22
CHAPTER
Wada Test
Adriana Palade
Patients with medically refractory epilepsy commonly undergo a Wada test (or intracarotid amobarbital procedure) as part of the presurgical workup for anterior temporal lobectomy. The Wada test is used to predict whether patients are at risk for postoperative amnesia or memory decline, to assess the risk of material specific memory deficits, and to lateralize memory dysfunction.
HISTORY
In 1949, Juhn Wada described the use of intracarotid amobarbital injection to determine the cerebral language dominance (1). This procedure was later modified at the Montreal Neurological Institute to also assess hemispheric memory and has since then become a standard component of presurgical evaluation for epilepsy surgery (2).
The cognitive outcome of epilepsy surgery or even tumor resection in the dominant hemisphere depends on the accurate localization and lateralization of the “eloquent” cortex, especially the language area. In addition, the Wada test has been used to predict residual motor functions before hemispherectomy.
However, during the past two decades, there has been a significant shift in the role of the Wada test in epilepsy surgery, such that the test is no longer performed as a routine procedure. Alternative noninvasive methods, most commonly functional magnetic resonance imaging (fMRI) and the evaluation of the available clinical data (seizure semiology, improved MRI quality, neuropsychological data) may provide enough information for language localization or memory reserve capacity determination.
PROCEDURE
The Wada test is performed in the radiology suite. Selective internal carotid catheterization is achieved via the femoral artery route. The Wada test is an invasive procedure accompanied by a significant rate of complications in up to 10% of the cases, ranging from encephalopathy (7%), seizures (1.2%), strokes (0.6%), transient ischemic attack (0.6%), and localized hemorrhage at the site of catheter insertion (3). The intracranial arterial anatomy may vary between patients, with persistent fetal circulation in some, hypoplastic anterior cerebral or vertebral arteries in others, and different amounts of blood flow to the contralateral hemisphere from the internal carotid artery (ICA). Most importantly, the posterior medial temporal lobe is supplied by the posterior circulation to a variable extent, so temporal lobe inactivation may be incomplete.
The dosage of amobarbital injected varies between centers from 75 mg to 150 mg, injected within a short period, typically of 3 to 5 seconds, with a single-bolus technique. The dose should be sufficient to produce anesthesia of the injected hemisphere, which is tested by assessing motor and speech performance during the injection. In preparation for the test, a patient is connected to an EEG machine using the standard 10-20 system of electrode placement. Typical stimulation techniques of eye opening and closure are performed, but none of the activating procedures is necessary. Before the injection, the patient lays on his/her back, raises his/her arms, and begins to count. The patient is instructed to continue counting until told verbally or with nonverbal maneuvers of manual pressure on the outstretched arms to stop counting and put the arms down. Within seconds of injection, the contralateral arm drops as hemiparesis/hemiplegia develops, indicative of an effective injection. When the dominant hemisphere is anesthetized, the patient develops a global aphasia and is mute for a period of 3 to 4 minutes, after which the speech gradually returns, with dysphasic and paraphasic errors seen first.
When the nondominant hemisphere is injected, the patient may continue to count or speak but usually exhibits dysarthria lasting several minutes. Occasionally confusion, inattention, perseveration, or agitation accompanies the hemiparesis and speech impairment, particularly when the language-dominant hemisphere is injected (4). Before administering the test items, especially if the patient is mute, several commands must be followed, such as opening/closing the eyes, raising the hand, or being able to look at the examiner. Over 18 items are presented during the test, which consist of objects, abstract designs, written words, and phrases to repeat. The patient is constantly asked to recognize each item that is presented and to remember it, allowing the examiner to assess the language function in addition to the memory. Intermittently throughout the test assessment of the strength of the extremity contralateral to the injection site is performed. Within 15 to 20 minutes, all testing items are presented, and simultaneously the contralateral hemiparesis usually resolves and ipsilateral slowing on EEG usually resolves.
