Vibrant Soundbridge Round Window Vibroplasty: Safety, Coupling Efficiency, and Speech Outcome of the Most Common Coupling Modalities

The Vibrant Soundbridge [1] (VSB; MED-EL GmbH, Austria) is an active middle ear implant which provides mechanical stimulation to the auditory ossicles or directly to the cochlear windows by itsfloating mass transducer (FMT). The VSB was originally developed to treat patients with sensorineural hearing loss (SNHL) by coupling the FMT to the incus [1‒3], and the procedure was termed incus vibroplasty. In order to extend the application of the VSB to conductive hearing loss (CHL) and mixed hearing loss (MHL), other coupling sites for the FMT such as the oval window [4, 5], stapes head [6‒8], or round window (RW) [9] were investigated and successfully implanted, allowing for the rehabilitation of patients with various middle ear or ear canal disorders.

The feasibility of the RW vibroplasty with direct placement of the FMT in front of the RW membrane (RWM) was demonstrated in 2005 by Colletti et al. [9], but patient outcomes showed a high variability [10‒13]. Main reasons were considered to be the mismatch in diameter between the FMT and the RWM [14] and the surgically challenging placement of the FMT at the RW, which may prevent sufficient signal transmission between the FMT and RWM. For this reason, the RW niche must be enlarged to fit the FMT, and the bone surrounding the RWM must be reduced to enable the contact between RWM and FMT, while simultaneously avoiding contact between FMT and the bony structures. To overcome these challenges, indirect coupling was introduced in 2006, using interposed materials between the RW membrane and the FMT such as cartilage, fascia [10, 15, 16], or a collagen membrane made from bovine pericardium (Tutopatch; Tutogen Medical GmbH, Germany) [17]. The thickness and shape of the interposed fascia depend solely on the surgeon and bear the risk of atrophy, while Tutopatch has a defined thickness and does not alter. To enable a more stable and standardized coupling with reproducible results, couplers such as the RW coupler (RWC) with a spherical titanium tip [14, 18‒20] and the RW soft (RWS) coupler with a conical silicone tip [21, 22] were developed in 2010 and 2014, respectively. Experiments in temporal bones demonstrated that the coupling efficiency (C_eff) depends on the static preload [17, 23] with an optimal coupling force in the range of 5–20 mN [19, 21]. Based on these findings, the custom-made devices Hannover coupler version 1 (HCV1) [24, 25] and version 2 [26, 27] (HCV2) were developed. The Hannover coupler is a mechanical cage for the FMT, with a small ball tip in the front. The Hannover coupler version 1, only applied in 2 cases, featured a simple spring at the rear end to provide a static force on the RWM, and the second version utilizes a more sophisticated s-shaped spring, which includes a visual preload indicator of the force applied (5 mN–20 mN), as well as an anchor to securely support the spring. The aim of the present study was to retrospectively evaluate the long-term outcome of the previously mentioned six most common VSB RW coupling modalities with regard to (1) safety and preservation of residual hearing, (2) C_eff, and (3) speech perception.

The long-safety and audiological outcomes of the VSB RW coupling were analyzed retrospectively in all patients implanted at the Hannover Medical School (MHH) with a VSB FMT placed at the RW between August 2006 and September 2021. All patient data were collected during routine visits.

Patients with a VSB implanted in at least one ear with the following most common coupling modalities were included in the analysis: the FMT was positioned in the RW niche (1) without material between FMT and round window membrane (RWM) (group Direct), (2) with interposed fascia (group Fascia), (3) processed collagen membrane made from bovine pericardium (Tutopatch^®, group Tutopatch) between the RWM and the FMT, or by using (4) the titanium RWC (group RWC), (5) the RWS coupler, or (6) the custom-made device HCV2. Patients were excluded from the analyses if the preoperative bone conduction threshold (BC threshold) (up to a maximum of 3 months before surgery) was not available or not valid, because the BC threshold was either (1) outside the measurement limits of the audiometer, (2) masking of the opposite ear was necessary but not possible, or (3) the patient was physically and mentally unable to perform the measurement. For the audiological evaluation, patients were also excluded if a complete postoperative data set consisting of the BC threshold, in situ threshold (vibrogram threshold [V_thr]), aided word recognition score (WRS) in quiet, and aided sound field (SF) threshold was not available at all selected time points (see audiological analysis).

