Awake craniotomy
Awake craniotomy refers to surgery that is performed on the brain while the patient is in a state of awareness and that allows for cooperation with functional testing of the cortex. It is usually performed when eloquent cortical tissue—tissue that is involved in motor, visual, or language function—is located in close proximity to the area to be resected. This may include resection of tumor or ictal foci in patients with epilepsy. The patient’s awake state allows for the mapping of the brain near the area to be resected, which avoids morbidity related to resection of the eloquent tissue and can reduce anesthetic interference with brain mapping.
Awake craniotomy has been shown to reduce the size of the resection, surgery time, postoperative neurologic deficits, early postoperative nausea and vomiting, and hospital stay. Vasopressor use and hypertension during head pinning are also decreased. Hospital stays have been reported to be as short as 1 day for patients with good functional status and uncomplicated tumors. , Lower post-intensive care unit inpatient costs were found in patients undergoing glioma resection under sedation versus general endotracheal anesthesia. Awake craniotomy has even been studied as a potential outpatient procedure with no reported adverse outcomes. This technique has been shown to have better patient acceptance. Despite the most common complaints about pain from the head holder, inadequate local anesthesia, and uncomfortable position, patients report better postoperative pain scores and use less opioids in recovery.
The anesthetic technique for awake craniotomy varies. The common goal is to ensure the best possible resection by keeping patients comfortable and safe with sedation and anesthesia before and after the awake interval, and by monitoring and guiding patients through conscious mapping and testing. General anesthesia—using a laryngeal mask airway (LMA), endotracheal intubation, or varying degrees of sedation—with discontinuation of anesthesia for the period of speech, memory, or motor testing, or a combination of these techniques, has been described , and found to be safe.
Approach to the Awake Craniotomy
Appropriate patient selection and preparation are the most important factors in the success of anesthesia for awake craniotomy. In the selection process, the following should be considered: age and maturity; anxiety, claustrophobia, or other psychiatric disorders; a patent airway; and a history of reflux or nausea and vomiting. Literature and experience suggest that hypertension, alcohol abuse, and lack of maturity may be risk factors for sedation failure. We recommend that children younger than 14 years not be considered, although the developmental status of the individual should be assessed. Furthermore, a patient with a potentially difficult airway or who demonstrates that there is a likelihood that they will obstruct under sedation is a poor candidate for awake craniotomy. Because the patient’s head will be in pins and positioned by the surgeon, sudden conversion to endotracheal intubation may be difficult. A plan must exist between the anesthesia and surgical teams for the management of the airway should patency be lost. Patients with a history of obstructive sleep apnea (OSA), difficult ventilation, or difficult intubation may be considered as higher risk for adverse outcomes, although current evidence does not support OSA as an independent risk factor for failure of awake craniotomy.
Preoperative consultation is essential. The anesthesiologist should clearly outline for the patient what to expect during the procedure, including the varying states of sedation and awareness, the positioning, the possible discomfort, and the testing process. A strong rapport between the patient and the anesthesiologist should be established prior to the procedure, and comfortable patient positioning, scalp block, proper anesthetic selection, and communication are paramount. The anesthesiologist must keep in mind the psychological state of the patient and attempt to alleviate anxiety and discomfort as much as possible to ensure the success of the technique and the surgery.
Positioning
Patient comfort and access during the awake period is important in a successful awake craniotomy. Lateral or semi-lateral positioning is commonly used to allow for patient comfort and to offer the anesthesia team ideal access to the patient. Adequate padding and pillows should be provided and pressure points carefully checked. The patient should confirm an acceptable level of comfort prior to sedation, as he or she will be in pins upon emergence from sedation and must remain still during the period when repositioning is not feasible. The patient must also be positioned and draped for ideal access by the anesthesia team who need to speak with the patient to test motor and sensation. Tenting the drapes upward from the patient on the side of the anesthesia team provides an area of access and may also reduce the patient’s sense of claustrophobia. Fig. 17.1 shows a configuration for setup in the operating room. The patient’s position is stabilized with the use of a deflatable beanbag or a backrest fixed to the operating table, and the patient is taped to the table.
Scalp Block
Performance of a reliable blockade of the innervation of the scalp is essential to the successful performance of an awake craniotomy. The technique for local scalp block for craniotomy is well-described. , Individually blocking the auriculotemporal, zygomaticotemporal, supraorbital, supratrochlear, lesser occipital, and greater occipital nerves is necessary to provide complete analgesia of the scalp. Ropivacaine and levobupivacaine can be safely used up to doses of 4.5 mg/kg and 2.5 mg/kg, respectively. Mepivacaine may also be used adjunctively if a faster setup is required. These blocks achieve peak plasma concentrations approximately 15 minutes after injection. Severe bradycardia after scalp block has been reported.
If general anesthesia is performed for the initial asleep period, necessary access, invasive monitors, and urinary catheters may be placed after induction. However, if sedation is selected, it should be deep enough for the patient to comfortably tolerate these invasive measures with minimal recall. Scalp block may also be performed at this time. The success of the scalp block is likely not known until the placement of head pins begins. Obvious physical response in the sedated patient or an increase in heart rate and blood pressure in the general anesthesia patient would indicate block failure. Boluses of propofol may be necessary to temporarily rescue the inadequately sedated patient.
Anesthetic Options
A number of different anesthetic techniques may be useful for awake craniotomy. Some providers may choose varying levels of sedation as tolerated by the patient. Others may choose a general anesthetic, with or without endotracheal intubation, in what is referred to as the asleep-awake-asleep technique.
One may choose to perform a general anesthetic from induction to completion of exposure and awaken the patient for neurocognitive and neurofunctional testing. If this method is chosen, it is important to remember that the patient will be emerging in head pins and bucking must be avoided to prevent patient morbidity. Conversely, one may choose sedation during the exposure period, keeping in mind that airway reflexes should be preserved.
Propofol, dexmedetomidine, and opioid infusions have been safely and successfully used for awake craniotomy. Volatile anesthetics have been used for general anesthesia during the asleep portion of the procedure. While the patient is in head pins, the provider should be aware of the danger of laryngospasm during emergence and extubation or LMA removal.
Total intravenous anesthesia (TIVA) is a viable choice for awake craniotomy. Propofol-only anesthesia with spontaneously breathing patients has been described as safe. Propofol is begun with a bolus of 0.5 mg/kg and continued at a rate of 75–250 μg/kg/min. Practitioners have safely and successfully used a combination of propofol or dexmedetomidine with an opioid such as remifentanil, sufentanil, or boluses of fentanyl. The rapid clearance of remifentanil makes it an appealing choice for quickly achieving the awake state. However, remifentanil may cause hypopnea in the spontaneously ventilating patient. A dose range of 0.01–0.1 μg/kg/min has been described, although when used alone, we have found efficacy as high as 0.2 μg/kg/min. In Manninen and colleagues’ 2006 study, propofol infusion combined with intermittent fentanyl yielded similar patient satisfaction, recall, and intraoperative complications to remifentanil, with a slightly higher rate of respiratory depression in the propofol and fentanyl group. Emergence from remifentanil-propofol has been described as approximately 9 minutes. Sedation with remifentanil infusion alone is performed in some centers, although no data have been published at this time. Alfentanil is known to induce epileptiform discharges in the hippocampal area and should be used with caution in patients with complex partial epilepsy. For patient comfort, an opioid infusion may be continued at a low dose during testing, titrated to effective patient cooperation. Longer-acting analgesia may be necessary prior to emergence, although emergence on low-dose opioid infusion continued to recovery may be used.
