Blog 2019-1: Percutaneous bone conductor or (active) middle ear implant to treat mixed hearing loss? No clear winner yet


To rehabilitate patients with mixed hearing loss, implantable devices like the Vibrant Soundbridge middle ear implant (VSB) or percutaneous bone conductors (BCDs) have been applied. Such implantable devices become the next option whenever conventional hearing devices (e.g. behind-the-ear device) or reconstructive surgery are not possible or probably not effective (e.g. in case of aural atresia or chronic draining ears). Using recent publications, describing clinical trails, a comparison is made between audiological outcomes obtained with either power BCDs and VSBs. Studies in patients with severe mixed hearing loss were considered only if the mean sensorineural hearing loss component was at least 35 dB HL (arbitrary choice).

Medline was used to find relevant studies, published between 2016 and early 2019. Of the 30 detected studies, 25 studies were excluded (see Materials and Methods at the end of this blog for details) and five studies remained:

Bush et al. (2016) published data 58 patients with mixed hearing loss using the VSB with its actuator coupled to either the round window (RW) or the oval window (OW); Lee et al., (2017) and Zahnert et al. (2019) also published such VSB data, on 19 and 24 patients, respectively. Bosman et al. (2018) published data on the application of a recently introduced percutaneous power BCD, viz. the Baha 5SP, applied in 10 patients; Bosman et al. (2019) published also data on another percutaneous power BCD, the Ponto 3SP device, applied in 18 patients.

The Methods and Materials section is added at the end of this blog.


Figure 1 presents the frequency-specific gain values related to the (bone-conduction) threshold; it should be noted that BCDs as well as VSBs with actuator coupled to the cochlea, directly stimulate the cochlea, bypassing the impaired middle ear (and thus the air-bone gap). Therefore, to assess the effectiveness of the device fitting, only the ‘effective gain’ is of importance, defined as the cochlear (bone-conduction) thresholds minus the aided thresholds (e.g. Busch et al., 2016; Snik, 2018).

The figure presents the ‘effective gain’ divided by the cochlear hearing loss thus the bone-conduction threshold, referred to as the gain-threshold (GT) ratio, plotted as a function of frequency (0.5 kHz to 4 kHz). If GT=0, it means that the aided threshold and the cochlear thresholds coincide, indicating no ‘compensation’ of the cochlear hearing loss. Regarding the VSB, the subfigure at the right shows the calculated results using Zahnert et al. (2019), Lee et al. (2017) and Busch et al. (2016), separately. As these curves show a similar trend, data were averaged after weighting for the number of patients per study. The mean VSB data in the left subfigure show the highest GT ratio at 2 kHz with roll-off in higher and lower frequencies. For the Baha 5SP, a broader frequency response is seen comparable to that of the Ponto 3SP; obviously, the latter device is less powerful (lower GT values).

The target GT ratio, according to the well-validated NAL-fitting procedure, is also indicated (Dillon, 2012; see also Materials an Method section at the end of this manuscript). Systematic differences are seen between the measured GT ratios and that NAL-target GT ratio, especially in the high and low frequencies. At 2 kHz the measured values are closest to the NAL-GT target. 


Figure 1. The gain-threshold (GT) ratio presented as a function of frequency. Results of the included studies are shown (individual data of the 3 VSB studies, presented in right subfigure, were averaged). According to the NAL-RP fitting rule, the target GT ratio is approx. 0.45 (indicated in the figure) 

The NAL fitting rule is well established for sensorineural hearing loss (Dillon, 2012). However, compared to conventional behind-the-ear devices, BCDs and middle ear implants have a limited dynamic range caused by a relatively low MPO (maximal power output). Therefore, a modified, pragmatic version of the NAL fitting rule has been introduced for these implantable devices (Snik et al., 2019; Materials and Methods section), meant primarily for validation purposes. Figure 2 shows target-aided


