Key Words
cochlear implantation - electrically evoked compound action potential - hearing preservation
- maximum amplitude - residual hearing - threshold
INTRODUCTION
In recent years, minimally traumatic surgical techniques, hearing preservation electrodes,
and hybrid electric–acoustic processors have expanded the indications for cochlear
implantation (CI). This trend is supported by growing evidence, indicating that combined
electric and acoustic stimulation lends to improved speech recognition in complex
listening environments, sound localization, music appreciation, and decreased listening
effort ([Turner et al, 2008]; [Buchner et al, 2009]; [Gifford et al, 2013]; [2017]). Hearing preservation after cochlear implant surgery is facilitated by electrode
design and surgical technique, minimizing insertion trauma ([Nadol et al, 1989]; [O’Connell et al, 2016]; [Wanna et al, 2018]; [Bruce and Todt, 2018]). Despite these advances, the degree of hearing preservation varies significantly
across individuals ([Gantz et al, 2016]; [Helbig et al, 2016]; [Hunter et al, 2016]; [Moteki et al, 2017]).
Postoperative acoustic hearing loss typically falls into two categories: immediate
and delayed. Immediate-onset hearing loss that is detected at the first postoperative
appointment is commonly attributed to surgical trauma or an acute inflammatory process
([Eshraghi et al, 2005]; [Carlson et al, 2011]; [Seyyedi and Nadol, 2014]). Presently, the status of acoustic hearing after CI is unknown before the activation
audiogram typically completed three weeks postoperatively. Electrocochleography has
been proposed for real-time feedback during electrode insertion, but it is not available
with all electrode designs or precisely correlated with hearing preservation outcomes
([Kim et al, 2017]; [O’Connell et al, 2017a],[b]). The causes of delayed-onset hearing loss have yet to be elucidated, but it has
been hypothesized that a foreign body reaction to the electrode array may be involved
([Seyyedi and Nadol, 2014]).
The electrically evoked compound action potential (ECAP) measured intraoperatively
is one of the first markers of auditory nerve function after cochlear implant surgery.
The ECAP represents a synchronous response from electrically stimulated auditory nerve
fibers, providing information regarding the status of the auditory nerve. ECAP measurements
are used intraoperatively to confirm auditory nerve, device integrity, and electrode
functionality and postoperatively by audiologists for speech processor programming.
ECAP measurements have been correlated with both detection thresholds (T-levels) and
maximum comfortable loudness (C-levels), with a greater correlation found with T-levels
([Brown et al, 1996]; [1998]; [2000]; [Hughes et al, 2000]; [Franck and Norton, 2001]; [Abbas et al, 2017]). The ECAP maximum amplitude has been linked to speech perception scores after CI
([Kim et al, 2010]; [2017]; [Schvartz-Leyzac and Pfingst, 2018]). In addition, patients with low behavioral audiologic thresholds and larger ECAP
amplitudes are more likely to have higher speech perception scores ([DeVries et al, 2016]).
Although postoperative ECAP measurements have been previously associated with speech
perception scores, the clinical utility of intraoperative ECAP measurements have been
limited to date. In patients with residual acoustic hearing in the apical cochlea,
the ECAP response can theoretically originate from stimulation of the spiral ganglion
cells in the modiolus and from direct intracochlear electrical stimulation of inner
hair cells in the scala media. Auditory stimulation resulting from electrical stimulation
of spiral ganglion cells is often referred to as electroneural hearing, whereas auditory
stimulation resulting from electrical stimulation of inner hair cells is referred
to as electrophonic hearing. Several studies have documented electrophonic auditory
activity with intact intracochlear structures such that the electrical stimulation
resulted in basilar membrane mechanical activation with peak activity corresponding
to the frequency of the stimulus ([Kirk and Yates, 1994]; [Nuttall and Ren, 1995]; [Xue et al, 1995]; [Sato et al, 2016]). Thus, in this patient population with preserved low-frequency acoustic hearing
and, hence, preserved apical inner hair cells, ECAPs are likely resulting from both
electroneural and electrophonic stimulation.