The alpha-2-agonist dexmedetomidine also may be used and has been recommended, due to its lack of interference with electrophysiologic testing, sedation with minimal respiratory effects, , and anxiolytic and analgesic qualities. A loading dose of 1 μg/kg is delivered over 10–15 minutes with an infusion rate of 0.1–0.6 μg/kg/h. Doses are higher in children. Dexmedetomidine use as a lone sedating agent has been described, as well as a combined anesthetic with propofol or opioids, or both. The use of remifentanil combined with dexmedetomidine has also been reported. Like remifentanil, dexmedetomidine may be continued at low doses during brain mapping and functional testing if needed for patient comfort. Dexmedetomidine is known to have a significant synergistic effect when used in combination with other sedative agents.
Droperidol and fentanyl were commonly used in the past, but neuroleptanalgesia has given way to faster acting and more quickly eliminated regimens.
Brain Mapping and Cognitive Testing
As patients emerge from deep sedation or anesthesia, the anesthesiologist must take responsibility for safely re-orienting the patient, providing a calming influence, and guiding him or her through the brain mapping and cognitive testing phase. The use of bispectral-index monitoring to shorten emergence has been described and may be useful. The patient may be disoriented for a brief period after sedation. Again, preoperative preparation becomes essential during this phase. As the patient is guided through the process, he or she must be reassured that involuntary movements and speech patterns may occur as a result of cortical stimulation by the surgical team. The anesthesia team must be prepared to address any anxiety and discomfort that may occur. Motor, sensory, cognitive, and speech testing may be performed during this time. The patient may be asked to verbally identify objects or pictures, read passages aloud, perform specific motor tasks, or identify paresthesias or other sensations. Cortical evoked potentials and electrocorticography (ECoG) may be used to identify functional tissue and seizure foci. The results of this testing will guide the surgeon in removing pathologic tissue with minimal disruption to eloquent tissue in order to reduce patient morbidity. Surgical resection then proceeds while the patient completes verbal tasks (speech area assessment) or performs motor tasks (motor area assessment). Seizure activity is also possible during this phase, and the anesthesiologist must be prepared for prompt treatment, possible airway intervention, and conversion to general endotracheal anesthesia.
When brain mapping and functional testing are complete, the patient should be sedated once more for closure of the dura, calvarium, and scalp. This period can be very stimulating to the patient and adequate sedation can usually be achieved with remifentanil, propofol, dexmedetomidine, or a combination of these agents, as has been described.
Adverse Events and Management
Seizures, respiratory depression, nausea, vomiting, anxiety, discomfort, and agitation may occur during awake craniotomy. As is commonly the case with sedation, airway obstruction, hypercarbia, and hypoxemia are all possible, and careful preoperative assessment of the airway is vital. We have also experienced laryngospasm with LMA during the asleep portion of the asleep-awake-asleep technique. An extensive review of anesthetic complications of awake craniotomies showed an 18.4% rate of hypoxemic events for patients undergoing sedation for the procedure compared to merely 1% in patients who received endotracheal intubation. Airway or ventilation complications occurred in just 2% when patients received propofol- only sedation. The rate of conversion to general anesthesia has been reported to be 2%. Dexmedetomidine appears significantly better than propofol for rate of respiratory depression. Dexmedetomidine has been described for rescue of a patient unable to tolerate awake brain mapping after a propofol- remifentanil sedation regimen, and is now used more commonly as a primary sedative. Airway-assist maneuvers and the use of oral or nasal airways are common in patients undergoing sedation and should be expected to treat transient obstruction.
Vomiting and aspiration are possible in the sedated patient. As the airway will be unprotected using this technique, administration of prophylactic antiemetics is advisable, and rapid treatment should be provided if nausea occurs. The incidence of nausea and vomiting is 4% for mixed sedation techniques and even less for the use of propofol. Once symptoms occur, they can be controlled with a hydroxytryptamine-3 receptor (HT-3) antagonist, such as metoclopramide 10 mg. Nausea can also result from inadequate analgesia of dural attachments and meningeal vessels. Additional local anesthetic should be administered by the surgeon and supplemental sedation administered by anesthesia.
Sedation with spontaneous ventilation may pose the problem of brain swelling, particularly when mass-effect already exists, due to hypopnea or periods of apnea and concomitant increase in PaCO 2 . However, spontaneous ventilation also may assist in keeping the brain relaxed due to maintenance of negative intrathoracic pressure and promotion of cerebral venous outflow. Mannitol or furosemide administration may be necessary to reduce swelling and improve the surgical field. Patient movement—with the head in pins or during craniotomy— can have morbid outcomes, including scalp and soft tissue injury, brain swelling from straining, and placing the cervical spine at risk. It is critical to anticipate possible patient movement—during times like emergence from sedation or as a result of seizure initiated during mapping or delirium—and control the movement quickly. Deepening sedation with propofol boluses may be effective, and conversion to general anesthesia must be considered if necessary. It is important to be aware that deepening sedation may result in hypopnea or apnea, and the anesthetic team must be prepared to take control of the airway.
Seizures may occur from electrical stimulation during brain mapping or from a patient’s underlying condition. Vigilance is critical because the untreated seizure while in head pins could be catastrophic. Seizure activity can be treated with propofol (0.75–1.25 mg/kg) or benzodiazepines, depending on the need for subsequent electroencephalograph (EEG) recording. A 4.9% incidence of seizures was reported with cortical mapping in an unselected series of 610 awake craniotomies. At the end of the procedure, benzodiazepines and phenytoin may also be used more freely.
Epilepsy surgery
Epilepsy is a disease of the brain characterized by: two unprovoked seizures greater than 24 hours apart; one unprovoked seizure; and a probability of seizures similar to the general recurrence risk after two unprovoked seizures occurring over the next 10 years, or diagnosis of an epilepsy syndrome. It is present in 0.5–2.2% of the general population. Because 30–40% of epileptics do not respond adequately to pharmacologic intervention, more than 400,000 people still have medically uncontrolled epilepsy in the United States. However, only 10–30% of patients with seizures refractory to medical management are appropriate candidates for seizure surgery, and only 1% eventually undergo the procedure.
Epilepsy is classified as partial, generalized, or psychogenic nonepileptiform seizures (PNES). Partial seizures are characterized by electrical disturbances localized to one area of one cerebral hemisphere. Simple partial seizures are not associated with a loss of consciousness, and generally last 1 minute or less. Complex partial seizures are characterized by a loss of consciousness or awareness and spread from their localized focus to other regions. Complex partial seizures may spread to become generalized. Generalized seizures have no demonstrated focal onset, although they may evolve from focal seizures, affect both hemispheres of the brain, and are characterized by a loss of consciousness. They are sub-categorized as generalized tonic-clonic (grand mal), tonic, myoclonic, absence (petit mal), and atonic. PNES are psychogenic episodes that may be characterized by seizure-like physical manifestations but have no corresponding epileptiform activity on EEG and are considered conversion reactions.
