Active middle ear implants (AMEIs) are widely used when hearing restoration with passive ME implants is limited and/or conventional hearing aids do not contribute to a sufficient hearing improvement (
). Lailach et al. 2019
Primarily, AMEIs have been developed for actuator coupling to the intact ossicular chain (OC) (especially to the incus) and thus for hearing rehabilitation in patients with sensorineural hearing loss. The development of coupling elements in analogy to passive ME implants as well as the verification of coupling to the RW membrane for “reverse stimulation” of the cochlea led to a broader spectrum of indications for patients with a defective OC and conductive or mixed hearing loss.
During the last decade, several generations of couplers have been introduced. Experimental human temporal bone (TB) studies have described the transmission of the floating mass transducer (FMT) (
Vibrant Soundbridge [VSB], MED-EL, Innsbruck, Austria) depending on the actuator coupling structure ( Colletti et al. 2005 ; Bornitz et al. 2010 ; Beleites et al. 2011 ; Bornitz 2011 ; Hüttenbrink et al. 2011 ; Beleites et al. 2014 ; Müller et al. 2017 ; ), the coupler generation ( Bornitz et al. 2021 ) or the residual ventilation of the ME ( Mlynski et al. 2015b ). Long-term clinical studies have investigated the postoperative hearing outcome after implantation ( Müller et al. 2019 Schmuziger et al. 2006 ; Baumgartner et al. 2010 ; ). However, the current literature still lacks the link between clinical data and TB experiments. Rahne et al. 2021
After treatment with AMEI, the postoperative speech comprehension of the patients is highly heterogenous. The broad interindividual spectrum leads to the question of appropriate preoperative and intraoperative predictors. One possible influencing factor to be evaluated is the coupling of the actuator. A stable connection between actuator and anatomical structure is the precondition for good signal transmission overall frequencies. A coupling deficit leads to higher hearing thresholds, deficient speech comprehension, or even reduced signal quality. While a coupling deficit can often be compensated by the amplification performance of the implant in cases of low-grade bone conduction (BC) component, patients at the lower edge of the indication spectrum dispose only of a low dynamic range so that a nonoptimal coupling is more important and leads to an unsatisfactory audiological result.
The aim of this study was to correlate clinical data on postoperative hearing outcome after vibroplasty with different latest generation of couplers—Clip coupler, incus short process (SP) coupler, incus long process (LP) coupler, round window (RW) soft coupler—in a large cohort with experimental data obtained by human TB experiments for the first time. A major aim is to facilitate the preoperative selection of a suitable coupler and to estimate the expected postoperative hearing outcome based on the preoperative audiological findings. The results should give an important contribution to patient care and patient safety to be able to advise patients as best as possible preoperatively about the opportunities and risks within the framework of optimal hearing rehabilitation. Furthermore, the study should address the question to which extent human TB experiments are suitable for assessing clinical audiological outcomes and which limitations there are in the correlation with clinical audiological data.
MATERIAL AND METHODS
Experimental Part: Assessment of the Transmission of 32 Human TB After Vibroplasty
A total of 32 human TB were extracted at the latest 10 days post mortem. Ethical approval for the use of human TB specimen has been obtained by the local ethics committee (EK 59022014). See
Table 1 for details on the specimens and the distribution within the individual target structure subgroups. Details on TB preparation have already been presented by ). Important details are described below. Müller et al. (2019
TABLE 1. -
Statistical details of the harvested specimen and the distribution within the individual coupling site subgroups
Right Ears/Left Ears
Average Age of Donors ± SD, yrs
Gender: Male Donors/Female Donors
61.3 ± 13.0
63.6 ± 9.3
51.4 ± 13.9
59.3 ± 17.1
*A total of 32 specimen were harvested. Some of the temporal bones were used for multiple measurements with different target structures.
After the extraction, all specimens were immediately frozen at −24°C until final preparation. During preparation and measurements, the specimens were stored in isotonic saline solution at 4°C. In all TBs, a mastoidectomy combined with an extended antrotomy and a posterior tympanotomy was performed. To facilitate the access to the stapes footplate, the facial nerve was partially resected. For laser Doppler vibrometry measurements, a foil with reflective polystyrene microbeads (approximately 0.5 mm²) was placed in the center of the footplate. A hole was drilled into the anterior wall of the external auditory canal in which a probe microphone was inserted and positioned above the umbo. A silicone tube was glued into the lateral end of the canal, in which an earphone could be inserted.
Experimental Setup and Data Processing
Measurements of the ME transfer function (METF) were obtained for the intact OC and for different reconstructions of the ME with different couplers (MED-EL, Innsbruck, Austria)—clip coupler, RW soft coupler, SP coupler, LP coupler—by means of laser Doppler vibrometry (Polytec GmbH, Waldbronn, Germany). METFs were measured with acoustic and mechanical excitation. The experimental and the analysis setup has already been described in detail (
). Important details are also given below. Müller et al., 2019
For acoustic excitation, an audiometric earphone (Telex-1470, Telex communications) was attached to the silicone tube in the external ear canal. The sound pressure was measured with a probe microphone (ER-7C, Etymotic Research, Elk Grove Village, IL) about 3 mm in front of the tympanic membrane (TM). For mechanical excitation of the OC, the FMTs were coupled to the OC.
The velocity of the stapes footplate was measured using a laser Doppler vibrometer (CLV 700 laser head and CLV 1000 controller unit, Polytec, Waldbronn, Germany). The laser beam was focused to the center of the stapes footplate. The excitation signal was a multisinus signal in the frequency range of 0.05 to 6 kHz with a resolution of 50 Hz. For acoustic excitation the equivalent sound pressure level of the signal was 95 to 105 dB SPL. For mechanical excitation, we used a driving voltage of the FMT resulting in a similar stimulation of the ME as acoustic excitation. The METFs were obtained in the frequency domain as complex averages of mostly 20 single measurement frames. METFs were either obtained as stapes displacement with reference to the sound pressure at the TM or as stapes displacement with reference to the driving voltage of the FMT (
). Müller et al. 2019
For data evaluation, the METF of the intact ME (METF
I) were used as baseline. Transfer functions with mechanical excitation (METF FMT) were displayed with reference to the corresponding METF I:
M E T F F M T r e l = 20 × l g ( M E T F F M T M E T F I )
[ M E T F F M T ] = n m ╱ V
[ M E T F I ] = n m ╱ P a
[ M E T F F M T r e l ] = d B ( r e f . 1 P a ╱ V .
