Neuromonitoring in surgery and anesthesia

INTRODUCTION — Electrophysiologic monitoring, or neuromonitoring, is used during surgery to assess the functional integrity of the brain, brainstem, spinal cord, or peripheral and cranial nerves. The goal of monitoring is to alert the surgeon and anesthesiologist to impending injury in order to allow modification of management in time to prevent permanent damage. In some cases, neuromonitoring is used to map areas of the nervous system in order to guide procedural management.

Neuromonitoring can include the recording of spontaneous activity (eg, electroencephalogram and spontaneous electromyogram) or evoked response to stimulus (eg, somatosensory evoked potentials, motor evoked potentials, triggered electromyography, and brainstem auditory evoked potentials). Frequently, multiple techniques are used together in order to increase the utility of monitoring and to overcome limitations of individual techniques [1,2].

Neuromonitoring has become common during many surgical procedures, and it has generally replaced intraoperative wake-up testing during spine surgery. Neuromonitoring is performed by a specialized team with specific expertise in the techniques that are used. In most instances, no "standard of care" exists for intraoperative neuromonitoring, and techniques are chosen by the surgeon and monitoring team in order to assess or protect structures at risk (table 1).

This topic will present an overview of neuromonitoring techniques, the effects of anesthetic agents on recorded signals, and the strategy for responding to electrophysiological changes.

MONITORING MODALITIES — Electroencephalography (EEG), electrocorticography (ECoG), electromyography (EMG), somatosensory evoked potentials (SSEPs), brainstem auditory evoked potentials (BAEPs), and motor evoked potentials (MEPs) are electrophysiologic monitoring techniques that are commonly used in the operating room to improve surgical decision-making and possibly reduce neurologic complications in both adult and pediatric patients [3-25].

Electroencephalography — EEG records electrical activity in the cerebral cortex and can detect ischemia and seizure activity, as well as assess the impact of the anesthetic agents on the brain.

Most often, electrodes are placed on the scalp in a standardized array prior to surgery. Each scalp electrode gives a continuous recording of spontaneous superficial brain activity 2 to 3 cm in diameter [26]. Occasionally, EEG electrodes are placed intraoperatively on the surface of the brain during seizure surgery (ECoG), during awake brain mapping, and for selected tumor resections. (See "Electroencephalography (EEG) in the diagnosis of seizures and epilepsy", section on 'Routine EEG technique'.)

Reductions in cerebral blood flow (CBF) produce rapid, characteristic changes in the EEG. With ischemia, progressive decrease in synaptic activity results in loss of high-frequency activity, loss of power, and ultimately EEG silence [27]. The severity of neurologic injury and expected EEG changes with decreasing cerebral blood flow are shown in a table (figure 1).

EEG monitoring is often used during carotid endarterectomy in order to assess cerebral perfusion during carotid crossclamping. In this setting, EEG slowing or asymmetry between the recordings on the operative and contralateral sides can provide evidence of ischemia. EEG monitoring for carotid endarterectomy is discussed separately.

EEG can also be used during intracranial surgery in order to evaluate the cortex for ischemia. A limitation of EEG monitoring is that surface EEG recordings do not detect ischemia in subcortical regions. If pharmacologic methods are used for metabolic suppression, EEG monitoring can confirm burst suppression, but monitoring for ischemia is then precluded [28].

Intraoperative ECoG, whereby cortical potentials are directly recorded from the brain, may be used during seizure surgery to identify the epileptogenic focus for resection [29]. These surgeries are conducted awake using local anesthesia (scalp block) or during light general anesthesia. EEG monitoring in this context is used to identify seizures from the tissue to be removed.

The EEG is also used to identify unwanted seizure such as that which may be evoked by cortical stimulation (eg, during cortical stimulation for MEP monitoring). (See 'Motor evoked potentials' below.)

Stereotactic electroencephalography is a minimally invasive method of monitoring with implanted leads to localize seizure foci in patients with focal seizures refractory to medical therapy. Leads can be implanted either with stereotactic technique or with robotic assistance during general anesthesia [30].

EEG monitoring for the depth of anesthetic effect is discussed separately. (See "Accidental awareness during general anesthesia", section on 'Brain monitoring'.)

Electromyography — EMG is used to monitor muscle activity, either by spontaneous or evoked compound muscle action potentials (CMAPs). EMG is monitored in muscles innervated by cranial or spinal nerves that are at risk during surgery [3,17]. Continuous, free-running EMG (often called "spontaneous EMG") can identify stretch or blunt mechanical trauma to cranial nerves or spinal roots, which can produce high-frequency trains called neurotonic discharges (figure 2) [17]. Muscle innervated by nonirritated and nonstimulated nerves should otherwise be quiet in their activity during general anesthesia [31]. (See "Overview of electromyography".)

Stimulus-triggered EMG (often called "triggered EMG") can help identify intact nerves. During tumor resection, triggered EMG is used to identify a nerve in order to avoid cutting or coagulating it. A mono- or bipolar stimulator is used within the surgical site to stimulate the nerve, and a resulting CMAP is recorded from the innervated muscle (figure 2) [17,31,32]. An increase in the latency between stimulus and CMAP and/or a reduction in the amplitude of the CMAP may indicate nerve injury. This technique is analogous to surface stimulation and measurement of CMAP performed as part of a diagnostic nerve conduction study. (See "Overview of nerve conduction studies", section on 'Motor nerve conduction'.)

EMG is commonly used in spine surgery involving instrumentation in order to help prevent postoperative radiculopathy (a more common complication than spinal cord injury) by identifying nerve irritation before injury [17,33]. Multiple muscles are usually monitored [17]. EMG triggered by stimulation of pedicle screws or pilot holes can be used to identify malpositioned screws that are too close to nerve roots.