Figure 2. The difference between prescribed aided thresholds and measured aided thresholds as a function of frequency. The modified NAL-procedure was used. Pooled VSB data are presented, and, separately, the data of the two percutaneous BCD studies

thresholds (obtained with this pragmatic validation rule) minus the measured aided thresholds. A value of 0 means that the result is adequate according to that rule; negative values indicate insufficient gain. Weighted means are presented of the pooled VSB studies and, separately, the results of the Ponto 3SP study and Baha 5SP study. In agreement with Figure 1, most obvious discrepancies are found at 0.5 kHz and 4 kHz, especially for the VSB device, in the order of 15-20 dB. The Baha 5SP seems to be the most adequate device with a rather flat curve around 0; according to Figure 2, the Ponto 3SP device lags behind by approx. 10 dB, what might be ascribed to the difference in MPO of the two BCDs. The MPO of the Baha 5SP is approximately 9 dB higher than that of the Ponto 3SP (Schwartz, 2018). The MPO of the VSB and the Baha 5SP are rather comparable (Chapter 2 in Snik, 2018).

Apart from aided thresholds, all the five studies presented aided speech recognition scores in quiet, however, measurement conditions and the tests used were divers. Nevertheless, in all the studies monosyllabic words were used. Bush et al. (2016) and Zahnert et al. (2019) presented word scores (WRS) obtained at a presentation level of 65 dB SPL. Bosman et al. (2018; 2019) reported phoneme scores (PS) of words presented at 60 dB SPL; which can be recalculated to a word score at 65 dB SPL (applying WRS=PS2.3 and using the slope of the speech performance-intensity curve: 6% per dB). Lee et al. (2017) measured word scores presented at the patient’s most comfortable listening level instead of at a fixed level and, thus, cannot be used for a comparison.

Table 1 shows the outcomes.

Table 1. Speech recognition scores

Study/test VSB, Zahnert VSB, Bush Ponto 3SP Baha 5SP
WRS 65 dB SPL 67% 78%
PS at 60 dB SPL calculated WRS at 65 dB SPL 75%




Note WRS: word recognition score, PS: phoneme score

In summary, for patients with mixed hearing loss, the achieved (effective) gain with either VSB or the percutaneous Baha 5SP show similar results in the mid frequencies (Figures 1 and 2), however, at 4 kHz and especially at 0.5 kHz, best results are found with the Baha 5SP. Regarding speech recognition, results are rather similar (Table 1; the WRS data at 65 dB SPL input). Therefore, from an audiological point of view, differences between devices are limited.

Closing remarks

It is known that the VSB outcomes depend on the efficacy of the actuator coupling to the cochlea (e.g. Busch et al., 2016). New coupling options are being developed (e.g. Lenarz et al., 2018), which might have a positive effect on the frequency response. Another way to deal with the peaked frequency response is to lower the gain of the VSB device in the mid frequencies. However, that is only an option if the implanted device has sufficient reserve.

For the percutaneous BCDs, the coupling of the processor to the skull is more or less standardized and therefore not a major source of variance between subjects.

To choose a device, the performance is highly relevant but also of importance is the status of the middle ear (can a VSB be placed?), invasiveness and complexity of the surgery, risk of medical and technical complications, longevity, cost etc. Regarding the Baha 5SP, with high volume setting, it might be advised to use the behind-the-ear processor away from the ear, with a clip connected to clothing, to deal with feed-back. That might also be considered as a draw back.

Note that the conclusions drawn are valid for the present devices/processors. If more powerful processors or more effective actuators are introduced or if better coupling techniques are developed, conclusions might change.