These relative transfer functions allow for better comparison between specimens.
The data of the RW coupling have been corrected with regard to the reverse stimulation of the cochlea according to the investigations of
) Details are described in Supplemental Digital Content 1, Stieger et al. (2013 and Supplemental Digital Content 2, https://links.lww.com/EANDH/B34 ). https://links.lww.com/EANDH/B35 Experimental Protocol
Step 1: Measurements of transmission at intact OC conditions (METF
I of each TB was measured for the intact ME using sound pressure excitation.
Step 2: Measurements of transmission after OC reconstruction (METF
The FMT of the VSB device was coupled to the OC or the RW with Clip coupler, SP and LP coupler, and RW soft coupler. Transmission after reconstruction was measured using mechanical excitation with the FMT, where the driving voltage was used as reference (METF
FMT). Statistical Analysis
Statistical analyses were conducted using OriginPro (OriginLab Corporation, Northampton, MA). Each subgroup (LP coupler, SP coupler, Clip coupler, and RW soft coupler) can be considered as an independent sample.
Explorative data analyses were initially performed for all measurements. The mean values, the standard deviations (SDs) and the 95% confidence intervals (CIs) of the mean were calculated.
Normal distribution of all data of all subgroups was tested and confirmed graphically in the mean speech range (0.5 to 4 kHz) using histogram. A parametric t-test was used to compare the means of two subgroups (SP coupler and LP coupler or Clip coupler and RW soft coupler) in the mean speech range respectively. A
p value of less than 0.05 was considered significant. Also “inference by eye” ( ) was used to compare the means of two subgroups based on the 95% CI of the means. If the proportion overlap of the 95% CI was less than 0.5, significant differences of the means could be assumed ( Cumming & Finch 2005 p < 0.05). Clinical Part: Assessment of 69 Patients After Vibroplasty
The study was proven and accepted by the local ethical review committee (EK 314082018).
Chart Review and Patients Selection Criteria
A retrospective chart review was performed. Based on the eligibility criteria for VSB implantation, patients aged 5 years and older were included in the analysis. All patients implanted with a Vibrating Ossicular Prosthesis (VORP) 502 or VORP 503 (VSB) in combination with the LP coupler, SP coupler, Clip coupler, or RW soft coupler at tertiary referral center between 2009 and 2019 were screened.
The coupling of the FMT was performed depending on the individual pathology. If the OC was intact and mobile, the FMT was coupled to the incus using an SP or LP coupler. The choice of SP or LP coupler was made depending on the surgeon’s preferences, as no recommendations regarding the best possible form of coupling existed for incus coupling until now. If the chain was defective and the stapes superstructure was present and the footplate was mobile, the FMT was coupled to the stapes head using a Clip coupler. If the chain was fixed or the stapes superstructure was missing, the FMT was coupled to the RW.
Subjects were included irrespective of the specific audio processor variant used (Samba, Amadé). Due to the retrospective design of the study, 18 patients without a complete data set were excluded. The minimum data requirement was the availability of preoperative pure-tone audiogram, aided sound field thresholds, vibrogram thresholds, and aided WRS at least 6 months postoperatively.
Audiological measurements were performed in an anechoic soundproofed room (DIN EN ISO 8253). Routine pure-tone audiometry was performed preoperatively and 6 months after surgery. The air conduction thresholds were measured at frequencies from 0.125 to 8 kHz with an AT900 clinical audiometer (Auritec GmbH, Germany). BC hearing thresholds were determined at frequencies between 0.125 and 6 kHz using the KLH 96 transducer (CB-Elmec GmbH, Germany).
At the implanted side, free-field (FF) thresholds were measured using narrow-band noise with center frequencies from 0.25 to 8 kHz. The pure-tone average (4PTA) was calculated across the frequencies 0.5, 1, 2, and 4 kHz. The effective gain (EG) was defined as difference between the aided 4PTAFF and the postoperative 4PTABC. The word recognition score (WRS) was measured preoperatively (unaided) and 6 months postoperatively at a fixed sound pressure level of 65 and 80 dB SPL using the Freiburger monosyllable word recognition test. The contralateral ear was plugged with a conventional earplug (single number rating [SNR] = 27 dB) and covered with an earmuff (Peltor Optime III H540A, SNR = 35 dB).
The measurement of the vibroplasty thresholds was similar to pure-tone audiometry. The difference was that the stimulus was presented via the FMT. Vibroplasty thresholds were determined by applying the Hughson-Westlake method. Results were reported on a decibel scale. Reference is the excitation voltage of the FMT at hearing threshold for a normal hearing subject and the FMT coupled to the long incus process (LP coupler). The pure-tone average (4PTA Vibroplasty) was calculated across the frequencies 0.5, 1, 2, and 4 kHz. The coupling efficiency was calculated by subtracting the vibroplasty threshold and the postoperative BC threshold.
Initial activation and fitting of the VSB was performed 6 to 8 weeks postoperatively. The initial fitting based on individual vibroplasty thresholds. Using these data, the default setting was based on DSL [i/o]. Subsequent fitting was performed manually, including repeated threshold measurements, speech audiometric measurements, and patients’ evaluation of speech comprehension.