Cranial nerve monitoring, which is utilized during acoustic neuroma and cerebellopontine angle tumor resections, microvascular decompression, skull base surgeries, thyroid and parotid procedures, and radical neck dissection, is a form of EMG, or intraoperative nerve conduction study [3]. Only cranial nerves with a motor component can be monitored with EMG (ie, cranial nerves III, IV, V, VI, VII, IX, X, XI, and XII).

Stimulus-triggered EMG can also be used for brainstem and motor-strip mapping during intracranial tumor surgery. For example, when anatomy is distorted by tumor, brainstem mapping can be performed by recording muscle activity that results from stimulation of cranial nerve motor nuclei using a hand-held stimulation probe. Stimulation can also be used to identify the motor cortex by recording EMG activity from the impacted area (eg, upper extremity, lower extremity, or face) [34-36].

Evoked potentials — Evoked potential monitoring is used to assess the integrity of the tested neural pathway. Somatosensory, visual, and brainstem auditory evoked potentials are used to monitor neurologic structures between peripheral sites where specific stimulations are applied, and responses are recorded from central locations. Motor evoked potentials monitor such structures by stimulating the motor cortex and recording from the epidural space (D-wave) or, more commonly, from distal muscles. Changes in evoked responses can result from technical, positional, pharmacologic, physiologic, or surgical causes [37]. (See 'Managing electrophysiologic changes' below.)

Morphology of the recorded waveform depends both on the site used for stimulation and the site used for recording. The amplitude and latency of the waveform can be analyzed to provide functional neurologic assessment (figure 3).

Amplitude is measured peak-to-trough in microvolts, and latency is measured from stimulus application to peak appearance in milliseconds [38]. Loss of, or change in, the waveform can indicate the need for modification of surgical strategy, patient positioning, and/or patient physiologic management in order to prevent or minimize neurologic system injury. (See 'Managing electrophysiologic changes' below.)

Somatosensory evoked potentials — SSEP is one of the most commonly used evoked potential monitoring modalities in the operating room [38,39]. An electrical stimulus is applied to a peripheral nerve, typically the median or ulnar nerve at the wrist for upper extremity SSEPs and the posterior tibial nerve at the ankle for lower extremity SSEPs, using needle or surface electrodes near the nerve [39]. Motor and sensory components of these large, mixed nerves are stimulated. Activation of the motor components results in visible muscle twitches in distal musculature (confirming stimulation and lack of significant muscle blockade), while activation of the sensory components results in responses that travel along the sensory pathway and ascend to the brain (figure 4).

The monitored neural pathway includes the dorsal root ganglia and the dorsal or posterior column of the spinal cord. Therefore, SSEP monitoring is particularly useful during posterior spine surgery [40].

Most commonly, responses are recorded over Erb's point, the popliteal fossa, the cervical spine, and the sensory cortex. However, responses can also be recorded at other points along the pathway, including over the peripheral nerves and spinal cord (figure 4) [26].

Recordings are obtained with an electrode pair called a "recording montage" (an active electrode and a reference electrode) to measure the voltage difference between the two sites [39]. The sensory cortex is monitored with scalp EEG electrodes placed according to the International 10-20 System of EEG electrode placement [39,41]. A significant decrease in the waveform amplitude or an increase in latency may indicate neurologic dysfunction, such as from surgical injury [39]. Criteria for warning the surgeon when monitoring SSEPs are a 50 percent reduction in amplitude or a 10 percent increase in latency [42]. Warning criteria for MEPs and BAEPs are less firmly established, however many centers consider a 50 percent reduction in amplitude significant [42-44].

SSEPs are used in a variety of spine, intracranial, endovascular, and cardiovascular surgeries (table 1).

Brainstem auditory evoked potentials — Brainstem auditory evoked potentials (BAEPs), also known as brainstem auditory evoked responses or auditory brainstem responses, are specialized sensory evoked potentials. An acoustic stimulus (a loud, repetitive click) is made in the ear canal using an ear insert device. The sound is transduced by ear structures, with information conducted to the brainstem via the eighth cranial nerve [40,45]. Recording electrodes are placed at the head near the ear (ie, mastoid process or ear lobe). Five main short-latency peaks (I to V) are usually seen within the first 10 milliseconds after stimulation [40].

Evaluation of BAEPs usually focuses on waves I, III, and V, which are thought to be generated near specific brainstem structures. Wave I is from cranial nerve VIII, wave III from the lower pons, and wave V from near the inferior colliculus (mesencephalon) (figure 5) [3,38,45,46].

BAEPs are used during posterior fossa surgeries to help assess brainstem function and preserve hearing when it is at risk [38,40,45,47,48].

Visual evoked potentials — Visual evoked potentials (VEPs) are specialized sensory evoked potentials of the visual pathway recorded from the visual cortex after flash stimulation of the retina through closed eyelids [40,49,50].

Usefulness of VEP monitoring has not been established, and concerns have been raised that traditional stimulation methods do not produce responses that follow the pathway of useful vision [49,51-53]. Although VEPs are susceptible to effects from general anesthetics and technical problems with stimulators make monitoring difficult [38,49], newer stimulating methods have been used successfully [51,52].

Motor evoked potentials — MEPs can be obtained by magnetic stimulation or, more commonly, by transcranial electrical stimulation, either to the scalp utilizing two needles or by direct stimulation of the surface of the brain (figure 6) [3,18].