Most epilepsy centers inject the site of the epileptic focus first and perform injections of both internal carotid arteries during the same day. It is common practice to record the EEG simultaneously with the Wada test to evaluate the extent of the physiologic effect of amobarbital, consisting of unilateral slowing. The EEG recording is extremely helpful if the patient becomes severely obtunded after injection, which is associated at times on EEG with bilateral slowing. This raises the suspicion of arterial crossover of the medication as an explanation of the obtundation. The clinical and physiologic (EEG) changes induced by amobarbital must resolve before testing of memory function.
LANGUAGE EVALUATION
Validity
Cerebral hemisphere dominance for language function was first documented by Broca, who described a patient with a lesion in the left inferior frontal gyrus who experienced expressive aphasia (5). The relationship between the language lateralization and handedness was also established during the same time period. It is well accepted that the left hemisphere is dominant for speech, not only for the majority of right-handed individuals, but also for the majority of left-handed ones and those who are ambidextrous. Nonetheless, the right hemisphere has been demonstrated to be dominant for language in some right-handed individuals who have crossed aphasia, explaining how lesions in the right hemisphere may result in aphasia (6). The right hemisphere is estimated to be dominant for language in 1% to 2% of the normal population.
The Wada test has commonly been described as the gold standard for the determination of cerebral hemisphere dominance for language against which novel language lateralization techniques are judged. However, in many patients, a combination of fMRI studies, ictal semiology, postictal recovery, and baseline neuropsychological measures is able to indicate language dominance with a high degree of accuracy (7,8).
Multiple studies have evaluated the relationship of handedness and language dominance. In an overwhelming majority of right-handed individuals (80%–100%), the left hemisphere is seen to be dominant for speech. There are a few right-handed individuals who are right hemisphere dominant for speech, with a range of 4% to 10%. However, the presence of a hemispheric lesion is an important influencing factor in language lateralization. If a left hemisphere lesion is present, the percentage of right-hemisphere-dominant subjects and those with bilateral speech representation increases to 12% and 7%, respectively (9). In the absence of a left hemisphere lesion, patients with left or mixed handedness appear to have left hemisphere dominance for speech in 64% to 70% of cases, right hemisphere language dominance in 15% to 20% of cases, and bilateral language representation in the remaining 15%. In the presence of a left hemisphere lesion, the right hemisphere is dominant for speech in 53% and 67% in the two series reported by Rassmussen (9). Bilateral representation was also found in up to 20% of left-handed individuals.
Patients with epilepsy have a much higher rate of altered language lateralization: 22% versus 6% altered lateralization seen in normal controls (10). These patients often have bihemispheric participation in language and increased variability in language dominance. Left-handed patients are most likely to have atypical language representation. Atypical language lateralization is more common in patients who have structural or functional extrahippocampal involvement, an early onset of epilepsy, usually before the age of 6 years (11), a short time period between the initial precipitating injury and onset of habitual seizures, and bitemporal and extratemporal interictal discharges on EEG (12).
Validation
The Wada test has been well accepted as the study of choice to define hemisphere dominance for language function before neurosurgical resective procedures, especially in epilepsy. It is also used as the gold standard against which novel language lateralization techniques are judged. The majority of these new techniques rely on the analysis of activation patterns during language tasks rather than on the evaluation of behavior and cognitive function following temporary and reversible pharmacological cerebral deactivation.
fMRI language tasks are most widely used in lateralizing language function and the paradigms have been found to provide valuable data in the assessment of hemispheric language dominance in patients with epilepsy (13). However, concerns have been raised about the reliability of these newer techniques in patients with atypical language representation. The interpretation of bilateral fMRI activation patterns is particularly challenging. There are large individual differences in fMRI language activation patterns in patients with atypical language representation demonstrated by the Wada tests (14). In addition, there is interhemispheric dissociation between speech production and comprehension (15). Speech production capacity is more likely to shift hemisphere (away from the seizure focus) than language comprehension. The Wada test can answer one question that the fMRI cannot address: Can a task still be performed if part of a particular hemisphere is removed?
Magnetoencephalography (MEG) is showing some promise in identifying both frontal and temporal language areas. Transcranial magnetic stimulation was introduced in the 1990s as a possible alternative to the Wada test, but it has been poorly tolerated by many patients and often produces significant discrepancies when compared to the Wada test results. Functional transcranial Doppler sonography in which the blood flow is measured in the bilateral middle cerebral arteries during the performance of a language task also shows good concordance with the Wada test language lateralization and may be better tolerated by special populations (16).