The statistical analysis was performed with IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY). Since a normal distribution could not be confirmed, nonparametric tests were used. The Wilcoxon test was used to test between pre- and postoperative audiological outcome parameters. The Kruskal–Wallis test was used to compare BC thresholds, FF thresholds, WRS, EG, and coupling efficiency between different coupling methods. Correlations were described by Spearman’s correlations coefficient. The correlation coefficients were evaluated as follows: 0.90 to 1.00 (−0.90 to −1.00) very-high-positive (negative) correlation, 0.70 to 0.90 (− 0.70 to − 0.90) high-positive (negative) correlation, 0.50 to 0.70 (−0.50 to −0.70) moderate-positive (negative) correlation, 0.30 to 0.50 (−0.30 to −0.50) low-positive (negative) correlation, and 0.00 to 0.30 (0.00 to −0.30) negligible correlation (
). Descriptive data are expressed as mean ± SD and 95% CI. A Hinkle et al. 2003 p value of less than 0.05 was considered significant. RESULTS
Transmission of the Intact OC
The METF with intact OC (
Fig. 1) was found to be within the borders of the ASTM ( Rosowski et al. 2007 ; ) for almost every single specimen. There are few specimens in each subgroup whose METFs are beyond the borders of the ASTM. Considering that no cosine correction of the data was performed, which would add 2 to 4 dB, the mean METF of each subgroup and the corresponding 95% CI are found to be within the borders of the ASTM. The mean METFs of all subgroup are comparable with each other with regard to the trend of the graphs and absolute values. Each specimen and each subgroup could be approved for subsequent evaluations. ASTM 2014 Fig. 1.:
METF with intact ossicular chain. LP coupler subgroup (n = 10), SP coupler subgroup (n = 8), RW soft coupler subgroup (n = 10), and clip coupler subgroup (n = 11). Mean amplitudes and 95% CIs of mean are shown in comparison to mean and 95% CI of the ASTM standard. The subgroups are comparable, the METF of every single specimen can be accepted for further measurement based on the borders of the ASTM (
Rosowski et al. 2007 ; ASTM 2014
). CI, confidence interval; LP, long incus process; METF, middle ear transfer function; RW, round window; SP, short incus process.
Transmission After Vibroplasty
Figure 2 depicts the characteristical transmission of the FMT after vibroplasty for all target structures. The FMT acts like a force-limited actuator without an abutment and is therefore only driven by its inertial load. The effective inertial load increases with the excitation frequency ( ). Therefore, the transmission of the FMT is insufficient in the low frequencies, whereas it shows a constant force in the high frequencies. The graphs of all target structures therefore show in general a similar course. Nevertheless, there are differences in the transmission among the individual couplers. These are discussed below. Note that the data of the RW soft coupler (bottom left) is corrected for reverse cochlear stimulation according to the data of Bornitz et al. 2021 ). In the reconstructions with the clip coupler, just the incus had been removed but the FMT usually had contact to the TM and/or malleus handle. Stieger et al. (2013 Fig. 2.:
Transfer function of the different couplers after vibroplasty. Upper left, LP coupler subgroup (n = 10), (upper right) SP coupler subgroup (n = 8), (bottom left) RW soft coupler subgroup (n = 10), and (bottom right) Clip coupler subgroup (n = 11). Every graph of each subgroup depicts the characteristic transfer function of the FMT with an insufficient transmission in the lower frequencies and high output in the higher frequencies with a maximum in the range of the resonance frequency at about 1.2 kHz. Note that the data of the RW stimulation (bottom right) are corrected for reverse stimulation of the cochlea according to the investigations of
Stieger et al. (2013
). LP, long incus process; RW, round window; SP, short incus process.
Figure 3 displays on the left that the transfer function of the LP coupler exceeds up to 8 dB above the transfer function of the SP coupler in the low-frequency range. These amplitude differences are significant from 0.65 to 0.8 kHz ( p < 0.05). In the high frequencies, the transmission of the LP coupler is up to 10 dB below the transmission of the SP coupler. These data show no significance. Across the entire frequency range the transmission of the SP coupler remains at the level of the mean transfer function of all couplers. On the right, there are clear inhomogeneities depicted when comparing the amplitude differences of the transfer function between the SP coupler and the LP coupler in the own data to current literature. Only the combined data of ) and Mlynski et al. (2015b ) are in general comparable with our own data. Schraven et al. (2016b Fig. 3.:
Left, The transfer functions of SP and LP coupler are displayed as amplitude differences with reference to the mean transfer function of all couplers. Below 1 kHz, the LP coupler’s transmission exceeds those of the SP coupler by up to 10 dB. The depicted 95% CI show significance of these amplitude differences only in a narrow frequency range between 600 and 800 Hz, where the largest amplitude differences occur. The data do not show significance at smaller amplitude differences. The data of the SP coupler correspond approximately to the average transmission of all couplers. From 1 kHz, the transfer function of the LP coupler is up to 10 dB below the transfer function of the SP coupler whose amplitude still remains within the range of the mean amplitude of all couplers. These differences are not significant. Right, The amplitude differences of the transfer function of the SP coupler with reference to the LP coupler are depicted compared to current literature data. The data of Bornitz (
Bornitz et al. 2010 ; Bornitz 2011 ; Bornitz et al. 2021
) have been generated by means of a finite-element model, all other data (
Mlynski et al. 2015b ; Schraven et al. 2016b ; Raphael 2021
) have been obtained from human TB experiments. The recent literature displays a high level of inhomogeneity. This aspect will be discussed further. Only the data of Mlynski are in general comparable to the own data. However, these data show a considerably larger amplitude difference of up to 20 dB compared to the own data. CI, confidence interval; LP, long incus process; SP, short incus process; TB, temporal bone.
Figure 4 depicts that the transfer function of the RW soft coupler remains below the transfer function of the Clip coupler from 400 Hz across the entire frequency range with a maximum amplitude difference of 15 dB in the high-frequency range from 1 kHz. These amplitude differences are significant from 1.1 to 1.7 kHz ( p < 0.05). The amplitude of the Clip coupler meets the mean transfer function of all couplers in the low frequencies, whereas it ranges 5 to 7 dB above the average in the high-frequency range. Fig. 4.:
The transfer function of Clip and RW soft coupler are displayed as amplitude differences with reference to the mean transfer function of all couplers. Note that the own data of the RW soft coupler are corrected for reverse cochlea excitation. Below 1 kHz, the Clip coupler’s transmission exceeds those of the RW soft coupler of up to 10 dB. The data show no significance of these amplitude differences. In the frequency range from 1 to 4 kHz the Clip coupler shows a significantly improved (
Cumming & Finch 2005
) transfer function of up to 15 dB compared to the transfer function of the RW soft coupler. Above 400 Hz, the average transmission behavior of the RW soft coupler is below the average transmission behavior of all couplers. The transmission behavior of the Clip coupler is above the average of all couplers from 1000 Hz. The data of the RW soft coupler show a much wider CI then the data of the Clip coupler. This variation in the data shows that the transmission behavior of the RW soft coupler in particular is influenced by various external factors. These factors are described in the discussion. CI, confidence interval; RW, round window.
