Effect of Anesthesia on Evoked Auditory Responses in Pediatric Auditory Brainstem Implant Surgery

Abstract

Objective:

Electrically evoked auditory brainstem responses (EABR) guide placement of the multichannel auditory brainstem implant (ABI) array during surgery. EABRs are also recorded under anesthesia in nontumor pediatric ABI recipients prior to device activation to confirm placement and guide device programming. We examine the influence of anesthesia on evoked response morphology in pediatric ABI users by comparing intraoperative with postoperative EABR recordings.

Study Design:

Retrospective review.

Methods:

Seven children underwent ABI surgery by way of retrosigmoid craniotomy. General anesthesia included inhaled sevoflurane induction and propofol maintenance during which EABRs were recorded to confirm accurate positioning of the ABI. A mean of 7.7 ± 2.8 weeks following surgery, the ABI was activated under general anesthesia or sedation (dexmedetomidine) and EABR recordings were made. A qualitative analysis of intraoperative and postoperative waveform morphology was performed.

Results:

Seven subjects (mean age 20.6 months) underwent nine ABI surgeries (seven primary, two revisions) and nine activations. EABRs were observed in eight of nine postoperative recordings. In three cases, intraoperative EABRs during general anesthesia were similar to postoperative EABRs with sedation. In one case, sevoflurane and propofol were used for intra- and postoperative recordings, and waveforms were also similar. In four cases, amplitude and latency changes were observed for intraoperative versus postoperative EABRs.

Conclusion:

Similarity of EABR morphology in the anesthetized versus sedated condition suggests that anesthesia does not have a large effect on far-field evoked potentials. Changes in EABR waveform morphology observed postoperatively may be influenced by other factors such as movements of the surface array.

INTRODUCTION

The auditory brainstem implant (ABI) is a neuroprosthetic device that electrically stimulates second-order auditory neurons of the cochlear nucleus using a multichannel surface array. The ABI was originally developed to provide sound sensations in patients with neurofibromatosis type 2 (NF2) and was approved by the U.S. Food and Drug Administration in 2000 for NF2 patients ages 12 and older.¹ New indications have included congenital deafness associated with hypoplasia or aplasia of the cochlea or cochlear nerve. Several North American clinical trials have examined the safety and feasibility of ABI surgery in nontumor children who are deaf and are not candidates for the cochlear implant.^2–4 Auditory benefits with the pediatric ABI range from environmental sound awareness in most subjects to open-set speech perception in a few cases.^2,5–7

ABI surgery is challenging. The convexity of the cerebellum and the narrow surgical corridor of the foramen of Luschka obscures direct visualization of the auditory brainstem during either translabyrinthine⁸ or retrosigmoid craniotomy.⁹ The surgical team relies on indirect anatomic landmarks such as the choroid plexus and root entry zone of the glossopharyngeal nerve to place the electrode paddle into the lateral recess of the IVth ventricle. Despite this blind surgical approach, image guidance has not yet been developed for ABI placement, although recent data suggests that computed tomography (CT) can resolve detailed positioning data of the ABI array.¹⁰ Consequently, electrophysiologic measures are needed to confirm accurate placement of the ABI array.

Intraoperative electrically evoked auditory brainstem responses (EABRs) following ABI placement were first described by Waring in 1999¹¹ and serve two key roles. EABRs generated during electrical stimulation of the cochlear nucleus are thought to indicate the ABI paddle position most likely to activate auditory neurons. These electrophysiologic data guide the surgical team and help maximize the number of electrodes eliciting an auditory sensation.¹¹ Continuous intraoperative EABR monitoring following final placement also ensures that the array did not shift during manipulation of the harness when closing dura and soft tissue.^8,9

At our center, pediatric ABI recipients undergo initial device activation and EABR testing under sedation or general anesthesia followed by awake activation the next day. EABRs identify electrodes most likely to elicit auditory sensation versus those likely to elicit side effects from neighboring nonauditory axons of passage.^2,7 A change or absence in postoperative EABRs (compared to intraoperative measurements at the time of device placement) may 1) indicate shifting of the ABI array position due to intracranial pulsations or brain growth and 2) predict modest or absent perceptual responses in the awake condition. Finally, monitoring of nonauditory stimulation under anesthesia is essential in children to eliminate electrodes associated with cardiovascular changes or noxious responses.