Although routinely obtained for cochlear implant patients at many centers, the role
of intraoperative ECAP measurements for traditional and hearing preservation CI has
not been previously reported. This study investigates the association between intraoperative
ECAP measurements after electrode insertion and postoperative audiologic outcomes
in patients with and without residual low-frequency acoustic hearing. Our primary
hypothesis was that patients with more robust intraoperative ECAP responses (e.g.,
lower ECAP thresholds and higher ECAP amplitudes) would have better rates of postoperative
acoustic hearing preservation.
MATERIALS AND METHODS
Patient Selection and Clinical Information
Institutional review board approval was obtained before initiation of the study. Adult
patients were included if they (a) underwent CI via mastoidectomy and facial recess
approach, (b) had intraoperative ECAP measurements, and (c) had postoperative activation
audiogram, confirming the presence or absence of residual low-frequency unaided air
conduction thresholds. Both hearing preservation candidates and traditional candidates
were included. Patients were excluded if records were incomplete, with one exception:
some patients only had available ECAP thresholds or amplitudes, and these patients
were included for their respective analyses. Notably, all patients were implanted
using a “soft surgical” technique to preserve cochlear structural preservation, irrespective
of the preoperative hearing preservation status. A round window approach was always
preferred and attempted, with an extended round window or a cochleostomy approach
performed as an alternative if a round window approach was not feasible. Intraoperative
ECAP measurements were recorded before the termination of general anesthesia using
a clinical protocol that attempted to deliver similar levels of charge across all
patients. ECAPs were recorded via standard clinical software for all implant manufacturers
using biphasic pulses with a monopolar electrode configuration. Stimulation rates
across manufacturers were 32, 80, and 80 pulses per second for Advanced Bionics (Valencia,
CA), Cochlear (Englewood, CO), and MED-EL (Innsbruck, AT), respectively. Pulse durations
were 32, 25, and 30 μsec per phase for Advanced Bionics, Cochlear, and MED-EL, respectively.
ECAP parameters were not chosen a priori, but rather reflect data collected per clinical
protocol, allowing this retrospective review. For Advanced Bionics, we collect intraoperative
ECAP data on odd electrodes plus electrode 16 using stimulation levels 100 to 500
CU, in 100-CU steps. For Cochlear, we collect intraoperative ECAP data on odd electrodes
using stimulation levels 190 to 230 CL, in 10-CL steps. For MED-EL, we collect intraoperative
ECAP data on odd electrodes plus electrode 12 using stimulation levels 0 to 1200 CU,
in 300-CU steps. Thus, the maximum number of points in the ECAP amplitude growth function
was five for all three manufacturers.
For reporting purposes, present levels at the ECAP threshold were first converted
from clinical levels to microamperes (μA) and ECAP thresholds were then determined
via linear regression of the ECAP amplitude growth function for each manufacturer.
Maximum ECAP amplitudes in microvolts (μV) were recorded from the intraoperative responses
using similar across-subject maximum stimulation levels per the institution’s clinical
protocol which is standardized for a given implant manufacturer. For Advanced Bionics,
Cochlear, and MED-EL, the maximum stimulation level used for intraoperative ECAP testing
is 1,216, 1,114, and 1,200 μA, respectively. The equivalent charge per phase for these
levels and corresponding pulse durations was 38.9, 27.9, and 36.0 nC for Advanced
Bionics, Cochlear, and MED-EL, respectively. ECAP amplitudes were reported for the
maximum stimulation levels used. Postoperative audiograms were completed at the initial
activation audiology appointment, approximately three weeks after surgery.
