The availability of amobarbital became sporadic beginning in the late 1990s. In addition, amobarbital is no longer manufactured by Lilly but is produced by Raxbury, which does not distribute the drug outside the USA. This had two direct effects on Wada testing. First, alternative anesthetic agents including methohexital, propofol, and etomidate were explored as agents to create hemispheric anesthesia for traditional Wada language and memory testing. The second effect, however, was to focus attention on issues regarding whether Wada testing should be the standard of care in the evaluation of all epilepsy surgery candidates.

Although formal surveying has not been done, the second most common anesthetic after amobarbital is likely methohexital (Brevital). Methohexital was introduced for Wada testing at the University of Florida due to its relatively short duration of action that would facilitate both hemispheres being studied on the same day because left and right amobarbital Wada injections were performed on different days at many surgical centers (Gilmore, Heilman, Schmidt, Fennell, & Quisling, 1992). The largest study using methohexital is from the University of Michigan (Buchtel, Passaro, Selwa, Deveikis, & Gomez-Hassan, 2002), and this drug has also been more recently adopted by the Cleveland Clinic (Lineweaver et al., 2006). In contrast to amobarbital administration, methohexital is typically given in divided doses of 3 mg followed shortly by 2 mg, although the reasons for this rather than a single mg injection are not clear and whether this approach is superior in any way to a single 5 mg injection has not been established. Single doses of 6–8 mg have been used without any apparent difficulty (Marla Hamberger, personal communication, December 5, 2009). In general, language and memory results using methohexital appear to be equivalent to those obtained using amobarbital. However, there appears to be increased seizure risk since methohexital is associated with a decreased seizure threshold (Loddenkemper, Moddel, Schuele, Wyllie, & Morris, 2007).

Propofol is a nonbarbiturate anesthesia with a very short duration of action and has been successfully used as a substitute for amobarbital. Dosing of 20 mg of propofol produces results comparable to 120 mg of amobarbital as reflected by the time to verbal and nonverbal responses (Mikati et al., 2009). However, due to its short duration of action, a number of patients require incremental injections to maintain the desired level of motor weakness. Approximately one-third of patients had an adverse event following propofol injection, with 12 % of all patients having increased muscle tone with twitching, rhythmic movements, or tonic posturing. Patients older than 55 or receiving a total injection dose greater than 20 mg were more likely to have significant adverse events, which in turn carries a risk of incompletion or inaccuracy of the Wada test.

Another nonbarbiturate anesthetic alternative to amobarbital is etomidate, which has rapid onset and short duration of action. Several approaches to etomidate administration have been successfully employed. At the MNI, a bolus of 2 mg etomidate that is administered followed by infusion insures a uniform period of anesthesia (Jones-Gotman et al., 2005). The bolus is followed by an infusion (0.003–0.004 mg/kg/min) at a rate of 6 mL/h. The use of a constant infusion allows direct control of the period of hemispheric anesthesia, insuring that all testing and memory item presentation is presented during maximum anesthesia rather than during recovery from anesthesia in which the medication effects are not fully known and may not be equivalent between hemispheres.

The difficulty in obtaining amobarbital has had the benefit of making centers ask the question of whether Wada testing can be or should be performed in all patients and to evaluate to what degree Wada testing should continue to be performed in preoperative epilepsy surgery evaluation. Advances in functional and structural imaging have provided predictors of memory outcome, thereby decreasing the sole reliance on Wada memory results in this context. However, there continues to be confusion about the role of the Wada test since so many procedural variations exist. Further, difficulty also exists given the variability in neuropsychological memory outcome measures and what the definitions of “significant change” are employed. Ultimately, these pieces of information are necessary to inform epilepsy programs which patient should be considered for Wada testing since the cost/benefit ratio will vary based upon clinical semiology, patient characteristics, and Wada protocol employed.

