instrument on the nerve, results in lower amplitude of the prominent negative peak. Only the initial positive deflection is seen after complete conduction block or transection of the nerve.

Otoacoustic Emissions

OAE are low-level sounds measured from the ear canals of humans and animals with intact cochlear function. Transient-evoked OAE and distortion-product OAE may be useful in differentiating sensory (cochlear) from neural (retrocochlear) hearing losses. Poor behavioral hearing thresholds and good emissions suggest a retrocochlear site of lesion. Reduced OAE indicates at least cochlear dysfunction.

Distortion-product OAE have been shown to be very sensitive to hypoxia and interruption of inner ear blood flow. Being sound waves, OAE are described by amplitude, frequency, and phase measures. Amplitude and phase are affected by manipulations that alter cochlear (and presumably outer hair cell) function. Subtle early changes in amplitude were measurable in some subjects within the first 10 seconds after internal auditory artery (IAA) occlusion. More recently, phase changes have been noted within 2 to 3 seconds of IAA compression, suggesting that phase measures may be the most sensitive OAE parameter for monitoring cochlear function during surgery.²⁷

Emissions were stable during various procedures over many hours. Inhalation and intravenous (narcotic) anesthetics did not seem to affect the recording of emissions. Acoustic noise has the potential to undermine the recording of OAE. This effect was particularly prominent when measuring emissions at frequencies less than 2 kHz.²⁸ Based on numerous recordings from various operating rooms, most background acoustic noise energy seemed consistently concentrated at less than 2 kHz. The authors recommended choosing patients for surgical OAE monitoring with preoperatively intact hearing and emissions greater than 2 kHz.²⁸

Some of the noise can be eliminated from the ear canal by sealing the probe and the meatus. A surgical scrub sponge has been used with some success.²⁹ Silicone, wax, and other substances also have been employed in an attempt to solve this problem. Distortion-product OAE have been monitored during loud suctioning of cerebrospinal fluid (relatively high frequency) from the operative field. Emissions could not be monitored during drilling (large-amplitude noise of low and high frequency). Mechanical interference with the conductive hearing mechanism would affect measurement of OAE. Fluid or debris such as bone dust in the external auditory canal or middle ear would interfere with emissions. For this reason, OAE monitoring is unsuitable for many transmastoid surgical approaches.

Transient-evoked OAE result in an amplitude-weighted frequency spectrum that may be repeated throughout the critical portion of the operation.²⁸ Generally, reliable transient-evoked OAE can be obtained in 30 to 40 seconds in ideal conditions and longer intervals in noisy situations.

Using a customized software PC-based system, distortion-product OAE have been monitored during surgery every 2 seconds. Distortion-product OAE seem to be more robust than other types of emissions and are expected to be present in ears with mild to moderate sensory hearing loss (pure tone thresholds ≤45 dB). Also, monitoring programs can be designed to monitor a single or numerous frequency locations of the cochlea. The best frequency for monitoring may change during the case, and the program can be modified to reflect this. Because of these properties, distortion-product OAE at this time seem to be the best OAE suited for intraoperative monitoring.

Because OAE reflect only cochlear function. ABR or direct CN VIII techniques would be appropriate to combine with emissions monitoring. Distortion-product OAE and ABR have been measured during vestibular schwannoma surgery using the same acoustic probe (ER 10-B, Etymotic Research Corp., Elk Grove Village, IL), which contained two speakers and a microphone.²⁹ Rapid switching between the two recording instruments allowed comparisons of the responses. The probe should be secured in the ear canal at the level of the meatus. The pinna, probe, and tubing may be prepared out of the field by folding and securing (with tape or suture) the pinna anteriorly. The tubing should be secured to the patient or operating table in a location to prevent pulling or other disturbance. Baseline measurements obtained after probe insertion and after sterile draping of the operative field help ensure successful monitoring.

Laser-Doppler Cochlear Blood Flow

Cochlear blood flow can be measured using laser Doppler flowmetry.³⁰ The technique involves positioning the end of a needle probe, housing emitter and collector optical fibers, against the cochlear promontory in the middle ear. The angle of the probe is adjusted to obtain maximal flow readings. Although the potential application of this technology is exciting, transferring laser Doppler flow measurement techniques developed in the animal model to human surgery is being done in research trials and is to be studied further before it can be practically applied during tumor removal operations.

Monitoring Other Cranial Nerves

Skull base tumors may involve many cranial nerves. CN III, IV, and VI, which control extraocular muscles, are at risk in tumors of the cavernous sinus. Motor portions of trigeminal nerve may be involved, and CN IX, X, XI, and XII can be involved in large skull base lesions.