Surgical management of epilepsy may be an option for patients with intractable epilepsy refractory to medical treatment. With successful surgical intervention, lifestyle improves, although most patients continue anticonvulsant therapy. Chin et al. reported that the rate of employment improved only modestly in their group of 375 patients, from 39.5% fully employed status preoperatively to 42.8% postoperatively; however, the rate of part-time employment nearly doubled, from 6.9 to 12.4%.
Anesthetic regimens have a significant effect on cortical mapping for epilepsy and may reduce or improve the effectiveness of testing and surgery. While many anesthetic agents have anticonvulsant properties, many also have varying profiles of proconvulsant or pharmacoactivating properties that can be useful in intraoperative localization of epileptogenic foci. Alternately, other agents may confound ECoG testing and lead to poor localization and less effective outcomes. Pharmacologic interactions between anticonvulsant medications and anesthetic drugs must also be taken into account. Pharmacoactivation of interictal epileptiform activities (IEAs) can be necessary in patients who do not demonstrate spontaneous interictal discharges during ECoG. The goals of the anesthetic regimen should be discussed with the neurosurgeon, neurologist, or neurophysiologist to determine if pharmacoactivation will be required. This may change during the procedure if the patient fails to generate IEAs spontaneously or under electrical stimulation. A goal-oriented anesthetic plan in concert with the neurosurgical team and knowledge of the activating properties of various anesthetic agents are essential.
Pharmacology of Anesthetic Agents
Proper sedation can be achieved through the use of a variety of anesthetic plans. In many cases, a general endotracheal anesthetic is preferred. In others, an awake craniotomy is performed for better functional testing and identification of seizure activity. Visualization of seizure activity that is similar to the patient’s typical seizures can be very helpful in identifying the true epileptogenic focus. Iatrogenic activation of IEAs may be achieved with administration of proconvulsant anesthetics and awareness of their anticonvulsant activities. EEG recordings support altering the activation and inhibition of the cerebral cortex with administration of anesthetic agents. For example, during light sedation, cortical activation with higher- frequency beta activity predominates, which progresses to slow-wave activity as sedative or anesthetic depth increases.
Sedative-Hypnotic Agents
As a group, sedative-hypnotic agents have the greatest variation and most confusing profile regarding effects on epileptogenic activity. Most agents can generate neuroexcitatory effects when used at low doses and neurodepressive effects when used at higher doses. Several induction agents, such as propofol and thiopental, can induce myoclonic movements not associated with EEG excitatory activity; whereas others, such as etomidate and methohexital, have been shown to generate both myoclonus and EEG-documented epileptiform activity in patients. , Motor stimulatory phenomena, such as myoclonus, opisthotonus, and tonic-clonic activity, may occur with varying frequency in both epileptic and nonepileptic patients during induction with these agents, but only a few agents actually produce cortical electrical activity suggestive of seizures.
Barbiturates and benzodiazepines have substantiated anti- convulsive properties and are recommended for treatment of refractory status epilepticus.
Propofol is among the most commonly used induction and maintenance agents in general anesthesia for epilepsy surgery and awake craniotomy. Propofol has been shown to depress ECoG recordings, decrease the frequency of spike activity, and produce a minimal effect on spontaneous IEAs. Propofol decreases the frequency of epileptogenic spikes and quiets existing seizure foci, particularly in the lateral and mesial temporal areas. One study demonstrated spike activation with low-dose propofol. There have been reported cases of tonic-clonic seizures with propofol, and myoclonic activity not related to excitatory EEG activity may be seen. Propofol may obscure spike wave activity for up to 20 minutes after termination of infusion and should be discontinued prior to ECoG testing.
Etomidate has been shown to activate EEG seizure activity at induction doses in patients with a history of epilepsy and may also generate myoclonic activity. It has been shown to have a high activation rate and demonstrates successful spike activation during intracranial electrode testing. At higher doses, etomidate may produce burst suppression and break status epilepticus. , To date, its use in intraoperative ECoG has not been studied.
Methohexital has been shown to activate EEG seizure activity in patients with epilepsy and may assist with activation of ictal foci during ECoG. It is associated with a high percentage of spike activation (50–85%), although with questionable specificity, showing up to 43% inappropriate activation in one study.
Dexmedetomidine may be a favorable agent for awake craniotomy due to its effects of sedation, analgesia, and anxiolysis; the absence of motor stimulatory effects; and the lack of respiratory depression. Dexmedetomidine does not affect background ECoG activity or IEAs and may be the best alternative for awake craniotomy. , ,
Ketamine may induce nonspecific activation of IEAs, especially in the limbic structure, and can activate seizure activity in patients with epilepsy. , It has been used to assist with activation of ictal foci during intraoperative ECoG. Ketamine appears to have a dose-dependent threshold for seizure generation, with most reported cases of clinical seizure activity occurring when doses larger than 4 mg/kg are administered. ,
Opioids
Synthetic opioids such as alfentanil, fentanyl, sufentanil, and remifentanil are commonly used in neurosurgical anesthesia because of their short duration of action and their ability to minimize cortical effects through continuous infusion. High doses of synthetic opioids have proepileptic properties. Standard maintenance doses of these agents do not significantly increase the risk of perioperative seizures or effects on ECoG. However, bolus doses of synthetic opioids, such as alfentanil and remifentanil, increase spike wave activity in the interictal foci of patients undergoing intraoperative ECoG. , Due to their high effectiveness and specificity, bolus doses of these agents are used to facilitate location of the ictal cortex through stimulation of spike wave phenomenon with concomitant depression of background EEG. Alfentanil has been shown to be the most effective and specific synthetic opioid for pharmacoactivation. Fentanyl has been associated with epileptiform electrical activity in subcortical nonictal cortical tissue and has been shown to be associated with contralateral activity. The clinical history of the use of synthetic opioids in large numbers of epileptic patients undergoing ablative procedures suggests that synthetic opioids can be used safely in this patient population without a significant increase in the risk of perioperative seizures. Morphine and hydromorphone used at clinically relevant doses do not appear to have significant proconvulsant activity.
Volatile Inhalational Agents and Nitrous Oxide
The epileptogenic potential of isoflurane, desflurane, and halothane appears low, and there have been no reported seizures when used in isolation. However, there are rare reports of myoclonic activity with a normal EEG. Convulsions with spike and wave activity on EEG have been reported with combinations of isoflurane and nitrous oxide (N 2 O). , Although N 2 O has been associated with seizure generation when used to supplement other agents, it appears to be fairly inert in both the development and the treatment of seizure activity in humans. Both N 2 O and isoflurane have been used for many years at multiple institutions with a good safety record in epileptic patients.
Enflurane, used with or without N 2 O, has been the most common offender, with reports of intraoperative and postoperative myoclonus and EEG-demonstrated epileptiform activity in both epileptic and nonepileptic patient populations. , , , , The incidence of EEG spike wave production with enflurane appears to be dose dependent. The end-tidal concentration that triggers maximum epileptiform activity is reduced during hypocapnia. Enflurane has fallen out of favor as new inhalational agents have become available, and it is now rarely used clinically in the United States. Enflurane should be avoided in patients with epilepsy unless the desired effect is to trigger seizures during ECoG.