Materials and methods

Search strategy

This review was carried out in accordance with PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses; Moher et al., 2009). Medline (Pubmed) was used to find relevant studies, published between 2016 and early 2019. Search string was: ((soundbridge OR ponto OR baha) AND (mixed hearing) NOT transcutaneous). Thirty manuscripts were identified; based on the abstracts, 20 manuscripts were excluded because: the focus was on cochlear implantation (1), comprised a review, no original data (5), case reports and how-we-do-it reports (4). Manuscripts dealing with medical/surgical issues (8), technical issues (1) and application in pure sensorineural hearing loss (1) were excluded as well. The full text of the remaining 10 manuscripts was reviewed. One more manuscript was excluded because the mean cochlear hearing loss was below 35 dB HL (Iwasaki et al., 2017), three manuscripts were excluded as the bone-conduction thresholds and/or aided thresholds were not presented frequency specific (essential information for the present analyses; Brkic et al., 2019; Olszewski et al., 2017; Monini et al., 2017) and one manuscript with just 5 relevant patients (Gregoire et al., 2018). Five studies were included; regarding VSB: Bush et al. (2016); Zahnert et al. (2019) and Lee et al. (2017). Regarding BCD, two studies were identified: Bosman et al. (2018) and Bosman et al. (2019).

Table 2 gives an overview of some relevant patients’ characteristics of these 5 studies (age and mean cochlear or bone-conduction thresholds). On the average, the patients using BCDs are older than those using the VSB.

Table 2. Some patient characteristics

Study; first author Lee, 2017 Busch, 2016 Zahnert, 2019 Bosman, 2018 Bosman, 2019
n 19 78 24 18 10
Device VSB VSB VSB Ponto 3SP Baha 5SP
Age, years 55±17 57±17 55 (34-75) 64 (22-80) 70 (59-77)
PTAbc dB HL 45 37 40 38 44

Note. n: number of patients; PTAbc: calculated mean bone-conduction threshold at 0.5, 1, 2, 4 kHz. Regarding the age at implantation, mean age is presented with standard deviation or the median age with range between brackets

Gain measurement

BCDs as well as the VSB with its actuator coupled to one of the cochlear windows stimulate the cochlea directly, bypassing the impaired middle ear. Therefore, the air-bone gap doesn’t play any role when fitting these devices and, to evaluate audiological benefit, the ‘effective gain’ is of importance. The ‘effective gain’ is defined as the cochlear (bone-conduction) thresholds minus the aided thresholds (e.g. Busch et al., 2016; Snik, 2018). The ‘effective gain’ assesses which proportion of the cochlear hearing loss is ‘compensated’ by the device. By definition, the often-used ‘functional gain’ equals the ‘effective gain’ plus the air-bone gap, and is mainly dominated by the width of the air-bone gap. In contrast, the ‘effective gain’ is a measure of the quality of the device fitting and can be compared to target gain values as prescribed by generally used, validated prescription procedures.

Given a certain hearing loss, obviously, the required ‘effective gain’ depends on the cochlear hearing thresholds (the higher the hearing loss, the higher the gain should be). Therefore, the ‘effective gain’ divided by the cochlear threshold (the gain/threshold ratio or GT ratio) is calculated; according to the widely used NAL-RP fitting procedure, this ratio should be approx. 0.45 for 1, 2 and 4 kHz, for proper amplification (Dillon, 2012).

However, for mixed hearing loss, that NAL-prescribed target GT ratio is not easily reached, owing to, amongst others, the limited highest sound level that can be produced by BCDs or VSBs (limited output), characterised by the ‘maximum output’ or MPO of these devices. Therefore, a pragmatically adapted NAL procedure has been introduced for these implants, to be used for validation purposes (Snik et al., 2019). This pragmatic procedure was based on 31 published clinical trials, regarding all types of BCDs and active middle ear implants and, consequently, this new procedure is device independent. The assumption is that appropriate (sufficient powerful) implantable devices have been chosen (regarding the MPO) and that amplification is linear. In short, aided thresholds should equal 0.55 times the cochlear threshold (following the NAL-RP fitting procedure) plus a margin of 5 dB with a minimum value of 25 dB HL at 1, 2 and 4 kHz (margins to deal with the limited MPO of the devices; Snik et al., 2019). Regarding 0.5 kHz, following the NAL-RP rule, calculated target aided threshold is 5 dB higher than that at 1 and 2 kHz, thus an additional margin of 5 dB is set at 0.5 kHz.