Between April 2009 and December 2019, 101 patients were implanted. Thirty-two patients were excluded due to different coupling types (19 patients) or incomplete data sets (13 patients). For this study, 69 subjects were selected. Demographic parameters were shown in
TABLE 2. -
Demographic data and statistical analysis of the clinical data
Age in Years, Mean ± SD (Range)
Kind of Hearing Loss
Unaided WRS 65 dB in %, Mean ± SD (95% CI)
Unaided WRS 80 dB in %, Mean ± SD (95% CI)
Aided WRS 65 dB in %, Mean ± SD (95% CI)
Aided WRS 80 dB in %, Mean ± SD (95% CI)
Postoperative BC (0.5–4 kHz) in dB, Mean ± SD (95% CI)
Aided Threshold(0.5–4 kHz) in dB, Mean ± SD (95% CI)
∆ vibrogram – BC (0.5–4 kHz) in dB, Mean ± SD (95% CI)
EG (0.5–4 kHz) in dB, Mean ± SD (95% CI)
LP (n = 8)
57.0 ± 18.5
SNHL (n = 8)
Malformation (n = 2)
Otitis externa (n = 6)
VORP 502 (n = 6)
VORP 503 (n = 2)
Amadé (n = 6)
Samba (n = 2)
32.1 ± 33.6
61.4 ± 17.5
75.0 ± 24.0
91.7 ± 7.9
38.2 ± 16.2
34.5 ± 5.7
8.9 ± 12.9
(−0.9 to 16.9)
−3.8 ± 14.8
SP (n = 9)
26.1 ± 26.2
SNHL (n = 2)
CHL (n = 6)
MHL (n = 1)
Malformation (n = 7)
COM (n = 1)
Otitis externa (n = 1)
VORP 503 (n = 9)
Samba (n = 9)
35.6 ± 34.8
61.1 ± 26.6
73.9 ± 20.6
93.6 ± 9.5
27.1 ± 14.0
31.7 ± 9.6
9.5 ± 6.5
4.6 ± 11.8
(−3.9 to 11.2)
Clip (n = 22)
50.1 ± 24.5
CHL (n = 5)
MHL (n = 17)
Malformation (n = 9)
COM (n = 13)
VORP 502 (n = 9)
VORP 503 (n = 13)
Amadé (n = 9)
Samba (n = 13)
37.5 ± 44.1
53.9 ± 38.4
66.8 ± 21.1
85.9 ± 19.4
36.1 ± 12.7
37.4 ± 8.6
5.2 ± 10.5
−0.31 ± 14.7
(−6.1 to 6.1)
RW (n = 30)
36.1 ± 23.5
CHL (n = 11)
MHL (n = 19)
Malformation (n = 6)
COM (n = 24)
VORP 503 (n = 30)
Samba (n = 30)
17.5 ± 28.5
41.4 ± 37.0
78.1 ± 15.1
92.3 ± 10.6
27.2 ± 13.1
33.1 ± 7.7
12.7 ± 11.0
5.9 ± 10.5
There were no significant differences regarding the postoperative speech intelligibility (aided WRS at 65 and 80 dB) among the coupling sites.
BC, bone conduction; CHL, conductive hearing loss; COM, chronic otitis media; EG, effective gain; LP, long incus process; MHL, mixed hearing loss; MW ± SD (95% confidence interval); RW, round window; SNHL, sensorineural hearing loss; SP, short process; WRS, word recognition score; VORP, vibrating ossicular prosthesis.
Complications and Revision Surgery
Four out of these 69 patients underwent revision surgery because of recurrent TM perforation or retraction. One patient who was implanted in 2012 had to be reoperated in 2019 because of an implant failure. One patient underwent cochlear implantation 5 years after VSB implantation due to an increasing sensorineural hearing loss.
Hearing Preservation BC (Preoperative Versus Postoperative)
The surgical procedure had only little effects on the BC (0.5 to 4kHz, PTA4). Mean preoperative BC was 29.3 ± 14.8 dB and increased significantly to 31.2 ± 14.4 dB 6 months after implantation (mean difference −1.9 ± 6.4 dB,
p < 0.05). Postoperative BC thresholds across the frequencies 0.5, 1, 2, and 4 kHz did not differ between the coupling groups ( p > 0.05, Table 2). EG and Coupling Efficiency
The EG was frequency dependent and tend to differ between the coupling groups (
Fig. 5). However, the differences failed to meet the defined significance level. Fig. 5.:
Postoperative effective gain (EG). The differences failed to meet the defined significance level. Tendentially, the lowest EG over the entire frequency range was observed in the group with the RW coupling. In contrast, superior effective gain was determined with LP coupling and with Clip coupling in the frequency range above 1.5 kHz. An overclosure was achieved in mid frequency range for the Clip coupling group (1.5–3 kHz) and for the LP group (1–3 kHz) and for the SP group (1.5–2 kHz). LP, long incus process; RW, round window; SP, short incus process.
The coupling efficiency did not show significant differences between the coupling modalities (
Fig. 6, p = 0.24). Fig. 6.:
Postoperative coupling efficiency. Higher values indicate a suboptimal coupling. In the RW coupling group, the difference between vibrogram and BC thresholds was larger compared to the other coupling modalities in the lower frequency range (0.5–1.5 kHz). However, the differences were not statistically significant. The best coupling efficiency was achieved in the Clip subgroup, especially above 2 kHz. LP, long incus process; RW, round window; SP, short incus process.
In patients with RW coupling, we additionally analyzed whether stapes footplate fixation had an effect on audiological performance. In 14 patients, coupling to the RW was performed because of a complete stapes fixation. 11 patients with RW coupling had a regularly mobile stapes footplate. Due to a facial prolapse stapes mobility could not be evaluated in five patients. Neither the coupling efficiency (11.8 ± 9.2 versus 12.6 ± 13.6 dB), nor the FF thresholds (34.9 ± 4.9 versus 31.6 ± 9.6 dB) and the EG (6.0 ± 11.1 versus 6.4 ± 7.1 dB) showed significant differences between the two groups for the PTA4 (0.5 to 4 kHz).