Preservation of postoperative low-frequency hearing was defined as having an air conduction
audiometric threshold at 250 Hz ≤90 dB HL. This threshold was chosen because the target
gain with acoustic amplification is theoretically achievable when figuring a half-gain
rule for acoustic amplification and typical low-frequency gain limits for conventional
hearing aids (range 40–45 dB). This frequency (250 Hz) was chosen for the following
reasons: (a) it is the lowest frequency for which clinicians can verify hearing aid
output for various prescriptive fitting targets and, thus, serves as a marker for
functional residual hearing, and (b) acoustic hearing low-pass filtered at 250 Hz
is the minimum bandwidth for which a significant additive benefit can be derived from
the addition of acoustic hearing from either the implanted or non-implanted ear in
adult cochlear implant recipients ([Keidser et al, 2011]; [Zhang et al, 2013]; [Sheffield and Gifford, 2014]; [Sheffield et al, 2015]). Although 90-dB HL was our criterion threshold for acoustic hearing preservation;
this does not mean that we would provide acoustic amplification for frequencies with
thresholds up to 90-dB HL for these individuals. Rather, the preservation of acoustic
hearing in the implanted ear was chosen as a surrogate marker for cochlear structural
preservation. Although we recognize that we cannot guarantee cochlear structural preservation,
one could reasonably argue that structural preservation must be present—at least to
some degree—for patients with residual acoustic hearing after CI.
Statistical Analysis
Analyses were performed with GraphPad Prism 7.0 (GraphPad Software, La Jolla, CA)
and SPSS Statistics version 25 (IBM, Armonk, NY). We first completed a linear mixed
model test of interaction effects for the following four variables: manufacturer,
electrode location (basal, middle, and apical), preoperative low-frequency pure-tone
average (LFPTA), and postoperative LFPTA. There were no statistically significant
four-way interactions among the aforementioned variables for both the ECAP threshold
[F
(254, 8) = 1.2, p = 0.44, ηp
2 = 0.97] and the ECAP amplitude [F
(159, 8) = 0.3, p = 0.99, ηp
2 = 0.86]. In the absence of interaction effects, we analyzed electrode location and
residual low-frequency data individually for both ECAP threshold and ECAP amplitude
using multiple t-tests, without assuming a consistent standard deviation (SD) between the groups.
Correction of alpha for multiple comparisons was conducted using the Holm–Sidak method,
and the effect size was measured using Cohen’s d (d). Means and SDs for outcomes reported herein are representative of all included patients,
whereas statistical comparisons of ECAP measurements reflect values for only those
patients who had data available for the specified ECAP parameter. Nominal data were
analyzed using a Fisher exact or chi-squared test. p values <0.05 were considered statistically significant.
RESULTS
Two hundred fifty participants, with and without preoperative low-frequency acoustic
hearing, who underwent CI with intraoperative ECAP measurements between 2011 and 2016
were identified. Patients were excluded for incomplete audiologic data (n = 21), history
of prior middle-ear surgery (n = 8), and cochlear ossification (n = 4) ([Figure 1]). Ultimately, 217 participants were included in the analysis, 79 with postoperative
residual low-frequency hearing and 138 control participants without postoperative
residual acoustic hearing, as previously defined. Females accounted for 44.3% and
47.1% of the patients with and without postoperative residual low-frequency hearing,
respectively (p = 0.8900) ([Table 1]). The median age was 66.1 and 66.5 years for patients with and without postoperative
residual low-frequency hearing, respectively (p = 0.9193). Electrode insertion was conducted through one of the three surgical approaches:
round window, extended round window, or cochleostomy; round window insertion was pursued
if anatomy allowed. The cohort with postoperative residual low-frequency hearing had
a larger portion of round window insertions than the cohort without postoperative
residual low-frequency hearing (81.0% and 65.2%, respectively, p = 0.0097). The cohort without postoperative residual low-frequency hearing had a
larger portion of cochleostomy approaches (21.7%) than the cohort with postoperative
residual low-frequency hearing (8.9%, p = 0.0103). The distribution of surgeons between the two groups was not statistically
different (analysis of variance, p = 0.687). The cohort with postoperative residual low-frequency hearing had an average
postoperative LFPTA of 76.5 dB; specifically, 62.3, 75.8, and 91.2 dB at 125, 250,
and 500 Hz, respectively. The cohort without postoperative residual low-frequency
hearing had an average postoperative LFPTA of 100.6 dB; specifically, 90.7, 102.1,
and 106.5 dB at 125, 250, and 500 Hz, respectively. Patients received implants from
one of the three manufacturers: Advanced Bionics (Valencia, CA), Cochlear Americas
(Englewood, CO), or MED-EL (Innsbruck, AT). The cohort with postoperative residual
low-frequency acoustic hearing had a larger portion of MED-EL electrodes than the
cohort without postoperative residual low-frequency hearing (53.2% and 37.0%, respectively,
p = 0.0230). Notably, the preoperative LFPTAs between the three manufacturers were
not statistically different (analysis of variance, p = 0.178).