Sevoflurane (not desflurane) has been reported to generate convulsions as well as electrical spike waves in both epileptic and nonepileptic patients. , The frequency of spike wave activity with sevoflurane increases with dose escalation and hyperventilation ( Fig. 17.2 ). , Hisada and colleagues reported that widespread neuroexcitatory activity associated with sevoflurane did not facilitate seizure focus localization in patients with temporal lobe epilepsy. Hyperventilation decreases the prediction specificity of leads with ictal spikes and should be employed cautiously during ECoG.
Muscle Relaxants
Long-term anticonvulsant therapy with phenytoin, carbamazepine, or both, is associated with resistance to the effect of nondepolarizing neuromuscular blockers, including pancuronium, vecuronium, metocurine, cisatracurium, and rocuronium, but less so with atracurium. , The etiology of this phenomenon is likely both pharmacodynamic and pharmacokinetic. ,
Anesthetic Management
Goals
Preoperative assessment of the patient’s neurologic condition, as well as comorbidities, is essential. Careful attention should be paid to anti-seizure medications. Intraoperative goals include maintenance of appropriate cerebral blood flood and perfusion, control of brain bulk, and rapid emergence from anesthesia for postoperative neurologic evaluation. In the event that seizure induction is desired, the goals of the anesthesiologist include selection of effective inducing agents and avoidance of patient injury. Careful postoperative monitoring of the patient’s neurologic status is required, and postoperative seizure control may be necessary.
Preoperative Evaluation
Neurologic History
The patient’s seizure history should be thoroughly understood prior to surgery. It may be difficult to discriminate seizure activity in the perioperative period from prolonged emergence or emergence delirium. Knowledge of the patient’s known seizure patterns may help to determine postoperative intervention. Prolonged emergence, characteristic motor activity, and poor responsiveness should raise suspicion for perioperative seizure activity.
The anesthesiologist should be vigilant for a number of medical conditions associated with epilepsy. Neurofibromatosis, also known as Von Recklinghausen’s disease, is an inherited condition that leads to tumor growth on nerve tissue. Variable expressivity means that the severity of this condition is wide ranging, from benign, asymptomatic tumors, to acoustic neuromas, significant intracranial lesions, and peripheral lesions. These tumors may involve cranial nerves or respiratory tract tumors leading to airway and respiratory compromise, including chronic aspiration, pulmonary fibrosing alveolitis, pulmonary hypertension, and cor pulmonale. Tuberous sclerosis is a disease causing widespread benign tumor growth in the brain, heart, lungs, kidneys, skin, and eyes. While it is less common than neurofibromatosis, tumors may lead to blockage of intraventricular cerebrospinal fluid (CSF) flow with hydrocephalus, cardiac dysrhythmias, intracardiac tumors, cerebral embolization, renal dysfunction, and arterial aneurysms. Intracardiac tumors, known as rhabdomyomas, are found in approximately 32.8–48% of tuberous sclerosis patients on echocardiography. , These patients should undergo a full preoperative cardiac evaluation. Down syndrome, Angelman syndrome, and Sturge–Weber syndrome are also associated with epileptiform activity. Open craniotomy is considered a moderate-risk procedure (indicating a less than 5% risk of cardiac events) with regard to its taxing effects on the cardiovascular system of the patient. Due to possible significant pneumocephalus up to 1 month after craniotomy, N 2 O should be avoided in patients who have undergone recent intracranial electrode placement.
Medication History
Medications for patients with epilepsy may present significant anesthetic considerations. Certain anticonvulsants significantly elevate dose requirements for both nondepolarizing muscle blockers and opioids. Both phenytoin and carbamazepine are associated with resistance to nondepolarizing neuromuscular blockade and elevated liver function parameters. The direct relationship between the number of anticonvulsants a patient receives and the dose of fentanyl required for intraoperative anesthetic maintenance further suggests that anticonvulsant therapy predisposes to opioid resistance. Elevated liver enzymes seen on liver function tests are commonly associated with anticonvulsant medications. Sedation and lethargy are common side effects of many antiepileptic agents, including newer agents such as lamotrigine and oxcarbazepine, and may potentiate the central nervous system- depressant effects of anesthetics. Chronic topiramate intake has been associated with intraoperative metabolic acidosis. Topiramate is associated with an asymptomatic non-anion-gap acidosis. Carbamazepine may cause a severe depression of the hematopoietic system and cardiac toxicity in rare cases. This drug’s metabolism is materially slowed by erythromycin and cimetidine, drugs that may be administered perioperatively. Likewise, a ketogenic diet, sometimes used as an adjunct anticonvulsant therapy, predisposes patients to metabolic acidosis. Valproic acid therapy results in dose-related thrombocytopenia and platelet dysfunction. However, additional bleeding risk during surgery is likely to be low in a patient taking valproic acid.
Patient Preparation
Regardless of the anesthetic approach selected, intraoperative awareness during ECoG is a possibility, due to reduced dosing of certain agents or the use of awake techniques. The patient should be reassured that this experience is usually described as a painless awareness. Careful explanation and reassurance to the patient and family of this and other risks, such as perioperative seizure, nausea, vomiting, and airway compromise, is essential. Neuropsychological impairment is commonly associated with epilepsy and psychiatric disorders, and impaired cognition is increased in this population. The anesthesiologist must be aware of these issues when selecting and preparing a patient for an awake technique.
Diagnostic Surgical Procedures for Intractable Epilepsy
Subdural grid electrodes may be placed for identification of epileptogenic foci in preparation for resection. A craniotomy is performed and the grid electrodes are placed under general anesthesia. Usual anesthetic concerns for craniotomy should be observed. Hyperventilation to relax the brain during exposure may be efficacious, but should be considered carefully against the risk of precipitating seizure activity in the epileptic patient. Hyperventilation may be less effective in patients with complex partial seizures, who may have lower CO 2 reactivity of cerebral blood flow than normal patients. Arterial line placement for blood gases and accurate blood pressure monitoring as well as adequate IV access are indicated. Since intraoperative testing is not performed, anesthetic techniques may be used without regard to their effect on EEG. As always, rapid emergence from anesthesia for neurologic assessment is preferred.
Placement of epidural (“peg”) electrodes may be used to include recording from deeper structures. It requires multiple burr holes and can be a lengthy procedure, depending on the number of electrodes to be placed. “Depth” electrodes for exploring subcortical regions of the brain require stereotactic placement. The procedure usually is uneventful and not associated with significant bleeding. A general anesthetic is most frequently used. Unless further monitoring is indicated for a medical comorbidity, only routine noninvasive monitoring is employed.
