Stability, survival, and tolerability of a 4.5-mm-wide bone-anchored hearing implant: 6-month data from a randomized controlled clinical trial

The test implant was the wide Ponto implant (diameter 4.5 mm, length 4 mm) and the control implant was the previous generation Ponto implant (diameter 3.75 mm, length 4 mm). Both implants used the same 6-mm abutment. The implants and abutments are developed and manufactured by Oticon Medical AB (Askim, Sweden) and are displayed in Fig. 1.

Out of all of the patients indicated for a percutaneous bone conduction device in our center, 57 adult patients with a total of 60 implants were consecutively included. Eligibility criteria were as follows: indication for a percutaneous implant; age of 18 years or older; bone thickness of at least 4 mm at the implant site; written informed consent given; abutment of 6 mm required (not longer); ability to participate in follow-up visits; no history of psychiatric diseases; no mental disabilities; no presumed doubt, for any reason, that the patient would not be able to attend all follow-up visits; no presence of diseases or use of treatments known to compromise bone quality at the implant site (e.g., radiotherapy, osteoporosis, diabetes mellitus).

The current study was designed as an open randomized controlled clinical trial in our tertiary referral center. The primary outcome parameter was implant stability measured as ISQ low values in the first 6 months after implantation. Secondary objectives were to compare ISQ high values in the same period, ISQ low and high values at all visits, time to stability dip (in ISQ low) if applicable, implant survival, soft tissue reactions during all visits, and quality of life outcomes.

The sample size was based on the primary efficacy variable. A weighted average of ISQ low values during the 6-month follow-up period was obtained by the mean area under the curve (AUC) calculation using the trapezoid rule with all ISQ low measurements over the first 6 months. Data from a similarly designed previous study [11] were used for the sample size calculation. An expected difference of 4.5 in the mean AUC of the ISQ low values of the test and the control groups, with unequal standard deviations of 2.8 and 5.5, respectively, were used for determining the sample size. A 2-sided t test with Satterthwaite’s correction for unequal variances was performed. For a power of 90 %, significance level of 0.05, and randomization ratio of 2:1, a total of 60 implants needed to be included.

Randomization was performed in a 2:1 ratio (test–control). A computer-generated list of random allocations was used. The group assignments were enclosed in sequentially numbered opaque sealed envelopes. The randomization was blinded to the patients and investigators until the surgery was performed. Patients were allocated in consecutive order. Blinding of the investigators after the group assignments were made was not feasible because the appearances of the implants and instruments used during surgery were clearly different. Because most patients were operated under local anesthesia, the patients were also not blinded.

Implants and abutments were placed in a single-stage surgical procedure. The linear incision technique with subcutaneous tissue reduction was applied in all cases [12]. Implants were alternately placed within or posterior to the incision line. In accordance with the study protocol, follow-up visits were scheduled at 7, 14, 21, and 28 days; 6 and 12 weeks; and 6 months. At each visit, resonance frequency analysis (RFA) was used to establish the implant stability quotient (ISQ). RFA uses magnetic pulses generated by the Osstell ISQ device (Osstell AB, Göteborg, Sweden) to excite the SmartPeg (type 55) that is connected to the abutment, which leads to vibration of the implant–abutment system. The intensity of these vibrations is analyzed by the device that computes the ISQ, which is an indication of the rigidity of the implant–bone interface [13]. Perpendicular measurements result in an ISQ high value and an ISQ low value. At each visit, the skin status was also assessed according to the Holgers classification [14]. Three weeks after surgery, the patients were fitted with the bone conduction device. The benefit of the bone conduction system was assessed using 3 questionnaires: the Abbreviated Profile of Hearing Aid Benefit (APHAB) [15], the Glasgow Benefit Inventory (GBI), and the Glasgow Health Status Inventory (GHSI) [16]. APHAB and GHSI outcomes were only included in the analysis when both the baseline screening before implantation and the 6-month evaluation had been completed. In cases where patients used hearing aids at the baseline evaluation, they were asked to complete the baseline questionnaire both for the aided and unaided conditions. The unaided condition was used as the baseline measurement for analyzing the benefit of the bone conduction system at 6 months.

Data management and statistical analyses were performed by external data managers and biostatisticians (Statistiska Konsultgruppen, Göteborg, Sweden) according to a predefined statistical analysis plan.

For implant variables, bilaterally implanted patients who received both a control and a test implant were included in both analyses. Patients who received 2 test or 2 control implants were represented by the mean of the 2 measurements for continuous variables or the worst value for categorical variables. For patient variables, bilaterally implanted patients who received both control and test implants were included in descriptive statistics but excluded in formal analyses on the patient level.

All tests were 2 tailed with significance levels of 0.05 and were executed using SAS v9.2 and v9.3 software (SAS Institute Inc., Cary, NC, USA).