Stimulation of the pyramidal and internuncial cells in the brain generate D (direct) and I (indirect) waves recorded from the epidural space; however, responses are most commonly recorded as compound muscle action potentials in peripheral muscle groups. For the upper extremity, the adductor pollicis brevis is usually monitored, while the tibialis anterior, lateral gastrocnemius, and/or abductor hallucis are monitored for the lower extremity [18]. Thus, MEPs monitor the corticospinal tract (ie, motor cortex, corticospinal tract, nerve root, and peripheral nerve) [2].

MEP can also be monitored as epidural D-waves. These are recorded from an epidural electrode and are specific for the corticospinal tract, but they do not differentiate laterality. The electrodes can be placed percutaneously, but they are usually placed by the surgeon in the operative field. This technique is most often used in intramedullary spinal cord tumor surgery [20,54].

For spine surgery, both MEPs and SSEPs are used to monitor spinal cord function to increase sensitivity. Motor and sensory tracts are anatomically distinct and have different vascular supply in areas of the cortex, brainstem, and spinal cord [3,22].

MEPs are monitored periodically throughout surgery and more frequently during critical surgical maneuvers [18]. As with the other evoked response techniques, the amplitude and latency of the response are monitored; a decrease in amplitude is a more common sign of impending neurologic compromise than an increase in latency [20]. Several criteria have been proposed for identifying significant intraoperative change; complete loss of signal is always considered significant [2,3,18,22]. MEPs are more effective than SSEPs for detecting motor injury since changes in the MEPs precede SSEP changes, usually allowing time to react in order to prevent neurologic damage.

Transcranial stimulation for MEP monitoring activates the muscles of mastication. After induction of anesthesia, one or two soft bite block(s) must be placed between the molars, making sure that the tongue and cheek are clear of the teeth. Severe tongue and cheek injuries are possible with repeated stimulation during surgery [55]. The bite block(s) must be checked again for proper positioning when the patient is turned prone for spine surgery and periodically throughout the surgery.

A number of patient factors can make MEP recording difficult, especially for the lower extremity, including diabetic neuropathy, hypertension, age extremes, and preoperative motor deficit [56,57]. However, with permissive anesthetic and neuromonitoring stimulating techniques, MEPs can be reliably obtained even in very young pediatric patients [58].

H-reflex — Intraoperative Hoffmann reflex (H-reflex) monitoring can be used as an adjunct to motor tract assessment. The H-reflex is a true reflex with an afferent arc mediated by large, fast-conducting group 1a fibers, and an efferent arc mediated by alpha motor neurons. It is elicited by electrical stimulation of an afferent mixed peripheral nerve and by then recording the muscle response. Most commonly, the tibial nerve is stimulated, with recording from the gastrocnemius or soleus muscles. The H-reflex can be monitored below the level where the spinal cord ends (where MEPs may be less useful) and in some patients in whom MEP cannot be monitored [13,26,59,60]. It monitors the pathway of the reflex and may be suppressed by injury of more cephalad motor tracts [60,61]. The sensitivity and specificity of H-reflex monitoring have not been fully explored. (See "Overview of nerve conduction studies", section on 'H reflex'.)

ANESTHETIC EFFECTS ON NEUROMONITORING — Most anesthetic drugs alter neural function by producing dose-dependent depression in synaptic activity. As such, the anesthetic effects vary with the location of the synapses, with greatest effects on motor evoked potential (MEP) muscle responses and cortical sensory potentials [31]. In general, inhalational agents have greater effects on all modes of neuromonitoring than do intravenous (IV) anesthetic drugs. Choice and dose of anesthetic drugs should be tailored to the individual patient and the modalities used. Every effort should be made to keep the level of anesthesia constant during critical monitoring periods in order to avoid confounding interpretation of changes.

All anesthetic drugs affect the electroencephalogram (EEG); this is the basis for the use of many monitors of the impact of anesthesia on the brain [62] (see "Accidental awareness during general anesthesia", section on 'Brain monitoring'). It is important for the neurophysiologist who interprets the EEG to know which drugs and doses are administered, but the anesthetic technique is not generally chosen because of effects on the EEG unless the procedure requires direct cortical recording (eg, for ablation of an epileptic focus).

Effects of commonly administered anesthetic medications on somatosensory evoked potential (SSEP) and MEP monitoring are shown in a table (table 2).

Volatile inhalation agents — Volatile halogenated inhalation anesthetics, including isoflurane, sevoflurane, desflurane, and halothane, cause a dose-dependent decrease in amplitude and an increase in latency of evoked responses, to varying degrees. Their effects are much greater on cortical responses than on subcortical responses [63-66].

●EEG – Except for halothane, volatile agents produce an initial increase in frequency of the frontal EEG, with decrease in frequency and amplitude at higher doses, burst suppression at about 1.5 minimum alveolar concentration (MAC), and electrical silence at high doses. Sevoflurane can produce seizure activity with high-dose mask induction [62].

●Somatosensory evoked potentials (SSEPs) – For patients without neurologic pathology, adequate SSEPs can usually be recorded at 0.5 to 1 MAC of volatile agents. However, for patients with any degree of baseline neurologic impairment, even low levels of inhalation agents may abolish potentials and make monitoring impossible.

●Motor evoked potentials (MEPs) – MEP responses are affected by even very low concentrations of volatile anesthetic agents. In general, total intravenous anesthesia (TIVA) facilitates MEP monitoring. However, inhalation agents at 0.5 MAC or less can be used in many patients, especially during intracranial surgery [67,68].

●Brainstem auditory evoked potentials (BAEPs) – BAEPs are resistant to the effects of inhalation anesthetics as compared with other forms of neuromonitoring [3,40,45,46,69].