The WRS increased from 28.2 ± 37.2% to 73.3 ± 19.1% at 65 dB SPL (
p < 0.001) after implantation and from 52.5 ± 37.1% to 90.2 ± 11.9% at 80 dB SPL ( p < 0.001). There was no significant difference between the analyzed groups ( Table 2). In the group with RW coupling, there was no significant difference in WRS at 65 dB between the group with fixed stapes (75.4 ± 17.9%) and the group with movable footplate (77.7 ± 11.2%). Generally, there was a low negative correlation between the WRS at 65 dB SPL (aided) and the postoperative BC threshold ( r = −0.33, p < 0.01), and negligible correlations between the WRS 65 dB and the coupling efficiency ( r = −0.03, p > 0.05), and EG ( r = 0.05, p > 0.05) (visualization in Supplemental Digital Content 3, ). https://links.lww.com/EANDH/B36 DISCUSSION
LP Coupler Versus SP Coupler
The transmission of the electrodynamic force-driven FMT depends on the coupling position. The simulation data (
Fig. 3, right side, dashed red graph ( )) obtained by means of a finite-element model ( Bornitz et al. 2021 Bornitz et al. 2010 ; ) confirm the experiments of Bornitz & Zahnert 2013 ), according to which a vibration nodal line runs in the area of the malleus head and the incus body in a wide frequency range. Coupling of the FMT to the incus body bears the risk of exciting ineffective vibration patterns (rocking movements) of the stapes footplate. Increasing the distance of the FMT to the incus ( Ball et al. (1997 Fig. 3, right side, dotted red graph ( )) additionally worsens the transmission, especially in the resonance frequency range of 1 to 1.5 kHz, and stimulates rocking movements of the stapes footplate. Although recent studies support the importance of rocking movements to the formation of cochlear action potentials (CAPs) ( Bornitz et al. 2021 Huber et al. 2008 ; ), the current doctrine states that in this case CAP are only triggered at increased stimulus thresholds and with lower sensitivity. Piston-like movements remain dominant for the formation of CAP. Compared to the simulation, the experimental data show a significantly better transmission when coupling to the short incus process. This is mainly due to the fact that the investigators are aware of the vibration characteristics of the chain and that the best possible coupling for the analysis can be chosen during the experimental measurements by means of real-time monitoring. Eiber et al. 2012
The coupling state—“loose” or “tight”—also influences the transmission. Despite a supposedly tight contact, relative movements can occur between FMT, coupler and actuator target structure, which result in an increase of the total harmonic distortion (THD) or the distortion factor with a consequent loss of transmission (
Bornitz & Zahnert 2013 ; ). The elastic crimping used in earlier SP and LP couplers did not create an optimal form closure due to the spring mechanism. Compared to the free-floating FMT, the THD increased when crimping on the long incus process. Perfect positioning—tight crimping and contact of the FMT with the stapes head—also showed increased THD ( Beleites et al. 2014 ). The data of Beleites et al. 2014 ) ( Raphael (2021 Fig. 3, right side, dashed green graph) show that additional adhesive forces between bone and coupler are needed for optimal coupling. There is no difference in the transfer function if the FMT is glued either to the long or to the short incus process. Improved attachment can also be achieved with the help of surgical cement ( ). Snik & Cremers, 2004
The current generation of SP and LP couplers described by Schraven (Schraven, Mlynski, et al., 2016) and Mlynski (Mlynski, Dalhoff, Heyd, Wildenstein, Hagen, et al., 2015; Mlynski, Dalhoff, Heyd, Wildenstein, Rak, et al., 2015) are equipped with a clip mechanism. The clip SP and LP couplers display across almost the entire frequency range a better transfer function compared to their precursor with a crimping mechanism. However, the current Clip SP coupler depicts up to 20 dB better transmission from approximately 600 Hz in comparison to the present clip LP coupler (
Fig. 3, right side, dashed blue graph). The superiority of the clip SP over the clip LP coupler is also reflected in our own data starting from the resonance frequency range. However, the amplitude difference only amounts to a maximum of 10 dB. In our own experiments, we possibly succeeded in positioning the clip LP coupler reproducibly close to the stapes head and parallel to the longitudinal axis of the stapes.
The parallel alignment of the FMT to the longitudinal stapes axis (
z axis) guarantees an optimal transmission function. ) indicated that coupling at the short incus process by means of SP clip coupler without sufficient parallel alignment of the FMT leads to transmission losses of up to 20 dB in the frequency range from 2 to 5 kHz compared to parallel alignment of the actuator. Mlynski et al. (2015b
In our clinical investigation, a similar postoperative WRS was obtained when comparing the LP and SP groups. In former clinical analysis, inconsistent results were found when comparing these two coupling sites. While
) found a significantly better WRS at 65 dB SPL for a SP group (76.1 ± 16.1%) compared to LP coupling (66.2 ± 23.5%) in a total of 42 patients, this could not be confirmed in two retrospective studies ( Schraven et al. (2018 Edlinger et al. 2021 ; ). In our analysis, the EG tends to be slightly higher in the LP group than in the SP group in the entire frequency range. However, this must be attributed mainly to the diversity of both groups. While the LP group exclusively included patients with sensorineural hearing loss, SP coupling was mainly performed in children with congenital ear canal stenosis or atresia (7/9 patients) and conductive hearing loss. In this patient group, due to the normal BC hearing level, the primary goal was to bridge the air bone gap with the AMEI. An additional amplification was initially not required, resulting in a lower EG in our SP group. Rahne et al. 2021
In general, the limitations of the determination of the EG as an outcome parameter should be pointed out. Besides intrinsic noise of the microphone, FF thresholds of AMEI are influenced by compression and feedback algorithms and the individual subjective acceptance of the patients, which may lead to a distortion of the EG. Since the calculation of the EG is based on the BC, the heterogeneous distribution of cochlear reserve in all subgroups must also be considered as a strong confounding factor. The dependence of the EG and the cochlear hearing loss for all coupling groups is visualized in Supplemental Digital Content 4,
In our clinical assessment, there were no meaningful differences between the SP and LP groups for coupling efficiency in the frequency range 0.5 to 2 kHz. In the frequency range >2 kHz; however, a tendency toward better coupling efficiency was found for the LP coupling. The tendentially somewhat less favorable coupling efficiency in the high-frequency range for incus coupling (LP and SP coupling) is consistent with the results reported by
) in a small case series. Mlynski et al. (2015a ), comparing the coupling coefficients of SP and LP coupling, showed a higher heterogeneity of results for the SP group, although no frequency-specific analysis was performed. When evaluating our clinical data, a limiting factor to consider is the small number of patients in both groups and the high variance of the results. Rahne et al. (2021