The safety of implantation was evaluated based on the number of adverse events (AEs) such as complications, revisions, and explantations. Available postoperative AEs up to the end of the data collection in August 2022 were extracted from medical files of each patient. AEs included complications, revisions, total revisions, and explantations. Events were classified as complications if they did not require revision or explantation, such as HL, pain, vertigo, or wound infections. Complications caused by the magnet strength of the audio processor that had been resolved by reducing the magnet strength were not considered. Revisions included all surgeries with manipulation (relocation) of the implant, the cable, or the FMT. Total revisions were defined as an explantation in combination with the reimplantation of a new VSB in the same surgery, independent of the new coupling modality or site. Total revisions with reimplantation of an implant other than a VSB were counted as an explantation. The AEs were then divided into four categories according to the cause of their occurrence: (1) implant related (implant or cable defects), (2) coupling related (e.g., FMT migration), (3) surgery related (wound healing disorder, infection, HL, or facial palsy), and (4) medical issue unrelated to implantation (cholesteatoma and other ear surgeries, which required revision or explantation).

A total of 111 RW vibroplasties were available for analysis, including seven ears with a total revision and change of coupling modality (see the flowchart in Fig. 1). Overall, 32 RW vibroplasties were excluded from the safety analysis. Twenty-five patients did not have a reliable preoperative BC threshold, 4 patients were implanted with a different type of FMT coupling in the RW niche than investigated here, and 3 patients were not able to perform the tests. Thus, 79 RW vibroplasties, including 4 bilateral implanted patients and 4 patients with two subsequent RW vibroplasties in one ear, which contributed both to the analysis of the respective coupling modality (see Table 2), could be included in the safety analysis.

Demographics, etiology, and previous middle ear surgeries for each group can be found in Table 1. The mean age at implantation was 55.0 ± 15.6 years (mean value ± standard deviation [MV ± SD]) in the range of 11–75 years. The mean preoperative BC threshold was 34.7 ± 12.3 dB HL, and the AC threshold 82.1 ± 15.5 dB HL. Reasons for treatment were chronic otitis media (n = 35 ears), cholesteatoma (n = 21 ears), otosclerosis (n = 9 ears), or malformation (n = 7 ears). In 3 patients (n = 3 ears), the cause was unknown. Due to these preexisting conditions, an average of 2–3 ear surgeries (including revisions) were performed prior to RW vibroplasty. The surgeries included tympanoplasties (n = 71 surgeries in 33 ears), subtotal petrosectomies (n = 65, 45 ears), radical cavities (canal wall down mastoidectomies, n = 22, 17 ears), VSB reimplantations with change of coupling modality (n = 8, 5 ears) (see Table 2), and other ear surgeries (n = 9, 2 ears).

The audiological performance was evaluated based on the comparison of average BC threshold, C_eff, and effective gain (EG) (all across 0.5, 1.0, 2.0, and 4.0 kHz) as well as aided WRS over time separately for the different coupling modalities. Hearing thresholds were determined using pure-tone measurements (AC, BC) or warble tones (aided SF threshold). The WRS in quiet was measured using the Freiburg monosyllable test at 65 dB SPL. All measurements were performed in a soundproof chamber, with the opposite ear plugged and muffled or masked if needed.

The in situ thresholds were determined directly via the processor during each fitting procedure using the Symfit^® fitting software. With the in situ threshold and the BC threshold, the C_eff was calculated for each patient (C_eff = V_thr – BC). Further, the EG was determined (EG = BC – SF).