●Visual evoked potentials (VEPs) – VEPs are sensitive to inhalation anesthetics. However, in a randomized trial of 60 patients who underwent spine surgery with VEP monitoring, flash VEP amplitudes were similar in patients who received a balanced anesthetic with 0.5 MAC sevoflurane, propofol, and remifentanil versus a propofol and remifentanil total intravenous anesthetic (TIVA) at similar anesthetic depths [70].

Nitrous oxide — The effect of nitrous oxide (N₂O) on neuromonitoring is similar to that of the volatile inhalation agents, and it is synergistic when administered with a volatile agent [63]. The effect of N₂O on the EEG is usually one of increasing frequency, but the effects vary depending on the other agents also being administered.

Intravenous agents — IV anesthetics such as propofol, barbiturates, and opioids have less of an effect on monitoring than inhalation agents, though very deep anesthesia, even with propofol, can affect waveforms, especially MEPs. When inhalation anesthetics must be avoided because of interference with monitoring, TIVA is used, usually with a combination of hypnotic and opioid medication. Except those noted below, all IV anesthetics cause a dose-dependent decrease in EEG amplitude and frequency, ultimately producing burst suppression and electrical silence at high doses.

●Propofol – While propofol decreases the amplitude of cortical evoked potentials, monitoring of evoked potentials in neurologically intact patients is usually possible at clinically relevant doses [71]. An advantage to the use of propofol is the ability to quickly titrate the level of anesthesia in response to evoked responses. Higher doses of propofol can have an effect on the spinal cord (glycine receptors) and alter MEPs, especially from the lower extremities [72,73]. A rare patient may develop lactic acidosis due to propofol infusion syndrome (PRIS).

●Barbiturates – SSEP responses are unaffected by administration of barbiturates (sodium thiopental, pentobarbital) even at very high doses [74-76]. MEP responses are quite sensitive to most barbiturates [77], however, MEP monitoring has been conducted successfully during TIVA with methohexital [78]. Small doses of methohexital increase EEG activity and have been used to identify seizure foci.

●Benzodiazepines – SSEP and MEP responses are unaffected by administration of benzodiazepines (eg, midazolam) at low doses used for premedication and amnesia [79]. Benzodiazepines are anticonvulsant and produce EEG slowing, but usually not burst suppression or electrical silence.

●Ketamine – Ketamine has been shown to enhance cortical SSEP and MEP amplitude and to partially reverse the depressant effect of N₂O on SSEPs [80,81]. Thus, the use of a ketamine infusion, which is often administered during spine surgery, may be beneficial when neuromonitoring is used. However, larger bolus doses of ketamine have been shown to decrease MEP amplitude [82]. The effects of ketamine on the EEG differ from those of most IV agents, as ketamine can increase EEG amplitude and frequency and evoke seizures in patients with epilepsy. Hallucinations have been reported in some patients who have received ketamine [83]. Thus it is often avoided toward the end of an anesthetic, and midazolam may be co-administered to reduce the incidence of hallucinations. (See "General anesthesia: Intravenous induction agents", section on 'Ketamine' and "Maintenance of general anesthesia: Overview", section on 'Ketamine'.)

●Etomidate – Etomidate increases the amplitude of cortical SSEP recordings without changes in peripheral evoked potentials or subcortical responses, and in muscle responses of MEP. This effect lasts throughout infusion of etomidate and has been used to enhance cortical SSEPs [79,84]. Etomidate may therefore be useful for induction of anesthesia and as a component of TIVA when neuromonitoring is planned. However, the use of etomidate may be limited by its association with adrenal suppression and concern about possible worsened outcomes in septic patients. (See "General anesthesia: Intravenous induction agents", section on 'Disadvantages and adverse effects'.)

The effect of etomidate on the EEG is also unusual in that low doses can evoke seizure activity in patients with epilepsy, and etomidate is used in electrocorticography (ECoG) to locate seizure foci. (See "General anesthesia: Intravenous induction agents", section on 'Etomidate'.)

●Dexmedetomidine – We do not usually administer dexmedetomidine during MEP monitoring. There is limited, conflicting literature on the effects of dexmedetomidine on MEP during anesthesia. In some small studies and case reports, SSEPs and MEPs were recordable at lower doses, but MEP recordings may be lost at higher doses [85-88]. In one retrospective study of dexmedetomidine infusion during pediatric spine surgery, even low dose infusion reduced MEP amplitude [89]. The effect of dexmedetomidine on SSEP is minimal.

Dexmedetomidine is not an anticonvulsant and produces an EEG similar to slow-wave sleep. It has little to no effect on ECoG monitoring [25].

●Opioids – IV opioids cause small, dose-dependent depression of SSEP and MEP responses, though even at very high doses of opioids, evoked potentials can be recorded [90-92]. Infusions of remifentanil, fentanyl, or sufentanil are commonly used as part of TIVA during neuromonitoring. Opioids tend to produce high-amplitude slow waves in the EEG.

●Lidocaine – While there is limited literature evaluating effects on neuromonitoring during currently used anesthesia regimens, it appears that lidocaine can be used as part of IV anesthesia without interference. In two reports, however, the addition of lidocaine to N₂O anesthesia was associated with reduced amplitude and increased latency of the SSEP [93-95]. Although high doses of lidocaine are proconvulsant, low doses are anticonvulsant, and infusions have been used to treat status epilepticus.

Balanced anesthetic approach — When SSEPs and MEPs are monitored in patients without significant neurologic impairment, we use a balanced anesthetic using both a low-dose inhalation anesthetic (eg, up to 0.5 MAC sevoflurane) and low- to medium-dose propofol (eg, propofol, 0 to 75 mcg/kg/min IV) with a relatively high-dose opioid (eg, remifentanil 0.1 to 0.3 mcg/kg/min), which offers several advantages:

●Movement with motor stimulation is reduced, which is particularly important during intracranial aneurysm surgery.