MEMORY EVALUATION
Validity
The primary purpose of testing memory during the Wada test is to identify those at risk for a postoperative amnesia.
Memory failure during the Wada test is diagnostic because it evaluates the capacity of the brain to support global, anterograde memory when certain brain structures are pharmacologically inactivated. The prognosis of the postoperative amnesia derives from the assumption that memory performance during the Wada test correlates with everyday anterograde memory, and the short-term effects of the Wada test are sufficiently analogous to the long-term effects of surgery on memory.
The incidence of severe postoperative amnesia after anterior temporal lobectomy is low with only 10 reported cases in the past 50 years. Most were reported before the MRI was available (17,18). Nine of these occurred following dominant temporal lobe resection, and all of the cases showed evidence of contralateral dysfunction or atypical dominance independent of the Wada test. However, the predictive value of the Wada test with regard to postoperative amnesia may not be known.
Criticism of the Wada test in predicting postoperative amnesia has increased after reports of false positive results. There are frequent reports in the literature of patients who failed both ipsilateral and contralateral injections during a Wada test and underwent unilateral temporal lobe resection. None of these patients developed global amnesia, although some reported postoperative memory decline. A large number of these patients also demonstrated a greater than 90% improvement in seizure control postoperatively (19,20). In addition, there are numerous reports of patients who failed the Wada test memory assessment but subsequently passed a repeat procedure and underwent surgery without any subsequent memory impairment (21). This certainly raises questions about the validity of this test for memory.
The literature lacks consensus regarding the predictive value of the Wada test with regard to postoperative memory decline. Baseline memory verbal assessment has been shown by many studies to be a particularly useful predictor of verbal memory outcome. Memory decline is more likely when ipsilateral memory is good or contralateral memory is poor (22).
Asymmetric memory performance on the Wada test may predict memory decline. Prediction of postoperative memory decline is based on concepts of functional reserve and functional adequacy. Functional reserve is linked to the integrity of the mesial temporal region contralateral to the seizure focus and represents the principal determinant of postoperative memory decline, particularly in predicting global amnesia. Evaluation of functional adequacy of the potential surgical side has greater reliability for material specific memory changes.
The most important variables that predict memory loss are baseline memory function as assessed by neuropsychological testing, the hemisphere subjected to surgery (the risk of severe amnesia exists when the language dominant hemisphere is targeted), and the existence of a structural lesion that will be removed.
AMOBARBITAL ALTERNATIVES
Although amobarbital has been widely used to evaluate the hemispheric dominance of language and memory before temporal lobe surgery in patients with medically refractory seizures, repeated shortage of this agent has prompted many centers into using alternatives such as etomidate, propofol, sodium methohexital, or pentobarbital. Most of the published studies involve retrospective chart reviews. Intracarotid administration of 2 mg of etomidate produces a clear EEG and motor effects in all patients, with a response very similar to use of amobarbital. Shivering was the most common reported side effect (23). Methohexital, pentobarbital, and propofol have all been studied as alternatives to sodium amobarbital for the Wada test. All these required a second injection due to their short duration of action (24). There was an increased risk of seizures with methohexital use, transient respiratory depression immediately after receiving pentobarbital, and increased tone with myoclonic jerks seen with propofol use.
REGIONAL CEREBRAL PERFUSION
The usefulness of the Wada test is limited because of inactivation of widespread hemispheric function. Therefore, selective Wada tests have been developed to deliver amobarbital exclusively in the brain portion intended to be resected. The distribution of amobarbital and its effect on regional cerebral perfusion during the Wada test using high-resolution hexamethyl propyleneamine oxime (HMPAO) SPECT coregistered with patient’s MRI data has been described in a pool of patients undergoing assessment for temporal lobectomy (25). The present study not only found complete mesial temporal hypoperfusion in 36% of patients, but it also found partial hypoperfusion in another 56% of patients, pointing out that the great majority of patients had some effect of the amobarbital injection on perfusion levels in the mesial temporal cortex. The lateral temporal lobe and pole showed the most perfusion, the mesial structures showed less. In mesial structures there was a variable pattern of hypoperfusion, with some patients showing preservation of anterior perfusion and some patients showing preservation of posterior perfusion. Preservation of posterior mesial perfusion can be explained by vascular anatomy, whereas anterior preservation is more difficult to explain. Noted as well was a lack of hypoperfusion in the basal ganglia that occurred despite the fact that intracarotid HMPAO delivery and therefore amobarbital delivery was similar to other structures. This apparent resistance to ambarbital-induced hypoperfusion is postulated to be due to relative paucity of GABAA receptors that normally bind barbiturates in these structures (26).