In contrast to the experimental data, the clinical examination showed a slightly worse coupling efficiency in the high-frequency range (>2 kHz) for the SP group. We attribute this to the fact that in the experimental setting the best possible coupling could be determined and analyzed with the help of the laser Doppler vibrometer measurement. However, wide 95% CI indicates heterogeneity in the experimental data. The clinical results therefore emphasize that an intraoperative real-time monitoring should be aimed for, especially in SP coupling. However, current methods using auditory brainstem responses are still in the study phase (
). Generally, the SP coupler leads to a shortened time of surgery and to reduced risks for facial nerve and chorda tympani ( Fröhlich et al. 2021 ). Schraven et al. 2018 Clip Coupler Versus RW Soft Coupler
In the case of chain defects, the only options for FMT placement are stapes (head or footplate) or RW coupling. The experimental data supports an advantage of coupling on the stapes head and footplate overall other coupling (
Fig. 4), with an average frequency response up to 10 dB larger than the average of all couplers. This can be attributed to a more favorable mechanical coupling both of the FMT to the bone (as compared to the RW) and to the inner ear (as compared to coupling further away along the OC). Compared to RW coupling, our data shows a significant superiority of the coupling on the stapes head from the resonance frequency. It must be noted that footplate velocity is not a correct measure for induced inner ear pressure gradient in the case of reverse excitation (via RW). This is mainly because the reverse impedance of the ME is higher than the inward impedance of the cochlea, as described by ). They also assume leakage through fissures in the scala vestibuli that they collectively call a “third window.” Both of these effects mean that stapes footplate velocity underestimates the actual differential pressure in the cochlea. Therefore our data have been corrected according to the investigations of Stieger et al. However, achieving effective coupling to the RW is much more challenging than stapes coupling. This is due to the small diameter of the RW and the RW niche, which even using an FMT soft coupler (as in our experiments) often causes a less than optimal coupling. This can also be seen in the different SD of the measurement data of Clip and RW coupling ( Stieger et al. (2013 Fig. 2).
The coupling should be performed parallel to the longitudinal stapes axis. For this purpose, either a parallel arrangement using a clip prosthesis (our data) or a serial arrangement using a Bell prosthesis (
) is possible. Deviations of 20° to 60° reduce the transfer function about 3 to 10 dB ( Huber et al. 2006 ). When coupling to the stapes head as well as to the footplate, the actuator should be covered with cartilage and stabilized to prevent loose coupling, to intensify the piston-like movements along the Bornitz et al. 2010 z axis and to suppress movements in other degrees of freedom. In contrast, excessive pressure on the FMT by the cartilage shield—in the case of insufficient ME ventilation—should be avoided as it leads to a stiffening of the annular ligament and to low-frequency transmission loss ( ). Müller et al. 2019
In addition to the below-average transmission of the RW soft coupler over the entire frequency range, our data also show a very wide 95% CI and thus a wide spread of the individual values in relation to the mean value (
Our clinical results showed no significant difference between forward (Clip coupling) and reverse stimulation (RW coupling) in speech comprehension. These results are consistent with previous clinical studies (
Zahnert et al. 2016 , 2018 ; ). However, the RW group showed a coupling deficit in the low-frequency range (0.5–1 kHz). Similar results were found in previous studies ( Spiegel et al. 2020 Busch et al. 2016 ; ). Comparing the coupling modalities in patients with a defect OC (couplings to the stapes head or RW), the best amplification (most pronounced at 1.5–3 kHz) was provided by the Clip vibroplasty. Similar results were reported by Schraven et al. 2016b ). The clinical and experimental results of RW vibroplasty showed a wide variance, since a large number of influencing factors must be considered. Various works have experimentally investigated the multiple influencing factors when coupling to the RW and have provided suggestions for improved and more reproducible transmission. Schraven et al. (2016a ) propagated an orthogonal positioning of the FMT vibration axis to the RW membrane (RWM) without contact of the FMT to bony structures. This may require drilling to widen the RW niche, as the mean width of the RW niche is 1.66 mm and the mean depth 1.34 mm ( Arnold et al. (2010 ), whereas the diameter of the FMT measures 1.8 mm and its length 2.3 mm. Coupling can also be improved by interposing connective tissue between the FMT and RWM ( Su et al. 1982 ) or by pressing the FMT against the hypotympanic wall ( Arnold et al. 2010 ). In our experiments, the RW soft coupler was used. Although it is adapted to the geometry of the RW, the coupling failed reproducibility as well. Nakajima et al. 2010 ), Maier et al. (2013 ), and Salcher et al. (2014 ) have clearly shown the dependency of the transmission on the preload of the RWM. However, it is difficult to assess the preload in the surgical site, since the RWM is difficult to see, especially in malformations, and free-drilling of the RWM can lead to significant noise trauma. The Hannover coupler ( Gostian et al. (2016 ), which is still in the development phase, is a promising tool to achieve an optimal transmission from a frequency range of 1 kHz with a defined preload of <10 mN on the RWM. Müller et al., 2017
Based on the clinical and experimental data, the coupling at the stapes head should be preferred whenever possible. The experimental as well as the clinical data show the best possible transmission of the FMT in the speech range above 1.5 kHz. However, it must be noted that the condition of the ME is often unknown preoperatively and the appropriate coupling site is only revealed intraoperatively. If the RWM remains the sole possible actuator target structure, the coupling deficit of the RW soft coupler in the low-frequency range should be taken into account in patient counseling, especially when sufficient information on ME status is available from previous surgeries. Particularly in patients at the lower indication limit, the low dynamic range may ultimately result in unsatisfactory speech comprehension. Especially for this target structure, the development of a real-time monitoring system would be preferable.
Factors Effecting Vibrogram–BC and EG
In general, the clinical data show a pronounced heterogeneity. The vibrogram normative data originate from TB experiments with the FMT (LP coupling) and they are assumed to have a scatter range of ±10 dB (estimated from other measurements; the reference data for the vibrogram are not disclosed). As a result, the coupling efficiency for the LP coupler should ideally be zero over the entire frequency range. For the other coupling options (SP, RW, and Clip), the efficiency indicates that those couplings are more or less efficient than the LP coupling. However, the vibrogram is a proprietary development of MED-EL. No data on calibration is available either by the company or in the literature. For these reasons, individual calibration to different coupler types is also currently not available.
