The system comprised of two parts, the external audio processor (AP) and an internal Vibrating Ossicular Prosthesis (VORP), which together convert sound in the environment into a vibratory signal delivered to the inner ear. The AP contains a microphone, which picks up sound and speech from the environment and converts it into a signal that can be transmitted across the skin to be received by the implanted internal receiver of the VSB. The implantable part consists of an internal coil, a magnet to hold the audio processor over the implant, a demodulator, a conductor link and a FMT. The FMT can be coupled to either the ossicular bones or to the round window.

Preoperatively, all subjects underwent comprehensive audiological and medical evaluations, including pure tone audiometry, sound field audiometry (with warble tones), speech audiometry, computed tomography of temporal bones and magnetic resonance imaging. All audiometric testing was performed in a standard audiometric booth at the Audiology Centre of the Prince of Wales Hospital. Air conduction thresholds were obtained at 0.25, 0.5, 1, 2, 4, and 8 kHz and bone conduction thresholds at 0.5, 1, 2, and 4 kHz. Speech recognition in quiet and in noise was measured by the Cantonese Hearing in Noise Test (CHINT) (8). Unaided speech thresholds were evaluated in quiet in three conditions: speech front (SF), speech implant side or speech ipsilateral (SI) and speech at non-implant side or speech contralateral (SC). The speech tests were also evaluated at a fixed level of 65 dB (A) of noise in three different noise conditions. The speech signal was always presented from the front (0⁰ azimuths). The noise was presented either in the front (NF), or from the implant side or noise ipsilateral (NI) or noise from the non-implant side or noise contralateral (NC). If a patient had better hearing in the non-implant ear, the non-implant ear would be occluded during testing to minimize the involvement of the non-test ear via air conduction. The equipment used included the GSI clinical audiometer (Grason-Stadler, Eden Prairie, MN, USA) with telephonic headphones and the Maico HINT system (the non-test ear via air conduction). The equipment used included the GSI clinical audiometer. During sound field testing and the speech testing, the subjects were positioned one meter away from the speaker. Apart from above measures, all subjects who were hearing aid users were also asked to fill in the Chinese version of the Abbreviated Profile of Hearing Aid Benefit (APHAB-CH) and the International Outcome Inventory of Hearing Aid (IOI-HA) to compare the performance between hearing aid and VSB.

The APHAB-CH is a Chinese self assessment inventory first described in 1995 to evaluate the disability associated with hearing loss and the reduction of the disability with amplification (9). The validated Chinese version of APHAB was used in the present study (10). The APHAB comprises 24 items in four subscales: ease of communication (EC), background noise (BN), reverberation (RV) and aversiveness of sound (AV). All items are statements about communication difficulties in daily life and respondents are asked to indicate their agreement with each statement on a seven-point scale. The mean rating for the six items in each subscale constitutes the subscale score, and the global APHAB score is computed by taking the mean of EC, BN, and RV subscales. Subjects were asked to complete the inventory at baseline, before implantation based on their experience with their hearing aid if any and then after 3 months and 6 months of use of the VSB.

IOI-HA is a brief seven-item questionnaire designed to evaluate the effectiveness of hearing aid treatment in everyday life. The aspects examined are: 1) hearing aid use (Use), 2) benefit (Ben), 3) residual limitation in activity (RAL), 4) satisfaction (Sat), 5) residual participation restriction (RPR), 6) impact on others (Ioth) and 7) quality of life (QoL). The validated Chinese version of the IOI-HA was used in the present study (11). The IOI-HA questionnaire involved the performance rating on a five-point scale.

This approach is suitable for patients without a working ossicular chain. We performed the operation under general anesthesia with facial nerve monitoring. Post-auricular incision was employed. For patients with wall-down mastoidectomy done before the surgery, we carefully elevated the meatal flap and exposed the round window by adequate inferior canal drilling. For patients without wall-down mastoidectomy done before the surgery, we approached the round window by posterior tympanotomy approach as stated above. When the round window niche was identified, the overhanging bony lip of the round window niche was removed by drilling. A small bony well was drilled at the hypotympanium region next to the round window area to harbor the FMT. A layer of perichondium was placed on the round window membrane, then the FMT was placed on the perichondrium and coupling with the round window. The coupling system was stabilized by addition perichondrium packing around the FMT. The VORP was fixed into the bony well created in the temporal region by non-absorbable stitches.

This approach is suitable for patients with functioning stapes and the anatomy of stapes allows a safe and stable coupling. We performed the surgery under general anaesthesia via the posterior tympanotomy approach. Left cortical mastoidectomy was first performed. The facial recess was opened with preservation of chorda tympani and facial nerve. Intra-operation facial nerve monitoring was needed throughout the drilling to prevent facial nerve injury. The facial recess was gradually enlarged by a diamond burr until the whole stapes suprastructure and the round window niche were visible. The tympanic membrane was kept intact throughout the surgery. Copious irrigation was needed to remove the retained blood clot and bone dust. The normal mobility of stapes was confirmed with the gentle needle palpation and the presence of round window reflex. The claw of the FMT was fashioned so as to allow secure crimpling onto the stapes head. The stability of the whole FMT-stapes system was confirmed with gentle needle palpation. VORP was fixed into the bony well created in the temporal region by non-absorbable stitches.

For both surgical techniques, intraoperative electrocochleography (ECochG) was performed using a GSI Audera evoked potential system (Grason-Stadler). Electrodes were placed on the promontory and measurements were performed before closing up. Alternating tone pips at 2 kHz were delivered to the implant via a custom-made audio processor. A series of ECochG was measured to ensure that the ECochG threshold was matched with the subject's pure tone bone-conduction threshold at 2 kHz.

The average surgery time for each patient was 3.3 hours (200±63.4 minutes). There were no intra-operative and postoperative surgical complications encountered for any patient. The device was activated 8 weeks after surgery and the Amade audio processor (MED-EL) was programmed with CONNEXX 6.1.1 software equipped with the Symfit database Rev. 5.0. The default Desired Sensation Level (DSL) amplification strategy was applied during first fit and the patients returned for a review and fine-tune twice at 2 week intervals until the most comfortable self-evaluated listening level was reported by the patients.

Three months and 6 months after activation of the implant, patients were asked to come back for postoperative assessment. They underwent pure tone audiometry, sound field testing and CHINT testing with the device turned on. The same test equipment and materials as those employed in the preoperative assessment were used during testing.