Given the importance of monitoring EABR changes during and after ABI surgery to help guide initial device mapping in children, it is critical to understand whether electrophysiological measurements are influenced by external perioperative factors such as the choice of anesthetic agent. Anesthesia has shown to have a dose-dependent effect on amplitude, threshold, and latency of sound-evoked ABRs in frogs,¹² mice,¹³ rats,^14,15 and lizards.¹⁶ In humans, specific anesthetic regimens have also been shown to have an impact on EABRs.¹⁷ For example, in a study of 12 children between the ages of 29 and 52 months, significant delays in ABR latencies were observed for wave V and interpeak intervals I to III, III to V, and I to V.

This study was undertaken because of the opportunity to compare within a single ABI subject EABRs in the anesthetized intraoperative condition to the sedated postoperative condition several months later. We hypothesize that the variance in postoperative EABR morphology compared to intraoperative measurements during ABI surgery cannot be explained entirely by differences in anesthesia. The goal of this study was to characterize the association of anesthetic approach with changes in postoperative EABRs in pediatric ABI patients.

MATERIALS AND METHODS

This study was approved by the Human Studies Committee at Massachusetts Eye and Ear (13-028H). All subjects were part of the safety and feasibility trial for auditory brainstem implantation in non-NF2 children approved by the Food and Drug Administration (NCT01864291). Inclusion and exclusion criteria for this trial are listed in Table I. Candidates who fulfilled these criteria underwent retrosigmoid craniotomy and placement of either the Cochlear Nucleus N24 ABI or Nucleus ABI541 (Cochlear Corporation). Mean age at primary ABI surgery was 20.6 ± 7.7 months (Table II), and average age at initial clinic evaluation was 18 ± 8.1 months (standard deviation). The first four subjects were implanted with the Cochlear Nucleus ABI24, and the last three were implanted with the Cochlear ABI541 after production of the Nucleus ABI24 was discontinued. All surgeries used a retrosigmoid craniotomy approach. Six subjects received an ABI on the right side and one on the left side. Subject one (S01) underwent revision surgery for device failure after impact from a mechanical fall, and another subject (S02) underwent surgery for spontaneous device failure approximately 1 year following initial implantation. Inhalation induction with sevoflurane and propofol for maintenance was utilized for all nine intraoperative EABR recordings (Table III).

TABLE I.

Subject Criteria for Pediatric Auditory Brainstem Implant.

Inclusion Criteria

1. Prelinguistic hearing loss with: MRI ± CT evidence of one of the following: Cochlear nerve deficiency Cochlear aplasia or severe hypoplasia Severe inner ear malformation Postmeningitis ossification When a cochlea is present or patent with a normally appearing cochlear nerve, lack of significant benefit from CI despite consistent use (>6 months) No or limited speech perception ability (limited to pattern perception on closed set testing materials using the CI) Lack of progress in auditory skills development

2. Postlinguistic hearing loss (<18 years of age) with both: No benefit from CI without possibility for revision or contralateral implant Postmeningitis ossification Bilateral temporal bone fractures with cochlear nerve avulsion Failed revision CI without benefit Previous open set speech perception and auditory-oral language skills

3. Ability to tolerate general anesthesia

4. Receipt of the appropriate meningitis vaccinations

5. No or limited cognitive/developmental delays, which would be expected to interfere with the child’s ability to cooperate in testing and/or programming of the device or in developing speech and oral language

Exclusion Criteria

1. Pre- or postlinguistic child currently making significant progress with CI

2. MRI evidence of one of the following: Normal cochlea and cochlear nerves or NF2 Brainstem or cortical anomaly that makes implantation unfeasible

3. Clear surgical reason for poor CI performance that can be remediated with revision

4. Intractable seizures or progressive, deteriorating neurological disorder

5. Lack of potential for spoken language development

6. Unable to participate in behavioral testing and mapping with their CI

CI = cochlear implant; CT = computed tomography; MRI = magnetic resonance imaging; NF2 = neurofibromatosis type 2.