●The addition of a 0.3 to 0.5 MAC inhalation agent may reduce the chance of awareness under anesthesia.

●Compared with TIVA, the addition of a 0.5 MAC inhalation agent allows reduction of the dose of propofol infusion, facilitating more rapid wakeup and earlier neurologic examination.

●Compared with TIVA, the chance of accidental interruption of the anesthetic for mechanical reasons (ie, kinked or infiltrated IV catheter or tubing such that IV agents no longer infuse) is reduced.

Inadequate evoked potentials — In most neurologically normal patients, adequate evoked responses can be obtained with a balanced anesthetic regimen, as described [67]. Rarely, inadequate potentials will require conversion of the anesthetic to TIVA, with elimination of the inhalation agent. These cases will require higher doses of propofol infusion and may benefit from the addition of ketamine and lidocaine, if appropriate, to avoid higher doses of propofol which may prevent recording of MEP. (See "Maintenance of general anesthesia: Overview", section on 'Adjuvant agents'.)

Neuromuscular blocking agents — MEP and electromyography (EMG) monitoring are affected by neuromuscular blockade; monitoring is not possible with complete paralysis [96]. The plan for administration of neuromuscular blocking agents (NMBAs) should be coordinated with the monitoring team and with the surgeon. The approach to the use of NMBAs when neuromonitoring is planned is as follows:

●Endotracheal intubation – When neuromonitoring must be performed very shortly after endotracheal intubation, a short-acting NMBA, or no NMBA, should be used for intubation. Succinylcholine can be administered to facilitate intubation, as recovery from paralysis will be complete within several minutes, in order to allow baseline testing prior to surgical positioning. When succinylcholine cannot be used (patients with burns, denervating injuries, neuromuscular disease), other NMBAs should be used with caution (as described below); otherwise, techniques that do not require NMBAs should be considered (eg, remifentanil 2.5 to 5 mcg/kg IV with propofol 2 mg/kg and ephedrine 10 to 15 mg IV, modified for patient factors). If rocuronium or vecuronium is used for intubation, sugammadex can be used to facilitate rapid reversal. However, use of reversal may then complicate need for subsequent muscle relaxation, such as during the exposure phase of spine surgery. (See "Rapid sequence induction and intubation (RSII) for anesthesia", section on 'Remifentanil intubation' and "Clinical use of neuromuscular blocking agents in anesthesia", section on 'Sugammadex'.)

Intermediate-duration (ie, 45 minutes of paralysis) nondepolarizing NMBAs (eg, rocuronium, vecuronium, cisatracurium) can be used for intubation when baseline neuromonitoring is not needed before positioning. These NMBAs are useful for craniotomy for aneurysm clipping or for patients with increased intracranial pressure. They facilitate placement of skull pins and positioning, but resolution of relaxation should be assured (and muscle relaxation reversed if necessary) before monitoring baselines are needed [97].

●NMBAs during surgery – During EMG monitoring, nondepolarizing NMBAs (ie, rocuronium, vecuronium, cisatracurium) should be avoided or monitored very closely. During MEP monitoring, the decision to use nondepolarizing NMBAs must balance the need to maintain enough neuromuscular function to record MEPs with the provision of enough relaxation to prevent excessive patient movement with stimulation.

If NMBAs are used, the degree of paralysis must be monitored, usually with a train-of-four (TOF) peripheral nerve stimulator, aiming for two of four twitches or a single twitch of 10 to 20 percent of baseline [96]. Importantly, the level of paralysis must be kept constant, which is best accomplished by the use of an infusion of the NMBA.

Ordinarily during anesthesia, the level of neuromuscular blockade is monitored using a TOF peripheral nerve stimulator, which may not be practical with the positioning for some surgical procedures. The neuromonitoring technologist can measure TOF in the individual muscles being monitored in order to more accurately measure NMBA effects [98]. Effects of NMBAs may vary between muscle groups. (See "Monitoring neuromuscular blockade", section on 'Nerves that may be monitored'.)

PHYSIOLOGIC EFFECTS ON NEUROMONITORING — Physiologic alterations can affect electrophysiologic (EP) signals, and the manipulation of physiological parameters can support the patient during surgically caused EP changes.

Blood pressure — Reduction in systemic and regional or local blood pressure can affect electroencephalography (EEG) and cortical evoked potential monitoring [99]. The EEG and somatosensory evoked potential (SSEP) changes associated with reduction in cerebral blood flow (CBF) and possible embolic episodes are the basis for monitoring patients who undergo carotid endarterectomy.

Even at normal mean systemic arterial pressure, local factors (eg, spinal distraction, vascular compromise from positioning, retractor pressure) may result in ischemia that may be reflected in neuromonitoring changes [100-102]. In addition, patients with baseline abnormal cerebral or spinal perfusion or autoregulation may be at higher risk of ischemia with any perturbation of blood flow. Thus, when monitoring changes occur, we increase mean arterial pressure as appropriate in order to increase tissue perfusion pressure. (See 'Managing electrophysiologic changes' below.)

Ventilation — Changes in partial pressure of oxygen (PaO₂) and partial pressure of carbon dioxide (PaCO₂) can affect neuromonitoring, either by changing oxygen delivery or due to changes in blood flow, especially in patients with compromised vascular anatomy. These effects are seen only at extremes of PaO₂ and PaCO₂ [103-107].

Temperature — Body temperature has been shown to alter EEG, SSEPs, brainstem auditory evoked potentials (BAEPs), visual evoked potentials (VEPs), and motor evoked potentials (MEPs). It has been suggested that core body temperature should be maintained within 2 to 2.5°C of baseline temperature [108].