Due to the method, the vibrogram and BC thresholds can at best be determined to an accuracy of 5 dB, so that a range in the difference of the vibrogram–BC of 20 dB is to be expected.
Individual outliers at only one frequency may indicate resonance-antiresonance problems. With clip coupling, large deviations in the vibrogram are possible due to differences in TM reconstruction and individual annular ligament prestress. The tympanic coverage, and especially the prestress, can produce low-frequency losses of 10 dB or more (
). Furthermore, differences in vibrogram thresholds can be expected depending on ME pathology. For example, OC malformation is more noticeable in patients with SP coupling compared to LP, clip, or RW coupling. The auditory canal volume itself can also lead to an influence of the vibrogram thresholds. Depending on the residual volume in front of the TM, an increase in the stiffness of the chain is to be expected, which can cause deviations of up to 10 dB in the vibrogram thresholds in frequencies up to about 2 kHz. In patients with small and stiffened TM, even losses of up to 20 to 30 dB are possible ( Bornitz et al. 2021 ). A reduction of the vibrogram–BC difference is only to be expected in the case of chain changes that also affect BC, for example, the BC notch at 2 kHz in otosclerosis patients. Beleites et al. 2014
When evaluating the EG, the individually tolerated amplification must also be taken into account. Patients with a longstanding, poorly compensated hearing loss or with a higher degree of hearing loss are very sensitive to loudness, so that there is only a small dynamic range. In this patient group, it is often necessary to reduce the amplification power (especially in the high-frequency range) for everyday listening situations. Therefore, the EG does not reflect the true performance of the AMEI in all patients. At this point, the maximum stable gain would be a parameter to be analyzed in future clinical studies depending on the coupling modality.
Summary of the Limitations
The influencing factors on the transmission known from the experimental setting (coupling position, coupling state, and alignment) are extended in the clinicial use by multiple individual biasing factors.
Patient-specific factors such as heterogeneous pathologies (sensorineural hearing loss, conductive hearing loss, mixed hearing loss, and atresia) with heterogeneous cochlear reserve or heterogeneous sensitivity to loudness influence the measurements of the EG, as do device-specific factors such as compression and feedback algorithms, and intrinsic noise of the microphone.
Patient-specific factors such as the morphology and condition of the ME as well as the outer ear (especially the external auditory canal), TM and OC reconstruction, and prestress of the annular ligament influence the vibrogram. Also, device-specific factors such as resonance-antiresonance phenomena affect the measurements.
Clinically, no significant differences for postoperative speech comprehension could be determined between the different coupling sites. The comprehensive look at both the clinical and experimental data permits the conclusion that the influence of coexisting influencing factors in the individual patient are crucial for the postoperative speech comprehension. Even superior coupling-related transmission properties in vitro can be repressed by numerous biasing factors in vivo.
The authors would like to thank MED-EL (MED-EL, Innsbruck, Austria) for providing the couplers and FMTs.
Arnold A., Kompis M., Candreia C., Pfiffner F., Häusler R., Stieger C. (2010). The floating mass transducer at the round window: Direct transmission or bone conduction? Hear Res, 263, 120–127.
ASTM. (2014). Standard practice for describing system output of implantable middle ear hearing devices. ASTM Volume 13.02 Medical and Surgical Materials and Devices (II), F2504 05.
Ball G. R., Huber A., Goode R. L. (1997). Scanning laser Doppler vibrometry of the middle ear ossicles. Ear Nose Throat J, 76, 213–222.
Baumgartner W. D., Böheim K., Hagen R., Müller J., Lenarz T., Reiss S., Schlögel M., Mlynski R., Mojallal H., Colletti V., Opie J. (2010). The
for conductive and mixed hearing losses: European multicenter study results. Adv Otorhinolaryngol, 69, 38–50.
Beleites T., Neudert M., Beutner D., Hüttenbrink K. B., Zahnert T. (2011). Experience with vibroplasty couplers at the stapes head and footplate. Otol Neurotol, 32, 1468–1472.
Beleites T., Neudert M., Bornitz M., Zahnert T. (2014). Sound transfer of active middle ear implants. Otolaryngol Clin North Am, 47, 859–891.
Bornitz M. (2011).
implantable actuators by means of a middle ear simulation model [Dissertation]. TU Dresden.
Bornitz M., Hardtke H. J., Zahnert T. (2010). Evaluation of implantable actuators by means of a middle ear simulation model. Hear Res, 263, 145–151.
Bornitz M., Lasurashvili N., Neudert M., Beleites T., Zahnert T. (2021). Ankopplung aktiver Mittelohrimplantate – biomechanische Aspekte. HNO, 69, 464–474
Bornitz M., Zahnert T. (2013). Aspekte zur Ankopplung implantierbarer Aktoren. Proceedings der 16. Jahrestagung der Deutschen Gesellschaft für Audiologie, 1–6.
Busch S., Lenarz T., Maier H. (2016). Comparison of alternative coupling methods of the
floating mass transducer. Audiol Neurootol, 21, 347–355.
Colletti V., Carner M., Colletti L. (2005). Round window stimulation with the floating mass transducer: A new approach for surgical failures of mixed hearing losses. Proceedings of the XVIII IFOS World Congress; 25–30th June, 2005; Rome.
Cumming G., Finch S. (2005). Inference by eye: Confidence intervals and how to read pictures of data. Am Psychol, 60, 170.
Edlinger S. H., Hasenzagl M., Schoerg P., Muck S., Magele A., Sprinzl G. M. (2021). Long-term safety and quality of life after vibroplasty in sensorineural hearing loss: Short/long incus process coupler. Audiol Neurootol, 27, 62–70.
Eiber A., Huber A. M., Lauxmann M., Chatzimichalis M., Sequeira D., Sim J. H. (2012). Contribution of complex stapes motion to cochlea activation. Hear Res, 284, 82–92.
Fröhlich L., Rahne T., Plontke S. K., Oberhoffner T., Dahl R., Mlynski R., Dziemba O., Aristeidou A., Gadyuchko M., Koscielny S., Hoth S., Kropp M. H., Mir-Salim P., Müller A. (2021). Intraoperative quantification of floating mass transducer coupling quality in active middle ear implants: A multicenter study. Eur Arch Otorhinolaryngol, 278, 2277–2288.