Nerves are localized using the same technique as used for the facial nerve electrode, employing subdermal needle electrodes placed with great care. Subdermal needle electrodes can be placed percutaneously in extraocular muscles that are innervated by a respective cranial nerve. Great care is needed to prevent injury to the globe. It is important to secure the electrodes well so that the location of the needle does not change when the patient is repositioned. The opposite forehead serves as a good location for reference electrode to avoid contamination with EMG potentials from ipsilateral facial muscles.³¹

EMG from soft palate (CN IX),³² false vocal cords or laryngeal muscles (CN X),³³ sternocleidomastoid or trapezius muscle (CN XI), or lateral tongue (CN XII) can be used for the lower cranial nerves. CN XII monitoring can be done by placing recording electrodes spaced 1 cm apart in the lateral tongue.³⁴ Each EMG potential is monitored on separate channels and clearly labeled so that the oscilloscope loudspeaker can be set to monitor the channel of importance at a particular time in a case.

CONCLUSION

Intraoperative monitoring of CN VIII has become the standard of care for skull base and CPA surgery. Cost-effectiveness analysis supports facial nerve monitoring in mastoid and middle ear surgery.³⁵ Innovative methods of cranial nerve monitoring using techniques such as intraoperative F wave²⁰ and transcranial electric motor evoked potential measurement^36,³⁷ are being investigated, but have yet to be widely adopted in otolaryngology. The latter technique is used in neurosurgery and vascular surgery, and provides the advantage of not relying on irritating stimulus of the nerve, but rather stimulation of the cerebral cortex. Their use may be limited by the requirement for only intravenous anesthetic agents because of suppression of the signal when inhalational agents are used.

For monitoring hearing, the monitoring paradigm must provide the surgeon with timely and accurate information to prevent, or repair, reversible injuries to auditory structures and to identify the maneuvers that cause hearing loss so that they may be avoided or modified in the future. Generally, the loss of a physiologic response correlates with poor postoperative hearing. Regardless of the monitoring technique used, the presence of an unchanged or reduced response is not predictive of hearing, which limits the prognostic accuracy of current monitoring techniques. There seems to be a consensus to use auditory monitoring during nerve section and microvascular decompression procedures where hearing preservation rates are quite high.

Refining techniques of intraoperative auditory monitoring may ensure better hearing outcomes after acoustic neuroma removal. It is important to separate and distinguish changes in sensory and neural auditory responses to understand how a particular surgical manipulation would affect hearing. In the future, probably a combination of two or more of the methods described in this chapter will be the standard. Ultimately, facial nerve function and hearing outcome depend on many factors, including the surgeon’s experience, tumor size and location, tumor invasiveness, surgical approach, and preoperative function. Refining techniques of CN VIII monitoring will ensure better outcomes in the future.

REFERENCES

1. Krause F. Surgery of the Brain and Spinal Cord. New York: Rebman; 1912.

2. Hilger J.A. Facial nerve stimulator. Trans Am Acad Ophthalmol Otolaryngol. 1964;68:74-76.

3. Delgado T.E., Buchheit W.A., Rosenholtz H.R., Chrissian S. Intraoperative monitoring of facial muscle evoked responses obtained by intracranial stimulation of the facial nerve: A more accurate technique for facial nerve dissection. Neurosurgery. 1979;4:418-421.

4. Moller AR, Jannetta PJ: Preservation of facial function during removal of acoustic neuromas: Use of monopolar constant-voltage stimulation and EMG. J Neurosurg 61:757-760, 2984.

5. Pratt R.L. Iatrogenic facial nerve injury: The role of facial nerve monitoring. Otolaryngol Clin North Am. 1996;29:265-275.

6. Hughes G., Bottomy M., Dickins J., et al. A comparative study of neuropathologic changes following pulsed and direct current stimulation of the mouse sciatic nerve. Am J Otolaryngol. 1980;1:378-384.

7. Hughes G.B., Bottomy M.B., Jackson C.G., et al. Myelin and axon degeneration following direct current peripheral nerve stimulation: A prospective controlled experimental study. Otolaryngol Head Neck Surg. 1981;89:767-775.

8. Chase S.G., Hughes G.B., Dudley A.W.Jr. Neuropathologic changes following direct-current stimulation of the rat sciatic nerve. Otolaryngol Head Neck Surg. 1984;92:615-617.

9. Choung Y.H., Park K., Cho M.J., et al. Systematic facial nerve monitoring in middle ear and mastoid surgeries: “Surgical dehiscence” and “electrical dehiscence. Otolaryngol Head Neck Surg. 2006;135:872-876.