Figure 1 Study design.
Table 1
Patient Demographics and Surgical Factors
|
Total
|
Residual Hearing
|
No Residual Hearing
|
p Value
|
|
n
|
217
|
79
|
138
|
-
|
|
Female (%)
|
46.1
|
44.3
|
47.1
|
0.8900
|
|
Age (years)
|
66.4
|
66.1
|
66.5
|
0.9193
|
|
Surgical approach (%)
|
|
Round window
|
71.4
|
81.0
|
65.2
|
0.0097
|
|
Cochleostomy
|
17.1
|
8.9
|
21.7
|
0.0103
|
|
Extended round window
|
11.5
|
10.1
|
13.0
|
0.5282
|
|
Implant manufacturer (%)
|
|
MED-EL
|
42.9
|
53.2
|
37.0
|
0.0230
|
|
Cochlear America
|
37.8
|
29.1
|
42.8
|
0.0584
|
|
Advanced Bionics
|
19.4
|
17.7
|
20.3
|
0.7227
|
Note: Statistically significant p values are in bold.
For all analyses, ECAP measurements were categorized by electrode location (apical,
middle, or basal), and values at each electrode were averaged within each group. The
average ECAP amplitude measurements (SD) for apical, middle, and basal electrodes
were 490.73 μV (429.54 μV), 348.63 μV (303.36 μV), and 243.36 μV (296.54 μV), respectively,
for the residual low-frequency hearing cohort, and 235.64 μV (300.172 μV), 213.23
μV (274.27 μV), and 183.13 μV (250.62 μV), respectively, for the non-residual low-frequency
hearing cohort ([Figure 2]). For the apical and middle electrode ECAP amplitudes, there was a significant difference
between groups with and without residual low-frequency hearing (p = 0.0001, d = 0.71 and p = 0.0088, d = 0.47, respectively), whereas the basal electrode ECAP maximum amplitudes were not
different between groups (p = 0.2180, d = 0.22).
Figure 2 ECAP maximum amplitude in patients with and without postoperative residual low-frequency
hearing. Statistically significant p values are marked with asterisks (*).
A similar analysis was conducted between the cohorts with and without residual low-frequency
hearing for ECAP thresholds. The average ECAP thresholds measured intraoperatively
for apical, middle, and basal electrodes were 330.95 μA (154.15 μA), 433.14 μA (172.12
μA), and 452.36 μA (242.64 μA), respectively, for the residual hearing cohort, and
400.58 μA (206.97 μA), 462.66 μA (197.12 μA), and 482.83 μA (193.44 μA), respectively,
for the non-residual hearing cohort ([Figure 3]). For the apical region, ECAP thresholds were significantly lower in the residual
low-frequency hearing group (p = 0.0099, d = 0.37), whereas there was no significant difference between groups for the middle
and basal electrode regions (p = 0.2640, d = 0.16 and p = 0.3225, d = 0.14, respectively).
Figure 3 ECAP threshold in patients with and without postoperative residual low-frequency
hearing. Statistically significant p values are marked with asterisks (*).