Anterior Cerebral Artery Wada Test
Newer techniques of inactivation of the mediobasal temporal lobe structures consist of temporary balloon occlusion distal to the origin of the anterior choroidal artery (AChA) and selective catheterization of the AChA. Only a small number of patients undergo a selective Wada test. Selection criteria included bilateral mesiobasal temporal lobe epileptogenesis, suggested by bilateral mesiobasal temporal lobe seizure onset, excessive bilateral interictal spiking, bilateral damage in structural or functional images of the temporal lobes, marked bilateral neuropsychological deficits in noninvasive neuropsychological testing, or unclear speech dominance with left-handedness. The main steps of the procedure are the introduction of a transfemoral balloon catheter, tolerance test of occlusion of the internal carotid artery by injecting contrast medium into the contralateral ICA with the compression of the ipsilateral ICA and temporary balloon occlusion distal to the origin of the AChA with injection of contrast and amobarbital. This allows for selective injection of the vascular territories of the AChA, the posterior communicating artery, and the ophthalmic artery.
The average amount of amobarbital injected with these techniques is 75 mg. As with the ICA Wada test, the patient is familiarized with the neuropsychological test procedure the day before or before the test using a similar but shorter version containing stimuli different from the real test. Continuous behavioral monitoring by videotaping the patient and synchronized EEG monitoring from scalp electrodes and preferably bilateral foramen oval electrodes are performed. Neuropsychological performances are recorded during the selective temporal lobe amobarbital test and among the three techniques: balloon techniques, superselective injection of the AChA, and the P2 segment of the posterior cerebral artery (PCA).
In comparison to preinjection performance, as expected, the postinjection performance decreases in both learning and recognition. The degree of postinjection performance depends on the type of selective amobarbital test. The left PCA selective amobarbital test has the most pronounced effect on learning and visual recognition. Verbal recognition performance decreased most with the left balloon technique. A more pronounced deficit for verbal than for visual material was seen following the left selective amobarbital injections, and the deficit was more pronounced for visual rather than verbal material following right selective injections. The selective temporal lobe Wada tests described here mimic and predict the effects of planned selective temporal lobe surgery much better than the ICA Wada test. There is a good reported correlation between memory assessed during amobarbital injection and the actual memory outcome following surgery, suggesting that the selective Wada tests have a good predictive value for memory outcome after amygdalohippocampectomy. Given the inherent variability of the vascular anatomy of the AChA, in cases in which this artery supplies only a small portion of the anterior hippocampus, the balloon technique appears preferable to the catheterization of the AChA. In cases with a very small AChA and a small P.Comm.A. The catheterization of the P2 segment of the PCA is probably the most reliable selective temporal lobe Wada test, especially in cases in which an extensive posterior mediobasal resection is intended (27).
Posterior Cerebral Artery Wada Test
Another procedure used by some epilepsy centers for assessing memory function is the selective injection of sodium amobarbital into the PCA. There are certain theoretical advantages of this procedure, especially the fact that the patients do not become aphasic and the drug can be delivered more effectively to the target regions, the ipsilateral hippocampus.
In this procedure, similar to the Wada test, a transfemoral catheter is used and a diagnostic angiogram is obtained in the common carotid artery ipsilateral to the proposed neurosurgical procedure, as well as in the left, right, or both vertebral arteries. The catheter is then placed in the vertebral artery and the tip is placed in the peduncular segment of the PCA with the aid of a guided wire.