For the audiological analysis, three postoperative time periods were defined. Short-term data were collected at the time of initial activation (IA) approximately 6 to 8 weeks after implantation. The mid-term data (2 years) were collected 1–3 years after implantation, and the long-term data (5 years) were collected 4–7 years after implantation. If more than one complete data set was available in the defined time span, the data set closest to the 2-year or 5-year appointment was chosen.

As the majority of patients from groups Direct, Fascia, and Tutopatch were implanted before the implementation of the in situ function in the fitting software, the in situ threshold at IA was not available and therefore not considered for these groups. For the groups RWS and HCV2, only a small number (n = 3 each) of long-term data sets (5 years) were available at the end of data collection in August 2022, and no long-term analysis was performed.

Only ears already included in the safety analysis were further considered for the audiological analysis. Due to missing data caused by loss to follow-up (n = 15), explantations or total revisions (n = 4), or incomplete measurements (n = 4), only 56 ears (52 patients) were included in the audiological analysis (Fig. 1). The mid-term data set (2 years) was collected after an average of 2.7 years, and the long-term data set (5 years) was collected after an average of 5.4 years after implantation.

The statistical analysis was performed using SigmaPlot 14 (Systat Software Inc.). Data sets were tested for normal distribution using the Shapiro-Wilk test. For comparison of data sets within groups, the paired t test was used in case of normal distribution, or else the Wilcoxon signed rank. When comparing data sets between two groups, the Welch’s t test was used in cases of homogeneity of variances; otherwise, the Mann-Whitney rank-sum test was applied. A significance level of p < 0.05 was used for all tests. Frequency-specific results were calculated, but reported as the MV across 0.5, 1, 2, and 4 kHz (PTA₄). MVs are given with the SD throughout the text.

Progressing HL after implantation can eventually result in a threshold exceeding the audiometer limit. To enable further analysis, a best-case estimate was defined as the audiometer limit plus 5 dB. This estimate was applied to six postoperative BC thresholds (n = 3, 500 Hz; n = 3, 4 kHz) and to three postoperative in situ thresholds (n = 3 patients; 4 kHz).

The safety analysis focused on the investigations of AEs in n = 79 RW vibroplasties up to 16 years after implantation (see Table 2). Overall, 23 AEs (29.1%) with 7 complications, 7 revisions, 5 total revisions, and 4 explantations were observed. The most AEs were found in group Tutopatch (n = 8, 36.4%) and the lowest rates in groups Direct and HCV2 (n = 1, 20%; n = 3, 20%, respectively). All complications were surgery related and occurred during or shortly after the surgery. Complications included HL, vertigo, wound infection, facial palsy, and pain at the implant area (coil). Revisions and total revisions occurred on average after 2 years and were mostly coupling related (n = 5) or surgery related (n = 4). Coupling-related AEs were caused by FMT migration (n = 3) or tissue growth (n = 2) which reduced the vibratory movement of the FMT. Explantations were medical related (n = 2, cholesteatoma) or device related (n = 2, implant or cable failure) and occurred after an average of 5.8 years. Four patients were reimplanted with another type of acoustic implant (n = 3, BC hearing aid; n = 1, Codacs™ [Cochlear Ltd., Sydney, Australia]). No statistically significant differences were observed between the coupling groups concerning the type, cause, or time of occurrence of AEs (Welch’s t test, p > 0.05).

The preoperative BC thresholds (baseline measurement) differed between coupling groups (Fig. 2a; Table 3). Statistically significant differences were found between the group with the largest HL (group Fascia, mean bone conduction pure tone average [BC PTA₄] 48.0 ± 11.3 dB HL) and the two groups with the smallest degrees of HL (group RWS, mean BC PTA₄ 32.2 ± 9.3 dB HL; group HCV2, BC PTA₄ 28.0 ± 9.5 dB HL; Welch’s t test, p = 0.019 and p = 0.049, respectively).