TABLE II.

Subject Characteristics.

	S01	S02	S03	S04	S05	S06	S07
Age at time of primary surgery	18 months	11 months	15 months	30 months	20 months	16 months	34 months
Gender	Male	Female	Female	Male	Female	Female	Male
Birth history	Preterm	Term	Term	Term	Term	Term	Term
	Oligohydramnios	CHARGE syndrome		Hyperbilirubinemia	Meconium aspiration		CHARGE syndrome
Family history of hearing loss	None	None	None	None	None	None	Maternal grandparent
Newborn screening	Absent ABR	Absent ABR	Absent ABR	Absent ABR	Absent ABR in one ear; inconsistent results in other ear	Absent ABR	Absent ABR
Prior intervention	None	None	None	None	Failed left cochlear implant trial	Failed hearing aid trial	None

ABR = acoustic auditory brainstem response; S = subject.

TABLE III.

Summary of Anesthesia/Sedation Used and EABR Comparisons.

	S01	S01R	S02	S02R	S03	S04	S05	S06	S07
Summary of Sedation/Anesthesia and EABR
Intraoperative anesthesia	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol	Sevoflurane & propofol
EABR (intra-op)	Present	Present	Present	Present	Present	Present	Present	Present	Present
Postoperative anesthesia	Dexmedetomidine	Dexmedetomidine	Dexmedetomidine	Dexmedetomidine	Sevoflurane & propofol	Dexmedetomidine convert to sevoflurane	Dexmedetomidine	Sevoflurane & propofol	Sevoflurane & propofol
EABR (postop)	Present	Not available*	Present	Present	Present	Present	Present	Present	Present
Comparison of Intraoperative and Postoperative EABR
Number of electrode pairs compared	4	0*	13	7	7	12	4	12	9
Waveform morphology	Similar for 2 electrode pairs		Similar for 10 electrode pairs	Similar for 4 electrode pairs	Similar for 5 electrode pairs	Major changes on 6 electrode pairs	Major morphology changes on all 4 electrode pairs	Major changes on 2 electrode pairs, slight changes on others	Major morphology changes on 8 electrode pairs
Waveform amplitude	Postop response larger amplitude for 1 pair		All postop responses smaller in amplitude	All but 1 postop response larger in amplitude	All postop responses slightly larger in amplitude	All postop responses larger in amplitude	All postop responses smaller in amplitude	4 postop responses larger in amplitude	1 intra-op response absent postop
Myogenic	1 pair with myogenic only in postoperative response		4 pairs with myogenic in intra- and postop responses. 1 pair with myogenic only in postop response	No myogenic in either intra- or postop	No myogenic in either intra- or postop pairs	No myogenic in either intra- or postop pairs	No myogenic in either intra- or postop pairs	No myogenic in either intra- or postop pairs	No myogenic in either intra- or postop pairs
Latency	Minimal differences		Minimal differences	Minimal differences	Minimal differences	Morphology changes prevent comparison	Morphology changes prevent comparison	Minimal differences	Morphology changes prevent comparison

^*Technical difficulties in postoperative recording prevented comparison.

EABR = electrically evoked auditory brainstem responses; intraop = intraoperative; postop = postoperative; R = revision; S = subject.