Gostian A. O., Pazen D., Ortmann M., Anagiotos A., Schwarz D., Hüttenbrink K. B., Beutner D. (2016). Loads and coupling modalities influence the performance of the floating mass transducer as a round window driver. Otol Neurotol, 37, 524–532.
Hinkle D. E., Wiersma W., Jurs S. G. (2003). Applied Statistics for the Behavioral Sciences (Bd. 663). Houghton Mifflin College Division.
Huber A. M., Ball G. R., Veraguth D., Dillier N., Bodmer D., Sequeira D. (2006). A new implantable middle ear hearing device for mixed hearing loss: A feasibility study in human temporal bones. Otol Neurotol, 27, 1104–1109.
Huber A. M., Sequeira D., Breuninger C., Eiber A. (2008). The effects of complex stapes motion on the response of the cochlea. Otol Neurotol, 29, 1187–1192.
Hüttenbrink K. B., Beutner D., Bornitz M., Luers J. C., Zahnert T. (2011). Clip vibroplasty: Experimental evaluation and first clinical results. Otol Neurotol, 32, 650–653.
Lailach S., Müller C., Lasurashvili N., Seidler H., Zahnert T. (2021). Active hearing implants in chronic otitis media. HNO, 69, 447–463.
Maier H., Salcher R., Schwab B., Lenarz T. (2013). The effect of static force on round window stimulation with the direct acoustic cochlea stimulator. Hear Res, 301, 115–124.
Mlynski R., Dalhoff E., Heyd A., Wildenstein D., Hagen R., Gummer A. W., Schraven S. P. (2015a). Reinforced active middle ear implant fixation in incus vibroplasty. Ear Hear, 36, 72–81.
Mlynski R., Dalhoff E., Heyd A., Wildenstein D., Rak K., Radeloff A., Hagen R., Gummer A. W., Schraven S. P. (2015b). Standardized active middle-ear implant coupling to the short incus process. Otol Neurotol, 36, 1390–1398.
Müller C., Zahnert T., Ossmann S., Neudert M., Bornitz M. (2019). Vibroplasty combined with tympanic membrane reconstruction in middle ear ventilation disorders. Hear Res, 378, 166–175.
Müller M., Salcher R., Lenarz T., Maier H. (2017). The hannover coupler: Controlled static prestress in round window stimulation with the floating mass transducer. Otol neurotol, 38, 1186–1192.
Nakajima H. H., Dong W., Olson E. S., Rosowski J. J., Ravicz M. E., Merchant S. N. (2010). Evaluation of round window stimulation using the floating mass transducer by intracochlear sound pressure measurements in human temporal bones. Otol Neurotol, 31, 506–511.
Rahne T., Skarzynski P. H., Hagen R., Radeloff A., Lassaletta L., Barbara M., Plontke S. K., Mlynski R. (2021). A retrospective European multicenter analysis of the functional outcomes after active middle ear implant surgery using the third generation vibroplasty couplers. Eur Arch Otorhinolaryngol, 278, 67–75.
Raphael F. (2021). Übertragungsverhalten elektromechanischer Aktoren bei Anregung unterschiedlicher Mittelohrstrukturen im humanen Felsenbeinmodell [Dissertation]. TU Dresden.
Rosowski J. J., Chien W., Ravicz M. E., Merchant S. N. (2007). Testing a method for quantifying the output of implantable middle ear hearing devices. Audiol Neurootol, 12, 265–276.
Salcher R., Schwab B., Lenarz T., Maier H. (2014). Round window stimulation with the floating mass transducer at constant pretension. Hear Res, 314, 1–9.
Schmuziger N., Schimmann F., àWengen D., Patscheke J., Probst R. (2006). Long-term assessment after implantation of the
device. Otol Neurotol, 27, 183–188.
Schraven S. P., Gromann W., Rak K., Shehata-Dieler W., Hagen R., Mlynski R. (2016a). Long-term stability of the active middle-ear implant with floating-mass transducer technology: A single-center study. Otol Neurotol, 37, 252–266.
Schraven S. P., Mlynski R., Dalhoff E., Heyd A., Wildenstein D., Rak K., Radeloff A., Hagen R., Gummer A. W. (2016b). Coupling of an active middle-ear implant to the long process of the incus using an elastic clip attachment. Hear Res, 340, 179–184.
Schraven S. P., Rak K., Cebulla M., Radeloff A., Grossmann W., Hagen R., Mlynski R. (2018). Surgical impact of coupling an active middle ear implant to short incus process. Otol Neurotol, 39, 688–692.
Snik A., Cremers C. (2004). Audiometric evaluation of an attempt to optimize the fixation of the transducer of a middle-ear implant to the ossicular chain with bone cement. Clin Otolaryngol Allied Sci, 29, 5–9.
Spiegel J. L., Kutsch L., Jakob M., Weiss B. G., Canis M., Ihler F. (2020). Long-term stability and functional outcome of an active middle ear implant regarding different coupling sites. Otol Neurotol, 41, 60–67.
Stieger C., Rosowski J. J., Nakajima H. H. (2013). Comparison of forward (ear-canal) and reverse (round-window) sound stimulation of the cochlea. Hear Res, 301, 105–114.
Su W. Y., Marion M. S., Hinojosa R., Matz G. J. (1982). Anatomical measurements of the cochlear aqueduct, round window membrane, round window niche, and facial recess. Laryngoscope, 92, 483–486.
Zahnert T., Löwenheim H., Beutner D., Hagen R., Ernst A., Pau H. W., Zehlicke T., Kühne H., Friese N., Tropitzsch A., Lüers J. C., Mlynski R., Todt I., Hüttenbrink K. B. (2016). Multicenter clinical trial of vibroplasty couplers to treat mixed/conductive hearing loss: First results. Audiol Neurootol, 21, 212–222.
Zahnert T., Mlynski R., Löwenheim H., Beutner D., Hagen R., Ernst A., Zehlicke T., Kühne H., Friese N., Tropitzsch A., Luers J. C., Todt I., Hüttenbrink K. B. (2018). Long-term outcomes of vibroplasty coupler implantations to treat mixed/conductive hearing loss. Audiol Neurootol, 23, 316–325.