The effect of surgery onto residual hearing preservation was tested by comparing the BC PTA₄ before (preop) and after the surgery (IA, 2 years and 5 years). The comparison of BC PTA₄ before the surgery and at the time of first activation (IA) showed a mean HL of more than 5 dB in group RWS and group HCV2, which increased over time and became significant after 2 years (RWS 7.6 dB, HCV2 8.8 dB; paired t test, p = 0.022 and p ≤ 0.001, respectively). The HL of 8.1 dB after 2 years in group Direct was not significant (paired t test, p = 0.587). The other coupling groups showed no significant change in residual hearing at any time point after the surgery (paired t test, p > 0.05).

The results of the C_eff analysis are displayed in Figure 2b and Table 3. The C_eff was determined for the groups Direct, Fascia, and Tutopatch after 2 years and 5 years and for the groups RWS and HCV2 at IA and after 2 years. A mean C_eff of approximately 25 dB was achieved by most groups at the investigated time points. An exception was found in the RWC group at IA (26.7 ± 8.1 dB) and the Tutopatch group at the 2 years date (26.6 ± 15.9 dB). The C_eff improved significantly in the group RWC by 6.1 dB from IA to 2 years (paired t test, p = 0.035) and by 3.8 dB from 2 years to 5 years (paired t test, p = 0.015). The C_eff of the other groups improved by <3.4 dB from IA to 2 years or 2 years to 5 years. In the Fascia group, however, a slight but nonsignificant deterioration of C_eff by 3.5 dB from 2 years to 5 years was determined. No significant differences were found between coupling groups at any time (Welch’s t test, p > 0.05).

The EG was determined as the SF threshold deducted from the BC threshold (see Fig. 2c; Table 3). A positive EG, indicating a closure of the air-bone gap, was determined in the groups Direct, Fascia, and HCV2 at IA and additionally in the groups RWS and RWC at the 2 years date. No significant differences in EG were found between the coupling groups at any time (Welch’s t test, p > 0.05). The EG improved in all groups from IA to the 2 years date, but the results were only statistically significant in group RWC and group HCV2 (paired t test, p = 0.01; p = 0.002, respectively). The EG of the Fascia group declined nonsignificantly by 3.2 dB from 2 years to 5 years.

The hearing benefit was evaluated by comparing the aided WRS over time and between coupling groups as shown in Figure 2d and Table 3. At the IA, the mean scores ranged from 50.0% (Fascia), 67.5% (HCV2), 71.9% (Tutopatch), 72.9% (RWC), and 73.1% (RWS) up to 80.0% (Direct). After 2 years, the improvement of WRS compared to the IA ranged from 2.5 percentage points (pp) (RWS) to 23.3 pp (HCV2), with the latter being statistically significant (HCV2, 67.5% vs. 90.8%; paired t test, p ≤ 0.001). The scores remained constant up to 5 years in the investigated groups (Table 3) with no significant differences between coupling groups (Welch’s t test, p > 0.05).

The current manuscript presents, to the best of our knowledge, the first comprehensive study that compares six different RW vibroplasty coupling modalities regarding safety, speech outcome, and coupling quality (efficiency and stability) for mid-term (2 years) and long-term (5 years) data.

For the safety analysis, we examined all AEs after implantation over an average follow-up period of 6.6 ± 4.3 years for 79 patients. We found no statistically significant difference between different coupling modality groups with respect to the quantity or type of AEs.

The overall AE rate was 29.1%, which is in the lower range of vibroplasty AEs reported by other studies. Vickers et al. [28] analyzed a cohort of 163 vibroplasty patients, pooling different coupling sites and couplers and reported an AE rate of 24.5% over 6 years, whereas Zahnert et al. [20] showed 33 AEs in a group of 24 vibroplasties with RWC, oval window coupling, CliP and Bell coupler over 3 years.

A subgroup of the AEs are coupling-related revisions. Revisions were assumed to be coupling related if the WRS dropped suddenly or over time despite stable BC thresholds and changes in V_thr were noticed at the same time. The coupling deficit was confirmed by the surgeon during the revision surgery. Our study included five events (6.3%) of coupling-related revisions. In 2 patients (Tutopatch, HCV2), the FMT was immobilized due to tissue growth, and in 3 patients (Fascia, RWS, HCV2), a loss of contact between the FMT and RWM was observed.