Following ABI array placement, intraoperative EABRs were obtained, and the array position was adjusted to optimize evoked response measurements. The method for eliciting intraoperative EABRs was similar to that previously described.¹⁸ Briefly, EABRs were recorded using an Xltek Protektor IOM system (Natus Inc.) triggered by manufacturer stimulating software. Bipolar electrical stimulation used single biphasic pulses (100 or 150 μs phase duration, 8 μs interphase gap) delivered to the ABI array through a processor and coil. The most frequent electrode pairs stimulated were either at the corners of the electrode paddle, that is, electrodes 2 and 8, or crossing the paddle, that is, electrodes 11 and 12 (Fig. 1). The number of stimulated electrode pairs vary between subjects for several reasons, including the amount of time available for recording intra- and postoperatively, the number of electrode pairs with responses, and the morphology of those responses. Even with EABR guidance during placement of the array, the physical orientation of implanted electrodes also varies across subjects. Between two and 10 electrode pairs were tested per subject (Table III). Responses were recorded using the electrode montage described by Waring¹⁵ (vertex (+) to nape (−) with the ground electrode at the hairline). Average hospital stay was 2.9 ± 0.9 days.

Fig. 1.

Schematic of Cochlear 21-channel ABI surface array. Example bipolar stimulation pairs used for electrically evoked ABI recordings. Electrode pair 2/8 (dark gray) illustrates a pair from the lateral superior quadrant. Electrode pair 17/18 traverses the electrode array at the medial aspect of brainstem. ABI = auditory brainstem implant.

The mean time of postoperative EABR evaluations was 7.7 ± 2.8 weeks (range: 6 to 12 weeks) after surgery. All subjects had EABR testing under light sedation or general anesthesia approximately 1 to 2 days prior to awake device activation. Intramuscular dexmedetomidine was used in six sedated evaluations (66.7%, 6 of 9) and general anesthesia with sevoflurane/propofol induction was used in three evaluations (33.3%, 3 of 9). One sedated evaluation was converted from dexmedetomidine to general anesthesia with sevoflurane after the initial sedation techniques failed. EABR responses were obtained for all postoperative evaluations, except one in which technical equipment problems prevented recording. Methods were similar to intraoperative recordings except for the use of a custom-programmed evoked potential system¹⁴ (instead of the Xltek; Natus Inc.). The same electrode pairs were stimulated using bipolar stimulation paradigm as previously described.

RESULTS

Subject Characteristics

Patient demographics, operative events, and anesthetic agent(s) used during ABI surgery and at initial activation are summarized in Table II. On high-resolution magnetic resonance imaging with parasagittal sequences, the most common structural anomalies were cochlear nerve (CN VIII) aplasia (100%, 7 of 7) with bilateral cochlear hypoplasia (85.7%, 6 of 7). One subject also had bilateral cochlear aplasia (14.3%, 1 of 7). Two subjects had CHARGE syndrome, and one subject had a family history of hearing loss.² One subject had a failed cochlear implant trial, and another subject had a failed trial with hearing aids.

Intraoperative (During ABI Surgery) and Postoperative (at Activation) EABRs

In total, seven children underwent nine ABI surgeries (7 primary, 2 revisions) and nine activations. EABRs were obtained in all surgeries and all but one postoperative recording. For all surgeries, intraoperative EABRs were used to adjust array position until the maximum number of responses with adequate waveform morphologies was achieved. Evoked responses associated with the final position of the ABI array were used for this study. Overall waveform morphology was similar to that seen in adult NF2 patients, with waves often identified as N1, P1, and P3, as well as later positive waves between 6 and 8 ms.¹⁸ The presence or absence of positive waves after 15 ms could not be assessed because the recording response window was 0 to 20 ms. Occasionally, some stimulation pairs elicited a large and sharp waveform around 4 ms (e.g., Fig. 2C). We term this waveform a myogenic response because in some cases it was associated with an increase in the activity on cranial nerves VII and/or X, as observed with intraoperative monitoring and accompanied by visible twitching of the face or throat muscles, respectively.