MEP waveforms can be detected during hypothermia to core temperatures of 31 to 34°C, with increasing latencies and stimulation threshold below 32°C [109-111]. These effects are relevant only for patients who are cooled (eg, cardiopulmonary bypass or circulatory arrest). However, regional hypothermia, such as that caused by a cool extremity after infusion of cold intravenous (IV) solution or by the spinal cord being exposed to cold irrigation prior to instrumentation, can also affect evoked potential monitoring [112].

Hematocrit — Anemia can affect neuromonitoring, by either reducing oxygen carrying capacity or changing the rheology of blood. The effect of anemia on SSEPs and VEPs has been studied using progressive isovolemic hemodilution. Latencies increased with hematocrits of 10 to 15 percent, and with those below 10 percent, amplitudes decreased and latencies increased further [113].

PATIENT POSITIONING EFFECTS — Patient positioning for surgery can cause neurologic and/or vascular compromise and can therefore affect neuromonitoring.

In certain patients, when severe neck flexion is required with positioning, such as for certain spine or intracranial posterior fossa surgeries, baseline motor and somatosensory evoked potentials are ideally performed after induction with the patient supine, and then repeated after turning prone (or sitting). Similarly, for patients with unstable cervical spines, signals are ideally obtained pre- and post-positioning. However, in many patients, baseline evoked potentials can be obtained after the patient is positioned for surgery. If recorded potentials deteriorate or are absent, changes to the position can be made before surgery begins.

Positional electrophysiologic (EP) changes can occur during surgery, after the initial positioning, if limbs are moved, if shoulders sag, or because of pressure on peripheral nerves. (See 'Managing electrophysiologic changes' below.)

ANESTHETIC STRATEGY — During neuromonitoring for spine or intracranial surgery, the choices and doses of anesthetic drugs must be tailored to the monitoring modalities. Usually, multiple monitoring techniques will be used, and the anesthetic management will be determined by the most restrictive modality. During electroencephalography (EEG) monitoring, it is important for the interpreting neurophysiologist to know which drugs and doses are administered, but the anesthetic technique is not generally chosen because of effects on the EEG unless electrocorticography (ECoG) is being used for seizure surgery. If general anesthesia is used for surgery with ECoG, a light anesthetic (eg, low-dose inhalation anesthesia, high-dose opioid infusion, dexmedetomidine infusion) should be used (table 2).

Our approach to anesthesia for the adult patient having spine or intracranial surgery with somatosensory evoked potential (SSEP), motor evoked potential (MEP), and electromyography (EMG) monitoring, and ECoG is as follows, with modifications made for patient factors (see "General anesthesia: Intravenous induction agents" and "Perioperative uses of intravenous opioids in adults: General considerations" and "Clinical use of neuromuscular blocking agents in anesthesia"):

Premedication — Patients can be premedicated in the usual way (eg, midazolam 0.01 to 0.02 mg/kg, fentanyl 1 to 2 mcg/kg). Benzodiazepines are avoided entirely for ECoG monitoring [114].

Induction of anesthesia — Unless special considerations apply, general anesthesia can be induced in most patients in the usual way. Use of neuromuscular blocking agents (NMBAs) to facilitate endotracheal intubation varies depending on the monitoring techniques used. (See 'Neuromuscular blocking agents' above.)

For pediatric patients, inhalation induction with sevoflurane can be used with rapid conversion to a maintenance anesthetic more favorable for the planned neuromonitoring modalities.

Maintenance of anesthesia — During surgery, each neuromonitoring modality will necessitate supportive anesthetic choices.

SSEP Monitoring — For SSEP monitoring, we use a balanced anesthetic approach with a relatively low-dose propofol infusion, a low-dose inhalation agent, and a relatively high-dose opioid infusion, as follows (see 'Balanced anesthetic approach' above):

●Propofol infusion (with inhalational volatile anesthetic) – 40 to 75 mcg/kg/min, discontinued 45 to 60 minutes before emergence

●Inhalation anesthesia – ≤0.5 MAC sevoflurane, although monitoring may be possible with up to 1 MAC in neurologically normal patients

•May require conversion to total intravenous anesthesia (TIVA) if SSEP cannot be recorded (see 'Additional maintenance considerations' below)

•Avoid N₂O due to its effects on both SSEPs and MEPs and its additive effects with other agents [79]

●Opioid infusion – Options:

•Remifentanil, 0.1 to 0.3 mcg/kg/min IV or

•Sufentanil, 1 mcg/kg/hour IV for the first hour, then decreased to 0.2 to 1 mcg/kg/hour IV, discontinued 45 to 60 minutes before emergence or

•Fentanyl, 10 mcg/kg/hr IV for first hour, then decreased to 1 to 10 mcg/kg/hour IV, discontinued 45 to 60 minutes before emergence

The use of opioid infusions during anesthesia is discussed separately. (See "Perioperative uses of intravenous opioids in adults: General considerations".)