The number of coupling-related revisions found in other studies varies widely. While Skarzynski et al. [12] showed a rate of 9.5% over 3 years in a group of patients with direct coupling (n = 21), Schraven et al. [29] even found a revision rate of 71%, with regard to their subgroup of RW vibroplasty without a coupler (Direct, Fascia, and Tutopatch), while in their group RWC, as in this study, no revision was necessary. The revision rates found in our study and in the literature suggest that the use of the RWC reduces the number of coupling-related revisions.

Another issue is AEs related to recurrent cholesteatoma in patients without subtotal petrosectomy, as shown by Sprinzl et al. [30] with a survival rate of 59.1% within a 6-year observation time. In this study, 20 out of 24 patients with cholesteatoma underwent subtotal petrosectomy at least 6 months prior to VSB implantation, which is a standard procedure in our clinic [31]. Over an average of 6.6 years, the survival rate was 75% in patients without subtotal petrosectomy and 89.5% in patients with subtotal petrosectomy. The two-step procedure presumably reduces the risk of a recurring cholesteatoma.

In addition to the previously reported AEs, we observed three device failures (3.8%) due to a mechanical impact on the implant site (1 patient) and unknown reasons (2 patients), resulting in one total revision and two explantations. Similar rates of device failures of 1.2% [28], 4% [32], and 7% [33] were reported in studies including large cohorts of 92 to 163 patients with multiple types of vibroplasties. Overall, the risk of device failure is low, and therefore, no correlation could be found associated with a specific coupling modality.

In total, out of the 79 RW vibroplasties analyzed, four explantations were performed after a mean time of use of 5.8 ± 1.5 years. These patients were not reimplanted with a VSB for various unrelated reasons, including progressive HL, medical, and personal reasons.

In addition, general risks of ear surgeries were found in single cases (each rate <5%) in this analysis with a combined AE rate of 13.9%. The risks include vertigo, facial nerve palsy, infections, and HL, which have been reported repeatedly in previous studies [16, 20, 30, 31].

Hearing preservation is an equally important aspect in evaluating the safety of ear surgeries. In our study, groups Fascia, Tutopatch, and RWC showed a stable BC threshold even after 5 years. However, groups Direct, RWS, and HCV2 showed a mean HL between 7.6 dB and 8.8 dB at 2 years. Although all frequencies were affected, the HL was more pronounced at 4 kHz and 6 kHz. The decline in BC PTA₄ in group RWS and HCV2 was already noticeable at IA (6 dB to 7.2 dB), unlike in group Direct, where the decline appeared between IA and 2 years (5.6 dB). Previous studies analyzing groups Fascia [16] and RWC [20] report a stable BC threshold up to 3.3 years after implantation. In studies including the RWS, the mean pre- or postoperative BC PTA₄ was either reported across a mixed patient group including different types of vibroplasties [19, 22], or the preoperative BC threshold [34] or any HL <10 dB was not reported [19].

To identify possible causes of HL, an extended screening of all patients with a HL between 5 dB and 10 dB within the first 2 years was carried out. However, no relationships/correlations were found between the HL of the patients and their demography, etiology, pathology, or number of prior surgeries. Merely, a HL shortly after implantation may suggest that it is surgery related or possibly coupler related.

The RWS and HCV2 have been designed with a small tip diameter to allow deeper placement into the RW niche than previous coupling modalities. The HCV2, in particular, was developed to enable application and monitoring of the applied preload. The pressure in the cochlear caused by the preload could also have a potential effect on the BC threshold. Moreover, the HCV2 is relatively long (6.2 mm) compared to the FMT with RWC (3.2 mm) and RWS (3.3 mm). This is due to the extended tip and spring, which requires an even larger RW niche than the other RW coupling types. As Pau et al. [35] stated, noise trauma might be caused by bone removal close to the RWM. There might be an additional risk of noise trauma when enlarging the RW niche on the opposite side of the RWM. Further research is required to evaluate these hypotheses.