Fig. 2.

Examples of EABRs recorded during auditory brainstem implant surgery and at activation. An illustration of waveform morphology changes noted in Table III. Each bracketed pair of waveforms contains the intraoperative waveform (top) and the postoperative waveform (bottom) from the same electrode pair for a given subject. (A) Example waveforms from two different subjects and electrode pairs illustrating similar intraoperative and postoperative EABR morphology. (B) Example waveforms from two other subjects illustrating major changes in waveform morphology and amplitude between intraoperative and postoperative EABR recordings. (C) Example waveforms from another subject illustrating major changes in waveform morphology primarily due to the addition of a myogenic wave (black arrow) in the postoperative recording. In all recordings, stimulus time begins at time 0 on the time scale (x-axis), and the amplitude scale (2 μV/division) is shown on the y-axis for each plot. EABR = electrically evoked auditory brainstem. [Color figure can be viewed in the online issue, which is available at www.laryngoscope.com.]

Intraoperative and Postoperative EABR Comparison

Figure 2 compares example waveforms of EABR recordings for single subjects during ABI surgery and at activation. Recordings from two subjects are shown in which the waveforms are generally the same in both conditions (Fig. 2A), and recordings from three other subjects are shown in which the waveforms are different (Fig. 2B, 2C). Waveforms from recordings during ABI surgery and at activation for each subject were qualitatively compared across four dimensions: general waveform morphology (number and shape of waves), overall waveform amplitude, the presence or absence of a myogenic-like response, and general wave latency. Overall, waveform morphology was similar during ABI surgery and at activation in half of the eight subjects (Table III). Of those that were similar, three patients (S01, S02, S02R) underwent EABR testing under sedation, and one was done under general anesthesia. Changes in myogenic responses were seen in only two subjects, and those changes were minimal (S01, S02). Latencies were similar when no waveform morphology changes were observed and were consistent with the change in recording equipment.

Four of the subjects had major changes in waveform morphology at ABI activation (Table III). The majority of these patients underwent general anesthesia as opposed to a single patient who was sedated at the time of initial device activation. EABR waveform amplitudes were larger in five subjects (S01, S02R, S03, S04, S06) and smaller in two subjects (S02, S05) at activation compared to recordings during ABI surgery.

DISCUSSION

This is the first study to record and compare EABRs from pediatric ABI patients in the intraoperative, anesthetized condition and the sedated/anesthetized, postoperative condition. We found EABRs were largely unchanged in the two conditions for some subjects, suggesting that anesthesia does not have a large effect on waveform morphology. Other subjects had different EABRs in the postoperative condition, and the interpretation in those cases is less clear.

EABRs for ABIs

The use of intraoperative EABRs to guide ABI electrode placement was first described for adult patients by Waring in 1995.¹⁹ Positive peaks at mean latencies of 0.4 ms and 1.45 ms and/or the presence of a three-wave response within the first 25 ms of stimulus onset were thought to indicate stimulation of the cochlear nucleus.^19–21 Similar waveforms were seen in pediatric patients in the present study and by O’Driscoll et al.²² As found in the present study, almost all previous studies of EABRs from ABI stimulation report variability from subject to subject and for different electrode pairs within a single subject.

There is a correlation between the presence of EABRs and sound perception in awake ABI users,^18,22 and these measures are predictive of auditory sensation at those electrodes.²² Although an increase in the number of “auditory” electrodes are associated with those contacts that evoke EABR waveforms, the absence of EABRs is not necessarily a predictor for lack of sound perception in either adults or pediatric patients.