Resection of Epileptogenic Brain Regions under General Anesthesia
Anesthetic planning for epileptogenic brain resection procedures depends greatly on the need for intraoperative brain mapping for seizure foci localization. In some cases, resection of epileptogenic foci is performed without brain mapping under general anesthesia. In such cases, the anesthetic goals are much like those of most open craniotomy procedures. If EEG is not planned, benzodiazepines may be given preoperatively for patient comfort. Monitoring should include direct arterial blood pressure monitoring and IV access should be adequate to replace rapid blood loss from dural sinuses. Brain relaxation is desirable to facilitate surgical exposure and resection. Maintenance of adequate cerebral perfusion without brain engorgement is an essential feature. As always, immobility is critical to the safety of the patient, as is adequate anesthesia to avoid patient awareness and pain. The anesthesiologist should always be prepared to control intraoperative seizures. Neurological evaluation in the immediate postoperative period is highly desirable. Therefore, the plan should consider anesthetic management that will allow for rapid emergence. This may include the use of the ultra-short-acting narcotic remifentanil, which allows for rapid emergence and early neurologic examination when compared to other opioids. However, addition of a longer acting opioid in the immediate postoperative period will be required. TIVA with propofol and remifentanil may be considered. Propofol’s property of better brain relaxation than isoflurane or sevoflurane at greater than half-monitored anesthesia care (MAC) in patients with mass lesions suggests its efficacy in craniotomies. However, these benefits may be less clinically significant when lower MAC doses of such volatile agents are used. Prospective studies have not been sufficiently powered to allow determination of the impact of anesthetic technique on neurologic and functional outcome after craniotomy as of this time. Antihistamines can activate seizure foci in patients with epilepsy and should be avoided as premedicants.
By contrast, when intraoperative brain mapping is anticipated, additional anesthetic goals and planning need to be considered. As described above, many anesthetic agents may promote or suppress epileptiform activity. The anesthesiologist must take care that medications administered to the patient will not interfere with intraoperative monitoring and the mapping of ictal foci. Likewise, it may be desirable in some instances to administer agents that will promote epileptiform discharges and improve mapping.
Barbiturate and benzodiazepine premedication should be avoided because it may elevate the seizure threshold, making ECoG recording of epileptogenic activity more difficult. An intubation dose of short-acting barbiturate during anesthesia induction is not contraindicated, but barbiturates should be avoided later in the procedure, as should intravenous lidocaine. Despite an isolated report of N 2 O-related diminution of epileptic foci during intraoperative ECoG, N 2 O can be used for these procedures. Ebrahim and colleagues recommended that propofol administration be stopped 20–30 minutes prior to ECoG, because it elicits high-frequency beta EEG activity ( Fig. 17.3 ) for as long as 30 minutes after discontinuation, although other investigators have reported that this type of EEG activity did not prevent ECoG interpretation. The use of low concentrations of isoflurane or desflurane is permissible when ECoG recording is planned, provided that these agents can be eliminated well before the start of corticography. Isoflurane may decrease the frequency and spatial distribution of epileptogenic spikes, although it is unclear whether this effect persists at low concentrations. Low-dose sevoflurane would be preferred, given its mild proconvulsant properties and short duration of action. When no potent inhaled anesthetics are in use, scopolamine, droperidol, and increased opioid dosing can be substituted to prevent intraoperative recall with virtually no effect on the EEG. Mild-to-moderate hypocapnia (PaCO 2 30–35 mmHg), however, is often necessary to assist in brain volume control and brain relaxation. If hyperventilation must be initiated during sevoflurane anesthesia, the anesthesiologist should be aware that the specificity of ictal lead prediction may diminish.
If cortical motor area stimulation is necessary for the surgeon to accomplish safe resection, particular attention must be paid to the management and dosage of neuromuscular blocking agents. As a general rule, neuromuscular blockade should be minimal to allow motor stimulation. If moderate residual neuromuscular block persists, a small dose of anticholinesterase can be administered to achieve its complete reversal.
Cortical stimulation for localization as well as light anesthesia and brain manipulation may lead to intraoperative seizures. Treatment of seizure activity during ongoing intraoperative ECoG, therefore, requires the use of short-acting anticonvulsants (such as methohexital) as one weighs the therapeutic goals of gross seizure control against the potential for interference with critical electrocortical monitoring. An alternative and highly effective means of suppressing seizures is irrigation of the cortical surface with cold saline.
When intraoperative EEG recordings fail to reveal seizure spikes, and in consultation with the surgeon and the electroencephalographer, the anesthesiologist administers anesthetics known to promote epileptiform discharges. These include methohexital (25–50 mg), , alfentanil (20 μg/kg), , and etomidate (0.2 mg/kg), all of which may be administered to help activate dormant foci. Alfentanil is the most effective of these agents, provoking abnormal EEG spike activity in 83% of patients, compared with 50% for methohexital. However, controversy exists over the correlation of pharmacologically elicited seizure spikes with the patients’ native epileptogenic foci. Severe bradycardia has been reported during amygdala- hippocampectomy that is not seen during routine anterior temporal lobe resection. This problem is thought to be the result of surgical limbic system stimulation resulting in enhanced neural vagal activity. ,
Cerebral Hemispherectomy
On occasion, the seizure foci are so diffuse as to require resection of substantial portions of an entire cerebral hemisphere. Frequently, this procedure is performed in children and can be associated with significant morbidity and mortality related to massive blood loss, electrolyte and metabolic disturbances, coagulopathy, cerebral hemorrhage, and seizures. Hemispherectomy requires a very large craniectomy, which increases the chance of bleeding and tearing of dural sinuses. Air embolism has also been reported and may lead to serious morbidity. Kofke and associates compared three different surgical techniques (anatomical, functional, and lateral) for hemispherectomy. Lateral hemispherectomy was associated with the lowest intraoperative blood loss, the shortest intensive care stay, and the lowest complication rate. Functional hemispherectomy had the highest rate of reoperation, whereas patients undergoing anatomical hemispherectomy had the longest hospital stays, greatest requirement for CSF diversion, and highest postoperative fever. Patients with cortical dysplasia had the largest intraoperative blood loss.
Continuous monitoring of blood pressure by arterial catheter is required, as is central venous access and monitoring of cardiac filling pressure. In addition, pressor and inotropic infusions should be readily available to combat low cardiac output states. Brian et al. report a series of 10 patients, aged 3 months to 12 years, whose intraoperative blood replacement amounted to 1.5 blood volumes on average. In seven patients, a coagulopathy developed intraoperatively and required administration of platelets, fresh frozen plasma, or both. Progressive hypokalemia requiring replacement occurred in four patients. Hypothermia and metabolic acidosis was observed in five patients. Urine output was a poor indicator of volume status because of frequent massive glycosuria. Zuckerberg and colleagues report several children younger than 5 years in whom severe decreases in cardiac index, bradycardia, increased systemic vascular resistance, and an alveolar- to-arterial gradient suggestive of neurogenic pulmonary edema developed after hemispherectomy with extensive subcortical resection. Removal of the endotracheal tube at the conclusion of procedures with large-volume resuscitation and with a high potential for postoperative complications would, therefore, seem unwise. Postoperative hemodynamic instability is common, and the airway may be compromised by seizure activity. Early postoperative recovery is best accomplished in an intensive care environment. As has been reported in adults, children undergoing major brain resection become hypercoagulable as early as during dural closure. Although the clinical significance of this finding is debated, thrombotic complications should be anticipated.
Vagal Nerve Stimulator Placement
Vagal nerve stimulation is a nonpharmacological intervention for patients with refractory epilepsy. A device based on cardiac pacemakers, the vagal nerve stimulator (VNS) emits electrical pulses from a generator, through an implanted wire, to an electrode wrapped around the left vagus nerve to modulate cerebral neuronal excitativity. It has been demonstrated to reduce seizure frequency. Proposed mechanisms of action include activation of the limbic system, locus ceruleus, and amygdala.