The focus of every treatment with a hearing aid is the rehabilitation of speech understanding, which is achieved with a WRS ≥75% following Mueller and Killion [36]. Except for group Direct, the score at IA remained below 75% in all groups and improved over time, resulting in a WRS ≥75% in all groups except group Fascia. The improvement over time can be explained by optimized fitting performed during clinical visits, habituation effects, as well as improving C_eff. From 2 years on, the WRS of all groups remained stable up to the 5-year appointment. As noted by Müller et al. [34], patients, who have received inadequate or no treatment for a long time prior to implantation, are often sensitive to loudness and refuse appropriate amplification. We observed only 3 patients (Tutopatch n = 1, RWS n = 2) with documented sensitivity to loudness who did not accept an optimized fitting even after years, and failed to improve speech reception (WRS ≤50%). However, most patients could adapt to amplification over a few fittings (habituation effect). This effect was particularly noticeable in group HCV2, in which 6 patients were initially sensitive to loudness. With increased amplification, some patients improved by up to 55 pp in WRS at 2 years, resulting in a statistically significant improvement in group HCV2. In group HCV2, the WRS improved statistically significant over time from IA to the best mean WRS of 90.8% ± 8.1% at 2 years. Groups Direct and RWC showed a WRS with low SDs at 2 years of 87.5% ± 8.3% and 82.5% ± 11.8%, respectively, suggesting reproducibility of audiological results. In comparison, Zahnert et al. [20] found a slightly lower mean WRS of 72.2% ± 15.6% after 3 years in a group of 9 patients with RWC. Larger variations can be found at 2 years in groups Fascia, Tutopatch, and RWS, with a mean WRS of 70.0% ± 23.3%, 80.0% ± 17.5%, and 75.6% ± 24.0%, respectively. Böheim et al. [16] reported an even lower mean WRS of 65% ± 30% after 3.3 years in a group Fascia including 12 patients. In regard to group RWS, varying results were reported, ranging from 53.8% ± 18.2% [37] to 78.1% ± 15.1% [34], indicating a certain unpredictability.

In clinical routine, we observed a ceiling effect of the aided SF threshold at approximately 25 dB HL. Nevertheless, an average speech recognition score ≥80% was provided. One possible cause is the limitation of the SF threshold by the intrinsic microphone noise at approximately 20 dB HL. In case of patients with a BC threshold approximately <20 dB, the noise is audible and masks softer sounds, thus limiting the aided SF threshold. This effect can result in a negative EG giving the false impression of insufficient amplification.

In order to compare the groups, it would be required to divide them into subgroups according to their BC PTA₄. For simplicity, we focused on changes in average EG over time. An overclosure of the air-bone gap (positive EG) was achieved at IA in groups Direct, Fascia, and RWS. Similar to speech recognition, the EG improved in all groups over time, with statistically significant improvements for both the RWC and HCV2 groups between IA and 2 years, resulting in a positive EG in all groups except group Tutopatch at 2 years.

The C_eff is a parameter evaluating the transfer function between an AMEI and the inner ear. The V_thr measurement via the Symfit^® software was developed by MED-EL based on the classical incus coupling [34, 38]. Thus, the C_eff indicates whether other types of vibroplasties are better or worse than the classical incus coupling. However, as a relative measure it is suitable for the comparison of different coupling modalities between groups and in the same type of vibroplasty, as shown in this study.

The lowest C_eff was found in group Direct (18.1 dB at 2 years, 14.7 dB at 5 years), although with only 4 patients, results should be interpreted carefully. Atturo et al. [39] demonstrated that the RW has a unique shape and size in each patient, and its diameter might be smaller than one of the FMT (1.5 mm). In case of direct coupling, contact may occur between the FMT and the bone surrounding the RW, resulting in a coupling deficit [40]. We have no information on the shape of the RWs of the implanted patients, but it can be assumed that the patients in group Direct had a, RW diameter ≥1.5 mm to place the FMT. The risk of diameter mismatch and the resulting challenges were the reasons for the adoption of indirect coupling, initially with fascia or Tutopatch and later with a coupler.