Anesthesia and EABRs

A number of studies have shown that anesthesia can influence the sound-evoked ABR in animals.^13–17 Van Looij et al. investigated the impact of anesthesia on ABR recordings in mice and found that ABR peak latencies, interpeak latencies, and thresholds were all significantly increased in anesthetized mice compared to the awake condition.¹³ Different anesthetic agents can have different and sometimes contradictory effects. For example, ABR recordings in guinea pigs after anesthetic induction with isoflurane show a dose-dependent decrease in waveform amplitudes but with increased latency. An anesthetic effect on the cochlea does not appear to cause these effects because cochlear responses from awake guinea pigs are not affected by anesthesia as long as cochlear temperature is normal.²³ Thus, the observed changes are likely caused by anesthetic effects on the ABR generators and their pathways in the central nervous system. In humans, a dose-dependent relationship has also been shown between anesthesia and ABRs. These results show that undesired increases in ABR latency may occur in anesthetized children compared to the awake condition.²⁴

In our study, we found that EABR morphology from subjects in anesthetized versus sedated states can be similar. One explanation may be that nonquantitative comparisons of ABR waveform morphology fail to detect small changes. Although the same subject was tested in both conditions, the change in recording equipment required by our experiments was a complicating factor. Secondly, although unlikely, a lack of EABR change could be explained by the fact that the anesthetics used in our study might not have the same effects as anesthetics used in previous studies of the ABR. The choice of inhaled sevoflurane induction and propofol maintenance was made based on several considerations. Patients for neurosurgical procedures such as ABI surgery are at a greater risk for elevated intracranial pressure due to hypoxia and hypercapnia, and care should be taken during anesthetic induction to avoid this. Induction can be achieved intravenously with thiopentone or propofol as well as through inhalation by facemask. In our subjects, induction was by inhalation of a volatile anesthetic agent such as sevoflurane and nitrous oxide, with a loading dose of propofol 2 mg/kg followed by an infusion of propofol at 200 mcg/kg/min. An intravenous infusion of propofol with sevoflurane allows for a motionless patient and uninterrupted and proper/accurate neuromonitoring, which are crucial during brainstem manipulation. Induction with sevoflurane followed by continuous infusion of propofol for maintenance was used for all surgical procedures in our study. Finally, another possibility for lack of EABR change is that the sedative used also has an anesthesia-like effect on the EABR; thus, changes might have occurred if testing occurred in the awake child. Intramuscular dexmedetomidine was used for six sedated activations, whereas general anesthesia with sevoflurane induction and propofol was used for three activations. A dichotomy exists between agents used for general anesthesia (sevoflurane and propofol) and sedation (dexmedetomidine).^24,25 General anesthetics cause complete loss of consciousness as well as analgesia, amnesia, and muscle paralysis. In contrast, sedatives are primarily used for light-to-moderate sedation with intact consciousness. Some sedatives such as dexmedetomidine also have analgesic properties.²⁶

Cases Where EABR Morphology Changed Following ABI Surgery

EABR waveforms from several subjects varied from intraoperative measures during ABI surgery to activation under anesthesia in two conditions. Changes in some EABR recordings from children have been previously reported.²² Anesthetics used during ABI placement and postoperatively at activation were no different from those in the subjects with no changes in EABR. One explanation is that a small change in the three-dimensional orientation of the ABI array could result in changes in electrically evoked responses. A change in the position of the device has been shown to dramatically change auditory outcomes.²⁷ Future studies at our center will test this hypothesis and build on our recent work that correlates perception with detailed positioning data of ABI arrays on CT in both adults and children.¹⁰ Finally, another possible explanation for changes seen on EABR following device placement could be a change in the size and orientation of the electrical field due to tissue changes after healing.

Limitations

Our study has several limitations. First, we report observations from a small cohort of pediatric ABI users, and therefore this study has limited power. Currently, pediatric ABI surgery is performed at only a few centers worldwide. Nevertheless, the consistency of our results increases the strength of our conclusion regarding the effects of anesthetic agents on electrically evoked responses. Another limitation is the retrospective nature of this study, which does not allow for controlling variables such as type of anesthetic agent used.