Of note, a statistical analysis of patients divided into three groups ([a] patients
without preoperative low-frequency hearing, [b] patients with preoperative low-frequency
hearing who experienced preserved low-frequency hearing postoperatively, and [c] patients
with preoperative low-frequency hearing who did not experience preserved low-frequency
hearing postoperatively) was conducted with similar results as the abovementioned
analysis. A comparison of the patients without preoperative low-frequency hearing
to the patients with preoperative low-frequency hearing that preserved low-frequency
hearing showed a statistically significant difference for ECAP maximum amplitude in
the apical and middle regions (p = 0.0017 and p = 0.0119, respectively), but no statistically significant difference in the basal
region (p = 0.1100). ECAP thresholds were not statistically significant between these two groups
for apical, middle, or basal electrodes (p = 0.1230, p = 0.3296, and p = 0.2373, respectively).
DISCUSSION
The primary objective of this study was to evaluate the association between intraoperative
ECAP measurements at the time of cochlear implant surgery and postoperative residual
low-frequency acoustic hearing. The abovementioned findings support our hypothesis
that patients with more robust intraoperative ECAP responses (e.g., lower ECAP thresholds
and higher ECAP amplitudes) would have better rates of postoperative acoustic hearing
preservation. Specifically, patients with postoperative residual low-frequency hearing
did exhibit significantly larger maximum ECAP amplitudes for apical and middle electrodes
and lower ECAP thresholds for apical electrodes during intraoperative ECAP testing.
These findings were observed when comparing ECAP data for patients with considerable
preoperative hearing (i.e., hearing preservation candidates) with both a group of
patients who did not have viable preoperative hearing for preservation and hearing
preservation candidates who ultimately did not have hearing preservation.
Previous studies provide some insight into the association between hearing preservation
and ECAP amplitudes in the middle and apical electrodes. Maximum amplitudes may serve
as a marker for neural health, as larger amplitudes are positively correlated with
spiral ganglion cell density ([Hall, 1990]; [Ramekers et al, 2014]). Correspondingly, ECAP amplitudes were found to be consistently smaller after deafening
in animal models ([Shepherd and Javel, 1997]; [Agterberg et al, 2009]). Our study did not reveal an association between ECAP amplitude in the basal electrodes
and hearing preservation; this may be related to the tonotopic organization of the
cochlea, as cochlear structures responsible for preserved low-frequency acoustic hearing
are generally located at the cochlear apex. Furthermore, most patients included did
not have preoperative acoustic hearing in the basal cochlea, and we theorize that
all ECAP responses resulting from intracochlear electrical stimulation in the basal
cochlea resulted from electroneural stimulation, and thus, we would not expect there
to be differences between groups. By contrast, for patients with preoperative acoustic
hearing in the apical cochlea, we hypothesized that a higher ECAP amplitude measured
at the time of electrode placement would be an indicator of residual hearing, which
could be inferred as resulting from an atraumatic surgical insertion. Although the
present data seem to support this hypothesis, verification requires animal studies
to document the degree of intracochlear insertion trauma, intraoperative ECAP amplitude,
and subsequent histologic analysis.
The findings of this study, taken into context of previous reports, suggest that the
intraoperative ECAP maximum amplitude, a marker for neural health, may function as
an indicator of intraoperative neural injury. Future studies will search for a specific
numerical value or cutoff of ECAP maximum amplitude at which hearing preservation
surgery is most likely successful. Intraoperative ECAP measurements could eventually
serve as immediate feedback for the surgeon regarding surgical technique and likelihood
for acoustic hearing preservation; however, additional prospective studies are required
before application in clinical practice.
ECAP thresholds, and their association with residual low-frequency hearing, are not
completely understood. Lower ECAP thresholds have been associated with a shorter distance
between the electrode and modiolus ([Gordin et al, 2009]; [Davis et al, 2016]). It follows that lower apical thresholds in our study would correlate with improved
proximity to the modiolus, although it is not clear whether better perimodiolar placement
would necessarily be related to higher rates of hearing preservation, unless perimodiolar
placement reflected a complete scala tympani insertion, as scala tympani electrode
location has been associated with increased rates of hearing preservation ([O’Connell et al, 2016]; [2017a],[b]). Further research and electrode imaging to verify scalar location are warranted.