The VNS is placed on the left side to avoid the vagal fibers that affect the sinoatrial node associated with the right vagus nerve and to reduce the likelihood of clinically significant bradycardia. The patient is positioned supine with the head turned to the right. The left vagus nerve is exposed, taking care not to injure the left carotid and jugular that flank the nerve within the carotid sheath. The generator pocket is created above the left pectoralis muscle. A tunnel is created, and the connecting wire from the generator is advanced through the tunnel to the nerve. The electrode array is attached to the nerve. The generator is connected, and the unit is tested before it is sewn into the pocket.
VNS placement is usually performed under general endotracheal anesthesia. Standard American Society of Anesthesiologists (ASA) induction monitors are used. Additional monitoring should be based on the patient’s comorbid status. Perioperative complications may include seizures; bradycardia; vocal cord paralysis or hoarseness from recurrent and superior laryngeal nerve injury or activation; and hematoma. Unilateral vocal cord paralysis has been reported as well, as has a predisposition to chronic pulmonary aspiration. Complete atrioventricular block and ventricular asystole have also been reported. If cardiac dysrhythmias occur, stimulation of the vagal nerve should be stopped immediately and additional rescue measures may be necessary.
Anesthetic management with regard to the patient’s seizure disorder is the same as has been described for other procedures under general anesthesia without mapping. It is recommended that patients take their seizure medications as scheduled prior to the procedure, and the anesthesiologist should be aware of interactions and effects with regard to anesthetic agents. Management of intraoperative and postoperative seizures may be necessary.
Emergence from Anesthesia and Postoperative Management
As with most intracranial procedures, rapid emergence from anesthesia is helpful for postoperative neurologic assessment. However, patients with seizure disorders and a long history of anticonvulsant use may experience lethargy and slower emergence from anesthesia. Intraoperative loading of phenytoin for treatment of seizures may increase the risk of delayed emergence from general anesthesia. Coughing and bucking on emergence are undesirable as they may increase the risk of intracranial bleeding and CSF leak. Hypertension should be avoided. If short-acting narcotic agents were used, postoperative pain control with longer acting agents may be necessary. The anesthesiologist should remain vigilant for changes in the patient’s mental status in the recovery phase that may indicate seizure activity, bleeding, or hematoma formation. Minor postoperative complications occur in 5.1–10.9% of patients undergoing surgical procedures for epilepsy. CSF leak was the most common minor complication. Neurologic complications involving speech, visual, motor, and memory deficits may occur. Cerebral edema may occur in patients with temporary subdural grid electrode implants. Nausea and vomiting occur in 38% of intracranial neurosurgical cases. Prophylactic administration of antinauseants is effective and advisable.
If patients have tapered or discontinued anticonvulsant medications prior to surgery, the anesthesiologist must be especially vigilant for postoperative seizures. Benzodiazepines and propofol may be administered to control seizure activity, and the airway may need to be secured. Administration of anticonvulsants such as phenytoin also may be necessary and is begun at 50 mg/min to a total dose of 20 mg/kg, assuming the patient has not previously been treated with phenytoin.
Epilepsy surgery
Epilepsy is a disease of the brain characterized by: two unprovoked seizures greater than 24 hours apart; one unprovoked seizure; and a probability of seizures similar to the general recurrence risk after two unprovoked seizures occurring over the next 10 years, or diagnosis of an epilepsy syndrome. It is present in 0.5–2.2% of the general population. Because 30–40% of epileptics do not respond adequately to pharmacologic intervention, more than 400,000 people still have medically uncontrolled epilepsy in the United States. However, only 10–30% of patients with seizures refractory to medical management are appropriate candidates for seizure surgery, and only 1% eventually undergo the procedure.
Epilepsy is classified as partial, generalized, or psychogenic nonepileptiform seizures (PNES). Partial seizures are characterized by electrical disturbances localized to one area of one cerebral hemisphere. Simple partial seizures are not associated with a loss of consciousness, and generally last 1 minute or less. Complex partial seizures are characterized by a loss of consciousness or awareness and spread from their localized focus to other regions. Complex partial seizures may spread to become generalized. Generalized seizures have no demonstrated focal onset, although they may evolve from focal seizures, affect both hemispheres of the brain, and are characterized by a loss of consciousness. They are sub-categorized as generalized tonic-clonic (grand mal), tonic, myoclonic, absence (petit mal), and atonic. PNES are psychogenic episodes that may be characterized by seizure-like physical manifestations but have no corresponding epileptiform activity on EEG and are considered conversion reactions.
Surgical management of epilepsy may be an option for patients with intractable epilepsy refractory to medical treatment. With successful surgical intervention, lifestyle improves, although most patients continue anticonvulsant therapy. Chin et al. reported that the rate of employment improved only modestly in their group of 375 patients, from 39.5% fully employed status preoperatively to 42.8% postoperatively; however, the rate of part-time employment nearly doubled, from 6.9 to 12.4%.
Anesthetic regimens have a significant effect on cortical mapping for epilepsy and may reduce or improve the effectiveness of testing and surgery. While many anesthetic agents have anticonvulsant properties, many also have varying profiles of proconvulsant or pharmacoactivating properties that can be useful in intraoperative localization of epileptogenic foci. Alternately, other agents may confound ECoG testing and lead to poor localization and less effective outcomes. Pharmacologic interactions between anticonvulsant medications and anesthetic drugs must also be taken into account. Pharmacoactivation of interictal epileptiform activities (IEAs) can be necessary in patients who do not demonstrate spontaneous interictal discharges during ECoG. The goals of the anesthetic regimen should be discussed with the neurosurgeon, neurologist, or neurophysiologist to determine if pharmacoactivation will be required. This may change during the procedure if the patient fails to generate IEAs spontaneously or under electrical stimulation. A goal-oriented anesthetic plan in concert with the neurosurgical team and knowledge of the activating properties of various anesthetic agents are essential.
Pharmacology of Anesthetic Agents
Proper sedation can be achieved through the use of a variety of anesthetic plans. In many cases, a general endotracheal anesthetic is preferred. In others, an awake craniotomy is performed for better functional testing and identification of seizure activity. Visualization of seizure activity that is similar to the patient’s typical seizures can be very helpful in identifying the true epileptogenic focus. Iatrogenic activation of IEAs may be achieved with administration of proconvulsant anesthetics and awareness of their anticonvulsant activities. EEG recordings support altering the activation and inhibition of the cerebral cortex with administration of anesthetic agents. For example, during light sedation, cortical activation with higher- frequency beta activity predominates, which progresses to slow-wave activity as sedative or anesthetic depth increases.
Sedative-Hypnotic Agents
As a group, sedative-hypnotic agents have the greatest variation and most confusing profile regarding effects on epileptogenic activity. Most agents can generate neuroexcitatory effects when used at low doses and neurodepressive effects when used at higher doses. Several induction agents, such as propofol and thiopental, can induce myoclonic movements not associated with EEG excitatory activity; whereas others, such as etomidate and methohexital, have been shown to generate both myoclonus and EEG-documented epileptiform activity in patients. , Motor stimulatory phenomena, such as myoclonus, opisthotonus, and tonic-clonic activity, may occur with varying frequency in both epileptic and nonepileptic patients during induction with these agents, but only a few agents actually produce cortical electrical activity suggestive of seizures.