Bosman AJ, Kruyt IJ, Mylanus EAM, Hol MKS, Snik AFM. On the evaluation of a superpower sound processor for bone-anchored hearing. Clin Otolaryngol. 2018;43:450-455. doi: 10.1111/coa.12989.

Bosman AJ, Kruyt IJ, Mylanus EAM, Hol MKS, Snik AFM. Evaluation of an abutment-level superpower sound processor for bone-anchored hearing. Clin Otolaryngol. 2019 Epub ahead of printing. doi: 10.1111/coa.13084.

Brkic FF, Riss D, Auinger A, Zoerner B, Arnoldner C, Baumgartner WD, Gstoettner W, Vyskocil E. Long-Term Outcome of Hearing Rehabilitation With An Active Middle Ear Implant. Laryngoscope. 2019;129:477-481. doi:10.1002/lary.27513.

Busch S, Lenarz T, Maier H. Comparison of Alternative Coupling Methods of the Vibrant Soundbridge Floating Mass Transducer. Audiol Neurootol. 2016;21:347-355. doi: 10.1159/000453354.

Dillon H. Hearing aids, 2012; Thieme Verlag, Stuttgart

Grégoire A, Van Damme JP, Gilain C, Bihin B, Garin P. Our auditory results using the Vibrant Soundbridge on the long process of the incus: 20 years of data. Auris Nasus Larynx. 2018;45:66-72. doi: 10.1016/j.anl.2017.02.007.

Iwasaki S, Usami SI, Takahashi H, Kanda Y, Tono T, Doi K, Kumakawa K, Gyo K, Naito Y, Kanzaki S, Yamanaka N, Kaga K. Round Window Application of an Active Middle Ear Implant: A Comparison With Hearing Aid Usage in Japan. Otol Neurotol. 2017;38:e145-e151. doi: 1097/MAO.0000000000001438.

Lee JM, Jung J, Moon IS, Kim SH, Choi JY. Benefits of active middle ear implants in mixed hearing loss: Stapes versus round window. Laryngoscope. 2017;127:1435-1441. doi: 10.1002/lary.26244.

Lenarz T, Zimmermann D, Maier H, Busch S. Case Report of a New Coupler for Round Window Application of an Active Middle Ear Implant. Otol Neurotol. 2018;39:e1060-e1063. doi: 10.1097/MAO.0000000000001996

Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097.

Monini S, Bianchi A, Talamonti R, Atturo F, Filippi C, Barbara M. Patient satisfaction after auditory implant surgery: ten-year experience from a single implanting unit center. Acta Otolaryngol. 2017;137:389-397. doi: 10.1080/00016489.2016.1258733.

Olszewski L, Jedrzejczak WW, Piotrowska A, Skarzynski H. Round window stimulation with the Vibrant Soundbridge: Comparison of direct and indirect coupling. Laryngoscope. 2017;127:2843-2849. doi: 10.1002/lary.26536.

Schwartz, J. MPO Available to the patient: A comparison of the Cochlear™ Baha® 5 Power, Cochlear Baha 5 SuperPower and the Oticon Medical Ponto 3 SuperPower. 26-4-2018.

Snik AFM. Auditory implants. Where do we stand at present? 2018;, see chapters 2, 3 and 4.

Snik A, Maier H, Hodgetts B, Kompis M, Mertens G, van de Heyning P, Lenarz T, Bosman A. Efficacy of Auditory Implants for Patients With Conductive and Mixed Hearing Loss Depends on Implant Center. Otol Neurotol. 2019;40:430-435. doi: 10.1097/MAO.0000000000002183.

Zahnert T, Mlynski R, Löwenheim H, Beutner D, Hagen R, Ernst A, Zehlicke T, Kühne H, Friese N, Tropitzsch A, Luers JC, Todt I, Hüttenbrink KB. Long-Term Outcomes of Vibroplasty Coupler Implantations to Treat Mixed/Conductive Hearing Loss. Audiol Neurootol. 2019;23:316-325. doi: 10.1159/000495560.