Because of the nature of retrospective investigation, this study has inherent limitations,
and thus, findings would benefit from confirmation through a controlled, prospective
study. Specifically, a comparison of surgical approaches between the postoperative
residual and non-residual hearing groups revealed a greater number of cochleostomy
approaches in the non-residual hearing group. Although the cochleostomy approach has
been associated with lower rates of hearing preservation than a round window approach
([Wanna et al, 2018]), it is possible that a cochleostomy was chosen more commonly for traditional cochlear
implant candidates or in more challenging cases. In addition, a comparison of manufacturers
revealed a greater number of MED-EL electrodes used in the residual low-frequency
group, despite no statistically significant difference in preoperative LFPTAs across
manufacturers; future prospective studies will focus on controlling for manufacturer
and insertion approach; however, evaluation of hearing preservation rates were outside
the scope of this study. Of note, pulse durations used for ECAP software are different
among manufacturers. In this study, stimulation levels were converted into charge
units, which were relatively comparable across manufacturers, although not identical.
Although the ECAP stimulation and acquisition parameters are inherently different
across the manufacturers, future research should attempt to equate stimulation levels
in charge across the manufacturers, allowing for more accurate across-device comparison.
Using current-generation software and hardware, with pulse durations of 32, 25, and
30 μsec per phase for AB, Cochlear, and MED-EL, respectively, one could set the upper
stimulation level for intraoperative ECAP at 400 CU for AB, 240 CL for Cochlear, and
1100 CU for MED-EL, and this would equate the stimulation level in charge at ∼33 nC
per phase for each manufacturer.
The retrospective study design also limits the evaluation of potential confounders
such as acute inflammation and surgical complications, which may occur between the
time of the intraoperative ECAP measurements and postoperative audiologic testing.
These events may impact audiologic outcomes in a manner that is not captured in this
study. Last, this study was not designed as a prospective hearing preservation study
to investigate hearing preservation rates or techniques at our center, as patients
with and without preoperative low-frequency hearing were included. Thus, the conclusions
drawn should be interpreted with caution until prospective studies investigating ECAP
as a part of hearing preservation protocol are performed.
As a result of minimally traumatic insertion techniques and advances in electrode
design, cochlear implant indications have expanded to include patients with residual
acoustic hearing. Hearing preservation outcomes vary significantly and the status
of residual low-frequency hearing is unknown until the first postoperative audiologic
appointment. Intraoperative ECAP measurements may provide immediate insight to residual
low-frequency hearing outcomes; patients with larger intraoperative ECAP maximum amplitudes
and lower ECAP thresholds are more likely to have postoperative residual low-frequency
acoustic hearing at their initial postoperative evaluation versus patients who are
classical candidates or those who have low-frequency threshold shift. Future prospective
studies will focus on clinical applications of intraoperative ECAP measurements and
the potential use of ECAP values as real-time feedback for surgeons.
CONCLUSIONS
The present study demonstrates that patients with postoperative residual low-frequency
acoustic hearing exhibited greater ECAP amplitudes for apical and middle electrodes
and lower thresholds at the apical electrodes. This finding is attributed to the preservation
of intracochlear neural substrate, namely, inner hair cells, with atraumatic electrode
insertion and the resulting ECAP responses resulting from both electrophonic and electroneural
stimulation. Indeed, the association between ECAP measurements and residual low-frequency
hearing may represent a potential immediate feedback mechanism for postoperative outcomes
that can be applied to all CIs. Additional animal studies could prove useful for verification
of these data following surgical insertion, ECAP assessment, and subsequent histology.
Abbreviations
CI:
cochlear implantation
ECAP:
electrically evoked compound action potential
LFPTA:
low-frequency pure-tone average
SD:
standard deviation