Barbiturates and benzodiazepines have substantiated anti- convulsive properties and are recommended for treatment of refractory status epilepticus.
Propofol is among the most commonly used induction and maintenance agents in general anesthesia for epilepsy surgery and awake craniotomy. Propofol has been shown to depress ECoG recordings, decrease the frequency of spike activity, and produce a minimal effect on spontaneous IEAs. Propofol decreases the frequency of epileptogenic spikes and quiets existing seizure foci, particularly in the lateral and mesial temporal areas. One study demonstrated spike activation with low-dose propofol. There have been reported cases of tonic-clonic seizures with propofol, and myoclonic activity not related to excitatory EEG activity may be seen. Propofol may obscure spike wave activity for up to 20 minutes after termination of infusion and should be discontinued prior to ECoG testing.
Etomidate has been shown to activate EEG seizure activity at induction doses in patients with a history of epilepsy and may also generate myoclonic activity. It has been shown to have a high activation rate and demonstrates successful spike activation during intracranial electrode testing. At higher doses, etomidate may produce burst suppression and break status epilepticus. , To date, its use in intraoperative ECoG has not been studied.
Methohexital has been shown to activate EEG seizure activity in patients with epilepsy and may assist with activation of ictal foci during ECoG. It is associated with a high percentage of spike activation (50–85%), although with questionable specificity, showing up to 43% inappropriate activation in one study.
Dexmedetomidine may be a favorable agent for awake craniotomy due to its effects of sedation, analgesia, and anxiolysis; the absence of motor stimulatory effects; and the lack of respiratory depression. Dexmedetomidine does not affect background ECoG activity or IEAs and may be the best alternative for awake craniotomy. , ,
Ketamine may induce nonspecific activation of IEAs, especially in the limbic structure, and can activate seizure activity in patients with epilepsy. , It has been used to assist with activation of ictal foci during intraoperative ECoG. Ketamine appears to have a dose-dependent threshold for seizure generation, with most reported cases of clinical seizure activity occurring when doses larger than 4 mg/kg are administered. ,
Opioids
Synthetic opioids such as alfentanil, fentanyl, sufentanil, and remifentanil are commonly used in neurosurgical anesthesia because of their short duration of action and their ability to minimize cortical effects through continuous infusion. High doses of synthetic opioids have proepileptic properties. Standard maintenance doses of these agents do not significantly increase the risk of perioperative seizures or effects on ECoG. However, bolus doses of synthetic opioids, such as alfentanil and remifentanil, increase spike wave activity in the interictal foci of patients undergoing intraoperative ECoG. , Due to their high effectiveness and specificity, bolus doses of these agents are used to facilitate location of the ictal cortex through stimulation of spike wave phenomenon with concomitant depression of background EEG. Alfentanil has been shown to be the most effective and specific synthetic opioid for pharmacoactivation. Fentanyl has been associated with epileptiform electrical activity in subcortical nonictal cortical tissue and has been shown to be associated with contralateral activity. The clinical history of the use of synthetic opioids in large numbers of epileptic patients undergoing ablative procedures suggests that synthetic opioids can be used safely in this patient population without a significant increase in the risk of perioperative seizures. Morphine and hydromorphone used at clinically relevant doses do not appear to have significant proconvulsant activity.
Volatile Inhalational Agents and Nitrous Oxide
The epileptogenic potential of isoflurane, desflurane, and halothane appears low, and there have been no reported seizures when used in isolation. However, there are rare reports of myoclonic activity with a normal EEG. Convulsions with spike and wave activity on EEG have been reported with combinations of isoflurane and nitrous oxide (N 2 O). , Although N 2 O has been associated with seizure generation when used to supplement other agents, it appears to be fairly inert in both the development and the treatment of seizure activity in humans. Both N 2 O and isoflurane have been used for many years at multiple institutions with a good safety record in epileptic patients.
Enflurane, used with or without N 2 O, has been the most common offender, with reports of intraoperative and postoperative myoclonus and EEG-demonstrated epileptiform activity in both epileptic and nonepileptic patient populations. , , , , The incidence of EEG spike wave production with enflurane appears to be dose dependent. The end-tidal concentration that triggers maximum epileptiform activity is reduced during hypocapnia. Enflurane has fallen out of favor as new inhalational agents have become available, and it is now rarely used clinically in the United States. Enflurane should be avoided in patients with epilepsy unless the desired effect is to trigger seizures during ECoG.
Sevoflurane (not desflurane) has been reported to generate convulsions as well as electrical spike waves in both epileptic and nonepileptic patients. , The frequency of spike wave activity with sevoflurane increases with dose escalation and hyperventilation ( Fig. 17.2 ). , Hisada and colleagues reported that widespread neuroexcitatory activity associated with sevoflurane did not facilitate seizure focus localization in patients with temporal lobe epilepsy. Hyperventilation decreases the prediction specificity of leads with ictal spikes and should be employed cautiously during ECoG.
Muscle Relaxants
Long-term anticonvulsant therapy with phenytoin, carbamazepine, or both, is associated with resistance to the effect of nondepolarizing neuromuscular blockers, including pancuronium, vecuronium, metocurine, cisatracurium, and rocuronium, but less so with atracurium. , The etiology of this phenomenon is likely both pharmacodynamic and pharmacokinetic. ,
Anesthetic Management
Goals
Preoperative assessment of the patient’s neurologic condition, as well as comorbidities, is essential. Careful attention should be paid to anti-seizure medications. Intraoperative goals include maintenance of appropriate cerebral blood flood and perfusion, control of brain bulk, and rapid emergence from anesthesia for postoperative neurologic evaluation. In the event that seizure induction is desired, the goals of the anesthesiologist include selection of effective inducing agents and avoidance of patient injury. Careful postoperative monitoring of the patient’s neurologic status is required, and postoperative seizure control may be necessary.
Preoperative Evaluation
Neurologic History
The patient’s seizure history should be thoroughly understood prior to surgery. It may be difficult to discriminate seizure activity in the perioperative period from prolonged emergence or emergence delirium. Knowledge of the patient’s known seizure patterns may help to determine postoperative intervention. Prolonged emergence, characteristic motor activity, and poor responsiveness should raise suspicion for perioperative seizure activity.
The anesthesiologist should be vigilant for a number of medical conditions associated with epilepsy. Neurofibromatosis, also known as Von Recklinghausen’s disease, is an inherited condition that leads to tumor growth on nerve tissue. Variable expressivity means that the severity of this condition is wide ranging, from benign, asymptomatic tumors, to acoustic neuromas, significant intracranial lesions, and peripheral lesions. These tumors may involve cranial nerves or respiratory tract tumors leading to airway and respiratory compromise, including chronic aspiration, pulmonary fibrosing alveolitis, pulmonary hypertension, and cor pulmonale. Tuberous sclerosis is a disease causing widespread benign tumor growth in the brain, heart, lungs, kidneys, skin, and eyes. While it is less common than neurofibromatosis, tumors may lead to blockage of intraventricular cerebrospinal fluid (CSF) flow with hydrocephalus, cardiac dysrhythmias, intracardiac tumors, cerebral embolization, renal dysfunction, and arterial aneurysms. Intracardiac tumors, known as rhabdomyomas, are found in approximately 32.8–48% of tuberous sclerosis patients on echocardiography. , These patients should undergo a full preoperative cardiac evaluation. Down syndrome, Angelman syndrome, and Sturge–Weber syndrome are also associated with epileptiform activity. Open craniotomy is considered a moderate-risk procedure (indicating a less than 5% risk of cardiac events) with regard to its taxing effects on the cardiovascular system of the patient. Due to possible significant pneumocephalus up to 1 month after craniotomy, N 2 O should be avoided in patients who have undergone recent intracranial electrode placement.