●NMBAs – No restrictions with SSEP, as muscle relaxation can improve SSEP responses [42]

MEP monitoring — For MEP monitoring, we use the same anesthetic approach as for SSEP monitoring with the exception of restrictions on NMBA administration, as follows:

●NMBAs (see 'Neuromuscular blocking agents' above)

•For induction, if prepositioning baseline MEPs are planned, and for patients with hemifacial spasm when the lateral spread response is monitored, options for intubation include:

-Succinylcholine 1 to 1.5 mg/kg IV or short-acting NMBA (eg, where available, mivacurium 0.2 mg/kg IV)

-High-dose remifentanil 2.5 to 5 mcg/kg IV with ephedrine 10 to 15 mg IV (administered first) and propofol 2 mg/kg

-Nondepolarizing steroidal NMBA (eg, rocuronium 0.7 to 1 mg/kg IV) followed by reversal of neuromuscular blockade with sugammadex (2 to 16 mg/kg IV depending on depth of neuromuscular blockade) prior to MEP signal acquisition

-Awake intubation, if indicated (see "Management of the difficult airway for general anesthesia in adults", section on 'Awake intubation')

-Lidocaine 4% (4 mL) can be applied through vocal cords to the trachea immediately prior to intubation to minimize tracheal irritation from the endotracheal tube and coughing during subsequent patient positioning. Note that topical lidocaine on the vocal cords may block the vocal cord EMG response during anterior neck procedures (eg, during thyroidectomy).

•If prepositioning baseline MEPs are not required, options for intubation include:

-Nondepolarizing NMBA for intubation (eg, rocuronium 0.4 to 0.6 mg/kg ideal body weight IV)

-Succinylcholine for rapid sequence induction

-Awake intubation in cases of airway concerns (see 'Neuromuscular blocking agents' above)

•We avoid NMBAs during maintenance of anesthesia unless they are requested for specific surgical indications (eg, abdominal exposure for anterior spine surgery). The dose and timing of NMBA are chosen in order to allow recovery or reversal of neuromuscular block during periods of high surgical risk.

●Bite block – We place a soft bite block between the molars and make sure that the tongue and soft tissue are pushed away from between the teeth in order to avoid injury during jaw clench. Also, we check the tongue position to make sure it is not pushed backward in order to prevent possible ischemia of the tongue.

EMG monitoring — When EMG is the only monitoring technique used, we use general inhalation anesthesia as follows:

●Inhalation agent – ≤1.5 MAC

●Opioid – No restrictions

●NMBAs – As for MEP

EEG and ECoG monitoring — When the EEG is employed for monitoring of cerebral ischemia (eg, carotid endarterectomy), prominent EEG activity should be maintained by avoiding an excessive depth of anesthesia (eg, using low-dose inhalation agent with opioid supplementation).

When EEG is being used during surgery to identify a seizure focus, inhalational agents are limited (eg, 0.5 MAC inhalation agent), benzodiazepines are avoided entirely, and propofol is avoided during the period of monitoring. A variety of anesthetic regimens can be used for ECoG monitoring, and no one regimen has been found to be superior [114]. Our anesthetic regimen preferences are listed below.

Options for ECoG with general anesthesia:

●High-dose remifentanil infusion (0.2 to 0.4 mcg/kg/minute) with 0.5 MAC volatile inhalation agent in addition to muscle relaxant

or

●High-dose remifentanil infusion with 50 to 70 percent nitrous oxide in addition to muscle relaxant

or

●Remifentanil infusion (0.05 to 0.3 mcg/kg/minute), dexmedetomidine infusion (0.2 to 1 mcg/kg/hour), and low-dose propofol (25 mcg/kg/minute) in addition to muscle relaxant

Occasionally, it may be necessary to enhance interictal epileptiform discharges pharmacologically [25]. In these cases, we prefer to administer alfentanil 20 to 100 mcg/kg IV. Alternatives include methohexital (25 to 50 mg IV) or etomidate (0.2 mg/kg) [114].

Additional maintenance considerations

●When acquisition of baseline recordings is likely to be difficult, such as for patients with severe diabetes, neurologic spinal disease, or edematous extremities, inhalation agents may be withheld until baseline recording is established.

●We usually use a processed EEG monitor to help guide the dose of propofol infusion during the procedure, and to assess for the time of extubation.

●If TIVA is required for neuromonitoring, we use the following, modified for patient factors (see "Maintenance of general anesthesia: Overview", section on 'Total intravenous anesthesia'):

•Propofol infusion – 80 to 150 mcg/kg/min

•Opioid infusion options (see 'SSEP Monitoring' above):

-Remifentanil (0.1 to 0.3 mcg/kg/min) or

-Sufentanil (0.2 to 1 mcg/kg/hour) or

-Fentanyl (1 to 10 mcg/kg/hour)

●When burst suppression is required during aneurysm clipping, we titrate propofol to the EEG recording (eg, increase the propofol infusion to 150 mcg/kg/min); after the aneurysm is clipped, we discontinue propofol and administer remifentanil 0.1 to 0.3 mcg/kg/min and a 0.5 MAC inhalation agent for the remainder of the case.

Supplementary anesthesia — If additional anesthesia is required (eg, for patients with chronic opioid use or to allow reduction of propofol dose because of interference with monitoring), options include:

●Ketamine bolus 0.5 mg/kg IV or infusion 0.3 to 1 mg/kg/hour IV, discontinued for final one to two hours of surgery

●Lidocaine infusion 1.5 mg/kg/hour IV

MANAGING ELECTROPHYSIOLOGIC CHANGES — Changes in EP responses during neuromonitoring can result from a number of factors, which can be organized into five categories: surgical, pharmacologic (eg, anesthetic agents), physiological (eg, hypoxia, hypotension, hypothermia, anemia), positional (eg, extreme head position, peripheral nerve compression, spine flexion or extension), and technical (eg, lead failure or dislodgement, electrical interference from operating room equipment, inaccurate locations of stimulating and recording leads) (table 3).

When a neuromonitoring change occurs intraoperatively, the surgeon, anesthesiologist, and neuromonitoring team should work together to determine the etiology of the change and to correct the cause. The following steps provide a framework for approaching a neuromonitoring change. However, a modified approach should be followed during certain surgical steps, such as temporary or permanent aneurysm clipping or placement of spinal instrumentation.