Even though the mean C_eff remained stable in group Fascia, we see some patients with a deterioration in the C_eff, which could indicate that fascia is degraded by the body [41, 42]. This assumption is supported by the observations made during one of the two reported total revisions in group Fascia. In this case of FMT migration after 4.9 years, no fascia was found in front of the RW during surgery. The Tutopatch group showed the worst mean C_eff with 26.6 dB ± 15.9 dB after 2 years and the largest variability among the groups. One possible reason might be that Tutopatch is placed directly in front of the membrane prior to FMT placement, which might hinder visual feedback and hence optimization of the contact during positioning.

Couplers have been developed to optimize the C_eff and, in particular, to reduce its variability. While in all other coupler groups the C_eff can be assumed to be stable, group RWC showed a mean C_eff of 26.7 dB ± 8.1 dB at IA, but decreased statistically significant at both 2 years and 5 years to 16.8 dB ± 5.7 dB, suggesting an improved transfer. We assume that the spherical shape of the RWC centers the coupler in the RWM, leading to an optimized coupling. While the C_eff in group RWS with 19.5 dB ± 8.8 dB at 2 years was comparable to the other investigated coupler groups, the mean WRS with 75.6% ± 24.0% at 2 years was inferior to most other modalities with a high variability between patients. A comparable WRS of 78.1% ± 15.1% was reported in Müller et al. [34], but simultaneously they reported a lower (better) C_eff of 12.7 dB ± 11.0 dB. This leads to the assumption of a floor effect, where a lower efficiency no longer provides additional benefits as suggested by Müller et al. [19]. The smallest variations and minor changes in C_eff were observed in group HCV2, indicating that the coupler remains in position regardless of alterations in the tissue or bone surrounding the RW.

We recommend a stringent documentation and evaluation of surgical boundary conditions in RW vibroplasty to enable identification of contributing factors, such as anatomical features of the RW, preexisting or due to previous surgeries and drilling-related characteristics such as speed, duration, and depth. This may help identify reasons for differences in results, particularly in relation to residual hearing.

This is the first comprehensive study comparing the most common different RW vibroplasties in terms of safety, residual hearing, and audiological outcomes. Although RW vibroplasty is a complex and challenging ear surgery, it is not associated with a higher risk than other vibroplasties or comparable complex ear procedures.

All investigated coupling modalities ensured a successful treatment benefit for conductive HL and mixed HL, providing rehabilitation of speech recognition. Moreover, a comparison between the groups showed that patients with RWC and HCV2 couplers tended to perform noticeably better in terms of aided speech understanding and C_eff than those in the other groups.

All patient data were collected during routine visits, and processing was anonymized in accordance with regulation (EU) 2016/679 of April 27, 2016, on the protection of natural persons with regard to the processing of personal data. The study was approved by the local Ethics Committee (Hannover Medical School, Approval No. 1897–2013) and is in accordance with the ethical standards of the Declaration of Helsinki. Written informed consent for anonymous use of data was obtained at the admission of patients.

All authors received travel support by MED-EL to conferences. The authors disclose no other conflicts of interest.

N.K. is funded by a MED-EL project grant to the Hannover Medical School. This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy – EXC 2177/1 – Project ID 390895286 and by the DFG as part of the program “Open Access Publication Costs.”

N.K.: writing – original draft, conceptualization, visualization, methodology, investigation, and formal analysis. S.B.: writing – review and editing, methodology, and supervision. T.L.: writing – review and editing, and resources. H.M.: writing – review and editing, conceptualization, methodology, supervision, and resources. All authors discussed the results and implications of this study.

The data supporting the study’s results are not publicly available as it could compromise the privacy of research participants but are available from the corresponding author (N.K.) upon reasonable request.