This clinical investigation was performed in accordance with the current version of the Declaration of Helsinki (Washington 2002, ISO 14155), Good Clinical Practice (International Conference on Harmonization Good Clinical Practice) and was approved by the local ethical committee.

Fifty-seven patients with a total of 60 bone-anchored hearing implants were included in the randomization, with 39 implants in the test group and 21 in the control group. Surgeries were performed between June 2012 and January 2014. Baseline demographic information is shown in Table 1. No significant differences were found between the test and control populations. Three patients received bilateral implants; 2 of these patients were randomized for both a test and a control implant, and 1 patient received 2 test implants. All randomized patients received their allocated treatment and could be included in the final 6-month analysis.

The mean AUC for ISQ low was 64.4 (SD 3.0; range 55.5–70.1) for test implants (n = 38) and 59.5 (SD 2.2; range 55.5–63.5) for control implants (n = 21). The difference between these groups of 4.9 ISQ points (95 % CI 3.4–6.4; p < 0.0001) was statistically significant. For ISQ high, a difference of 3.2 (95 % CI 1.7–4.7; p < 0.0001) was observed during the 6-month follow-up, with a mean AUC of 65.8 (SD 2.7; range 57.0–70.5) for the test implant and 62.6 (SD 2.8; range 56.9–66.8) for the control implant. At all follow-up visits, statistically significant differences in mean ISQs between both groups were recorded. The results are displayed in Fig. 2. The mean increase in ISQ from baseline is statistically significant in both groups; however, the increase in ISQ from baseline for the test implant is statistically significantly stronger compared to the increase for the control implant. The ISQ dip at 42 days for the test implant can be ascribed to a single implant that displayed a very low ISQ (ISQ low 46, ISQ high 52) but remained clinically stable and presented with an ISQ within the normal range at the next follow-up appointment.

No dip in mean ISQ was observed, as the ISQ high and ISQ low values were higher than the baseline ISQ values (at surgery) at all follow-up visits.

Implants were loaded 3 weeks after surgery (with a 2-day range) in all but 1 patient (loaded at 24 days). This early loading moment did not seem to influence ISQ values, as these progressed positively in both implants.

At 6 months, a mean increase in the ISQ low from the time of surgery of 4.5 (SD 4.6; range −4 to 29) was observed for the total group (n = 59), which was significantly different from the ISQ low at the time of surgery (p < 0.0001). The mean increase was 5.2 (SD 5.0; range −4 to 29) in the test group and 3.2 (SD 3.7; range −3 to 13) in the control group. The mean difference in the increase in ISQ low between both groups was statistically significant (95 % CI −0.5 to 4.5; p = 0.03).

No implants were lost during the follow-up period. In each experimental group, 1 implant required surgical revision of the soft tissue; 1 patient who suffered from psoriasis presented with insufficient skin healing after surgery and the other patient presented with skin partially overgrowing the abutment. Three implants (7.9 %) in the test group and 2 implants (9.5 %) in the control group developed adverse skin reactions (Holgers grade 2–4). Results related to soft tissue reactions are displayed in Fig. 3. The analysis of soft tissue statuses throughout the follow-up period revealed findings of Holgers grade 0 in 86.8 % (test) and 89.0 % (control) of visits, Holgers grade 1 in 12.1 % (test) and 9.1 % (control) of visits, Holgers grade 2 in 1.1 % (test) and 1.3 % (control) of visits, Holgers grade 3 in 0.0 % (test) and 0.6 % (control) of visits, and no Holgers grade 4 cases over all of the visits. Two out of the 5 patients who presented with adverse skin reactions suffered from skin diseases. Furthermore, no statistically significant differences were noted in other postoperative complications: bleeding or hematoma [1 test (2.6 %) vs 1 control (4.8 %) implant], pain or numbness [4 test (10.5 %) vs 2 control (9.5 %) implants], and wound dehiscence [2 test 5.3 % vs 3 control (14.3 %) implants]. Additionally, skin height did not differ between the groups.

The GBI questionnaire was completed 12 weeks after surgery. Eight patients completed the questionnaire outside of the defined visit window (mean of 22 days after the planned visit date). These results were still included in the final analysis. No differences were observed in the outcomes between the test and control groups. The results are shown in Table 2.

All patients completed the APHAB and GHSI questionnaires 6 months after surgery. However, 5 patients did not complete baseline questionnaires and were consequently excluded from the benefit analysis. One additional patient did not complete the baseline APHAB, while another 3 patients were excluded from the benefit analysis using the GHSI because of incomplete data on the 6-month questionnaire. The outcomes of these questionnaires are displayed in Fig. 4. For the GHSI, significant improvement was observed for the total and general scores, but not for the social and physical subscales. The APHAB indicated that there was statistically significant improvement on all of the subscales in the aided condition compared to the unaided condition.