Medication History
Medications for patients with epilepsy may present significant anesthetic considerations. Certain anticonvulsants significantly elevate dose requirements for both nondepolarizing muscle blockers and opioids. Both phenytoin and carbamazepine are associated with resistance to nondepolarizing neuromuscular blockade and elevated liver function parameters. The direct relationship between the number of anticonvulsants a patient receives and the dose of fentanyl required for intraoperative anesthetic maintenance further suggests that anticonvulsant therapy predisposes to opioid resistance. Elevated liver enzymes seen on liver function tests are commonly associated with anticonvulsant medications. Sedation and lethargy are common side effects of many antiepileptic agents, including newer agents such as lamotrigine and oxcarbazepine, and may potentiate the central nervous system- depressant effects of anesthetics. Chronic topiramate intake has been associated with intraoperative metabolic acidosis. Topiramate is associated with an asymptomatic non-anion-gap acidosis. Carbamazepine may cause a severe depression of the hematopoietic system and cardiac toxicity in rare cases. This drug’s metabolism is materially slowed by erythromycin and cimetidine, drugs that may be administered perioperatively. Likewise, a ketogenic diet, sometimes used as an adjunct anticonvulsant therapy, predisposes patients to metabolic acidosis. Valproic acid therapy results in dose-related thrombocytopenia and platelet dysfunction. However, additional bleeding risk during surgery is likely to be low in a patient taking valproic acid.
Patient Preparation
Regardless of the anesthetic approach selected, intraoperative awareness during ECoG is a possibility, due to reduced dosing of certain agents or the use of awake techniques. The patient should be reassured that this experience is usually described as a painless awareness. Careful explanation and reassurance to the patient and family of this and other risks, such as perioperative seizure, nausea, vomiting, and airway compromise, is essential. Neuropsychological impairment is commonly associated with epilepsy and psychiatric disorders, and impaired cognition is increased in this population. The anesthesiologist must be aware of these issues when selecting and preparing a patient for an awake technique.
Diagnostic Surgical Procedures for Intractable Epilepsy
Subdural grid electrodes may be placed for identification of epileptogenic foci in preparation for resection. A craniotomy is performed and the grid electrodes are placed under general anesthesia. Usual anesthetic concerns for craniotomy should be observed. Hyperventilation to relax the brain during exposure may be efficacious, but should be considered carefully against the risk of precipitating seizure activity in the epileptic patient. Hyperventilation may be less effective in patients with complex partial seizures, who may have lower CO 2 reactivity of cerebral blood flow than normal patients. Arterial line placement for blood gases and accurate blood pressure monitoring as well as adequate IV access are indicated. Since intraoperative testing is not performed, anesthetic techniques may be used without regard to their effect on EEG. As always, rapid emergence from anesthesia for neurologic assessment is preferred.
Placement of epidural (“peg”) electrodes may be used to include recording from deeper structures. It requires multiple burr holes and can be a lengthy procedure, depending on the number of electrodes to be placed. “Depth” electrodes for exploring subcortical regions of the brain require stereotactic placement. The procedure usually is uneventful and not associated with significant bleeding. A general anesthetic is most frequently used. Unless further monitoring is indicated for a medical comorbidity, only routine noninvasive monitoring is employed.
Resection of Epileptogenic Brain Regions under General Anesthesia
Anesthetic planning for epileptogenic brain resection procedures depends greatly on the need for intraoperative brain mapping for seizure foci localization. In some cases, resection of epileptogenic foci is performed without brain mapping under general anesthesia. In such cases, the anesthetic goals are much like those of most open craniotomy procedures. If EEG is not planned, benzodiazepines may be given preoperatively for patient comfort. Monitoring should include direct arterial blood pressure monitoring and IV access should be adequate to replace rapid blood loss from dural sinuses. Brain relaxation is desirable to facilitate surgical exposure and resection. Maintenance of adequate cerebral perfusion without brain engorgement is an essential feature. As always, immobility is critical to the safety of the patient, as is adequate anesthesia to avoid patient awareness and pain. The anesthesiologist should always be prepared to control intraoperative seizures. Neurological evaluation in the immediate postoperative period is highly desirable. Therefore, the plan should consider anesthetic management that will allow for rapid emergence. This may include the use of the ultra-short-acting narcotic remifentanil, which allows for rapid emergence and early neurologic examination when compared to other opioids. However, addition of a longer acting opioid in the immediate postoperative period will be required. TIVA with propofol and remifentanil may be considered. Propofol’s property of better brain relaxation than isoflurane or sevoflurane at greater than half-monitored anesthesia care (MAC) in patients with mass lesions suggests its efficacy in craniotomies. However, these benefits may be less clinically significant when lower MAC doses of such volatile agents are used. Prospective studies have not been sufficiently powered to allow determination of the impact of anesthetic technique on neurologic and functional outcome after craniotomy as of this time. Antihistamines can activate seizure foci in patients with epilepsy and should be avoided as premedicants.
By contrast, when intraoperative brain mapping is anticipated, additional anesthetic goals and planning need to be considered. As described above, many anesthetic agents may promote or suppress epileptiform activity. The anesthesiologist must take care that medications administered to the patient will not interfere with intraoperative monitoring and the mapping of ictal foci. Likewise, it may be desirable in some instances to administer agents that will promote epileptiform discharges and improve mapping.
Barbiturate and benzodiazepine premedication should be avoided because it may elevate the seizure threshold, making ECoG recording of epileptogenic activity more difficult. An intubation dose of short-acting barbiturate during anesthesia induction is not contraindicated, but barbiturates should be avoided later in the procedure, as should intravenous lidocaine. Despite an isolated report of N 2 O-related diminution of epileptic foci during intraoperative ECoG, N 2 O can be used for these procedures. Ebrahim and colleagues recommended that propofol administration be stopped 20–30 minutes prior to ECoG, because it elicits high-frequency beta EEG activity ( Fig. 17.3 ) for as long as 30 minutes after discontinuation, although other investigators have reported that this type of EEG activity did not prevent ECoG interpretation. The use of low concentrations of isoflurane or desflurane is permissible when ECoG recording is planned, provided that these agents can be eliminated well before the start of corticography. Isoflurane may decrease the frequency and spatial distribution of epileptogenic spikes, although it is unclear whether this effect persists at low concentrations. Low-dose sevoflurane would be preferred, given its mild proconvulsant properties and short duration of action. When no potent inhaled anesthetics are in use, scopolamine, droperidol, and increased opioid dosing can be substituted to prevent intraoperative recall with virtually no effect on the EEG. Mild-to-moderate hypocapnia (PaCO 2 30–35 mmHg), however, is often necessary to assist in brain volume control and brain relaxation. If hyperventilation must be initiated during sevoflurane anesthesia, the anesthesiologist should be aware that the specificity of ictal lead prediction may diminish.