The potency of anesthetic effects of amobarbital is reduced by carbonic anhydrase inhibitor medications such as topiramate or zonisamide (Bookheimer, Schrader, Rausch, Sankar, & Engel, 2005; Burns et al., 2009; Kipervasser et al., 2004). Thus, amobarbital dose adjustments may be required, or testing patients off their carbonic anhydrase medications may be preferred.

The total duration of amobarbital anesthetic effect depends both on the initial drug dose, whether additional incremental injections are administered (to produce a flaccid contralateral hemiplegia), and individual patient differences in drug responsivity. However, behavioral recovery tends to be rapid, rarely lasting more than 10 min. In our experience, comprehension of complex two-stage commands involving inverted syntax is typically the last language-related ability to recover, and repetition is the next most sensitive language measure to mild residual medication effects. Typically, the last motor deficits to resolve include pronator drift, decreased motor speed, or asterixis contralateral to the side of injection. Return of language and motor functions to baseline levels is necessary prior to assessing memory.

Patients undergoing Wada testing have a pretest baseline evaluation that serves two purposes. The first is to familiarize the patient with the specific techniques that will be administered during the test. The second is to establish a baseline level of function, which is particularly useful for language tasks since they may be affected by schooling, clinical history, as well as having idiosyncratic regional dialects. With few exceptions, patients will obtain perfect scores of 8/8 for memory testing using our recognition approach. The order of injection between the presumed ipsilateral and contralateral sides is alternated across patients in order to ensure that there is not a systematic confound associated with injection order. For example, to the extent that there are residual amobarbital effects lasting beyond the typical testing window, always beginning the ipsilateral injection may tend to decrease WMA magnitude. Both hemispheres are typically tested on the same day, with a minimum of 30 min separating the two hemispheric studies. Patients are forewarned that they may experience arm weakness or difficulty talking, although memory for either impairment is often poor. Patients are also told that the medication makes some people feel a little bit drunk or sleepy. This is included to reduce anxiety and to create an expectancy to help the patient interpret subjective drug-induced effects. After beginning to tell patients about potential subjective drug effects, we have decreased the number of “confused” or mildly agitated responses to nearly zero.

Patients begin counting repeatedly from 1 to 20 with their hands held up, their palms turned upward, and fingers spread. An injection of 100 mg amobarbital is administered by hand over a 4–5 s interval via a percutaneous transfemoral catheter. Following demonstration of hemiplegia and evaluation of eye-gaze deviation, the patient is requested to execute a simple midline command (e.g., “stick out your tongue”). We do not require a complete flaccid hemiplegia in order to proceed with cognitive testing and do not base our timing of stimulus presentation according to EEG-based information. If there is only minimal change in muscle tone, an incremental injection of 25 mg may be used. However, we try to avoid more than one additional dose increment since it has been our experience that multiple injections increase the likelihood of patient obtundation compared to an identical total dose delivered with a single bolus. Complete lack of motor deficit suggests misplacement of the catheter.

Following evaluation of simple comprehension for midline commands and eye gaze, eight common objects are presented for 4–8 s each in the visual field midline and ipsilateral to the injection, and the object names are repeated twice to the patient. On average, this occurs beginning approximately 30–45 s following injection. Examples of Wada memory items include a combination of ordinary household items (e.g., fork, mousetrap), small toys (e.g., doll, plastic shark), and plastic food (e.g., hotdog, pizza). At times, due to patient confusion, inattention, or nonresponsiveness, the patient’s eyes are held open. Language, discussed below, is assessed in detail following presentation of memory items. We also impose a minimum of 10 min between drug administration and memory assessment. Recognition memory of material presented during the procedure is tested after amobarbital effects have worn off as demonstrated by return baseline language performance on all tasks described below, return of 5/5 strength, and absence of pronator drift, tactile extinction, asterixis, and bradykinesia.

Language assessment contains five common language domains (viz., counting disruption, comprehension, naming, repetition, and reading). Although we have developed a formalized approach to calculate a language laterality ratio [(left score minus right score)/(left score plus right score), (L − R)/(L + R)], this laterality ratio is used for research purposes and is not routinely reported clinically (Loring et al., 1990b).