With the catheter in place but before the injection the patient is presented with a list of four words read aloud. The patient is asked to recall the list of four words. At the author’s center, for the memory testing procedure, amobarbital 75 mg is used to inject into the PCA. Hemianopia is used as a marker of the functional effect of the amobarbital in the PCA system. The EEG rarely shows brief bursts of theta frequency slowing in the ipsilateral posterior quadrant and is therefore an unreliable marker for the clinical efficacy of the injection. When the hemianopia is confirmed, free recall of the objects given before injection is attempted, followed by a recognition test. Immediately, a second list of six words is given with several learning trials. The hemianopia is tested every minute to assess whether the drug is still active. Hemianopia usually resolves in about 5 to 6 minutes and free recall of the second lists of words, followed by recognition test is completed. The primary assessment involves both learning and delayed recall.
Since the PCA amobarbital test does carry an increased risk of stroke and requires particular expertise in cerebral angiography, a stepwise procedure involving both the ICA and then the PCA is usually attempted. Since the intracarotid approach involves a lower risk, it is reasonable to perform this procedure initially on all patients. If the patient fails the ICA Wada test for memory, an additional test involving the PCA approach might be considered, especially in patients with normal hippocampi or dual intrahemispheric pathology. If the patient passes the PCA Wada test, it would be reasonable to consider surgery.
Middle Cerebral Artery Wada Test
The middle cerebral artery (MCA) Wada test involves placing the guiding catheter in the ICA below the skull base and navigating the microcatheter into the M1 segment distal to the origin of the lenticulostriate arteries or into a single MCA branch. The final catheter placement is checked by angiographic series. The MCA Wada test is performed by using the same standardized protocol that has been described for the ICA test.
The indication for the MCA Wada test is the assessment of the residual motor function before intended hemispherectomy, resection of epileptogenic lesions in or adjacent to the motor cortex or Wernicke’s area, electrical status epilepticus of sleep (ESES), or partial-onset seizures with quick bilateral synchrony to the opposite side to evaluate whether ongoing bilateral epileptiform activity could be suppressed by the unilateral injection (28). In patients with developmental or early acquired hemispheric lesions involving the motor cortex, in which the functional hemispherectomy is intended, if the simple tasks (fingers flexion, thumb opposition of the thumb) are preserved, the question of disconnecting the affected motor cortex from the healthy brain without producing contralateral hand plegia cannot be answered. The type of the lesion and the timing of the insult are important parameters when predicting a patient’s level of function after hemispherectomy (29). Patients with developmental lesions appear to carry a lower risk of additional motor deficits compared to patients with acquired lesions in early childhood. The prediction of postoperative deficits is most difficult in patients with relatively good motor function who had hemispheric insults perinatally and beyond the neonatal period but before completion of myelination (end of third year of life). In these cases, transcranial magnetic stimulation is difficult to perform because of distorted regional anatomy. fMRI evaluations in hemiparetic patients are often disturbed due to motion artifact.
Selective MCA Wada tests are of lesser importance in patients with acquired lesions in or adjacent to the motor cortex or the classic language areas and subdural grid electrode stimulation or fMRI are more appropriate to lateralize language functions. Although the accuracy of these investigations to detect different language representation sites within one hemisphere is yet to be determined. In patients with ESES or Landau-Kleffner syndrome, ictal epileptic discharges are usually bilateral, necessitating catheterization of the single MCA branches, usually bilaterally with higher angiographic risk.
Selective catheterization of the single MCA trunk or branches is associated with a risk of permanent neurological deficits by inducing thromboembolic vessel occlusion or vasospasm. This risk is most likely significantly higher than with ICA Wada test. One also must be aware that even most accurate amobarbital injections do not guarantee that the drug reaches the target area. While an expected and regularly occurring contralateral hemianopia in the PCA Wada test is proof that amobarbital reached its target area, neurologic deficits are often absent in MCA Wada tests of patients with perinatal hemispheric infarcts.
In presurgical workup of epilepsy patients selective MCA Wada tests are rarely performed. The main indication is the functional inactivation of the motor cortex before functional hemispherectomy, particularly in patients with partially preserved fine motor control of the contralateral hand and finger.