●Confirm the change – When encountered, neurophysiologic changes should be reconfirmed by repeat signal acquisition (if time allows) before a call for action is initiated. Changes in other monitored modalities will also help in the evaluation.

●Alert the team – Management of a change in neuromonitoring should be a team effort. The surgeon, anesthesiologist, and neuromonitoring personnel will all play a role in diagnosis and management.

●Troubleshoot the cause – Depending on the monitoring modality in use, the differential diagnosis may be narrowed by identifying the change as cortical versus subcortical, unilateral versus bilateral, and focal versus global. In general, surgical, technical, and positional etiologies result in more localized changes, while anesthesia and physiologic etiologies tend to result in more global effects on neuromonitoring. However, anesthesia-related changes in evoked potentials may be more pronounced on the side of preexisting asymmetric neurologic deficits (eg, nerve root compression or unilateral spinal cord pathology). The temporal relationship between monitoring changes and surgical and anesthesia maneuvers may also help with diagnosis.

At certain stages of surgery, such as temporary or permanent clipping during cerebral aneurysm surgery, spinal instrumentation, or crossclamping of the carotid, surgical causes are very high on the list for differential diagnosis. Since prompt correction of etiology of change is necessary in order to avoid neurologic deficit, it is important not to waste precious time on unlikely causes of the change. This is why it is important at such stages not to make anesthetic changes.

●Optimize physiology – While the cause is being investigated, the anesthesiologist can help minimize potential neurologic injury by optimizing physiologic variables (eg, raise blood pressure if appropriate, increase fraction of inspired oxygen).

●Correct the cause – If the cause is identified, it should be corrected, when possible. Further treatment may depend on subsequent monitoring and the clinical situation. Some examples include the following:

•If evoked potentials deteriorate during positioning for cervical spine surgery, the head may be repositioned, and, if the potentials do not recover, the patient may be awoken for neurologic exam and the surgery aborted.

•If hypotension or relative hypotension is the cause, increase blood pressure.

•If EEG or SSEP changes occur with carotid crossclamping during carotid endarterectomy, shunt placement should be considered expeditiously, and the blood pressure should be raised immediately as possible. (See "Anesthesia for carotid endarterectomy and carotid stenting", section on 'Hemodynamic management' and "Carotid endarterectomy", section on 'Carotid shunting'.)

•If MEP or SSEP is changed or lost after temporary clip during aneurysm surgery, release the clip if possible, raise the blood pressure as appropriate, and finish surgery as quickly as possible.

SUMMARY AND RECOMMENDATIONS

●Electroencephalography (EEG) – EEG provides a functional assessment of the cerebral cortex, including the occurrence of ischemia, seizure activity, and the degree of anesthesia drug effects. It is commonly used during carotid endarterectomy in order to assess cerebral perfusion during carotid crossclamping (figure 1). EEG monitoring is also used in the form of electrocorticography (ECoG) during craniotomy for seizure focus ablation (table 1). (See 'Electroencephalography' above.)

●Electromyography (EMG) – EMG is used to monitor spontaneous muscle activity or evoked compound muscle action potentials (CMAPs) in muscles innervated by cranial or spinal nerve roots that are at risk during surgery. Triggered EMG is used during surgery in order to monitor nerve integrity. These techniques are commonly used during posterior fossa and spine surgery, during cortical and brainstem motor mapping, and during anterior neck procedures (eg, thyroidectomy) (table 1). (See 'Electromyography' above.)

●Somatosensory evoked potentials (SSEPs) – SSEPs are among the most commonly employed evoked potential monitoring modalities in the operating room. SSEPs are often used to monitor the sensory pathway in spine and intracranial surgeries (figure 4 and table 1). (See 'Somatosensory evoked potentials' above.)

●Brainstem auditory evoked potentials (BAEPs) – BAEPs are primarily used intraoperatively during posterior fossa surgeries in order to help assess brainstem function and preserve hearing when it is at risk (figure 5). Compared with other intraoperative neuromonitoring modalities, BAEPs are very resistant to general anesthesia (table 2). (See 'Brainstem auditory evoked potentials' above.)

●Motor evoked potentials (MEPs) – MEPs monitor the corticospinal tract. MEPs are monitored along with SSEPs during intracranial and spine surgery (figure 6). MEPs can detect ischemia in an area of the brain that is not covered by SSEPs. Changes in MEPs occur earlier than SSEP changes, which may allow more rapid correction of the cause and possible prevention of neurologic injuries (table 1). (See 'Motor evoked potentials' above.)

●Effects of anesthetics – Anesthetic agents can affect neuromonitoring (table 2). In general, evoked responses are affected more by volatile inhalational agents (eg, isoflurane, sevoflurane, desflurane) and by nitrous oxide (N₂O) than by intravenous (IV) agents. MEPs are very sensitive to inhalation agents, SSEPs are moderately affected, and BAEPs are resistant to the effects of inhalation anesthetics. The level of anesthesia should be kept constant during critical monitoring periods in order to avoid confounding the interpretation of changes. (See 'Anesthetic effects on neuromonitoring' above.)

●Effects of neuromuscular blocking agents (NMBAs) – NMBAs should be avoided or monitored very closely during EMG and MEP monitoring. If NMBAs are used the degree of neuromuscular blockade should be monitored using a train-of-four (TOF) nerve stimulator, aiming for a constant level of paralysis at two of four twitches or single twitch 10 to 20 percent of baseline. (See 'Neuromuscular blocking agents' above.)

●Managing neuromonitoring changes during anesthesia – When a neuromonitoring change occurs, surgical, anesthetic, physiologic, positional, and technical causes should be investigated using a team approach. A checklist is provided (table 3). (See 'Managing electrophysiologic changes' above.)