Percutaneous bone-anchored hearing implant surgery: inside or outside the line of incision?

For this cohort study, all adult patients (aged 18 years or older) who received any type of percutaneous BAHI at our clinic between 1 January 2010 and 31 January 2014 were identified from our Bone Implant database. Patients operated by two surgeons, A and B, were selected. Both surgeons use the same standardised linear incision technique; they were trained and work in the same centre. Surgeon A places the implant outside the line of incision on a consistent basis, while the other surgeon B consistently uses the technique with placement of the implant inside the line of incision.

Eligibility criteria were: one staged procedure with tissue reduction, initial placement of the implant (no previous implant loss or removal) and availability of the medical record including at least one postoperative visit at our outpatient clinic.

In the selected study period the simplified linear incision technique with subcutaneous soft tissue reduction was consistently used [14]. In this procedure, a longitudinal incision of approximately 30 mm is made with the optimal site of implantation being approximately 50 to 55 mm posterosuperiorly to the ear canal. The next step is the exposure and mobilisation of the periosteum after sharp dissection of the subcutaneous tissue. Subsequently, the implant is placed and there will be resection of subcutaneous tissue over an area of approximately 2 cm around the incision. The remaining periosteum will be removed. In the final step of the surgical procedure, surgeon A punches the skin next to the incision while surgeon B punches the skin in the line of incision, consequently placing the implant outside or inside the line of incision, respectively .

The first postoperative visit was a week after surgery, when the healing cap and gauze with antibiotic ointment were removed, followed by an inspection of the incision. All patients received topical therapy with hydrocortison/oxytetracycline/polymyxine B for 2 weeks during the first postoperative visit. This visit was followed by an appointment for fitting of the sound processor after 3–6 weeks. Further follow-up was in general after 3 and after 12 months. Extra visits could be initiated by physicians or patients depending on arising problems or individual needs. In addition, some patients visited the outpatient clinic more often because they participated in clinical trials [16, 17]. At each visit, there was registration of the degree of skin reaction, using the Holgers grading system, and therapeutic intervention if applicable [9].

Data were obtained from our Bone Implant database and patients medical charts. Information about incision technique, surgeon and implant type was collected. Unless otherwise described in the operative report, it was registered that surgeon A placed the implant outside the incision line and surgeon B placed the implant inside the line of incision.

All follow-up visits by one of our physicians, residents or specialised nurses were included in the analysis; consultation by telephone was not included. The notes from the physical examination were used to determine the presence and timing of a soft tissue reaction. A Holgers classification 2 or higher was registered as an adverse soft tissue reaction and a Holgers classification of 1 or higher was classified as a soft tissue reaction. The reason for this distinction was because of the clinical consequences of the adverse soft tissue reaction, namely an indication for (local) treatment. Additionally, if the Holgers notation was missing but there was notation of any redness, swelling, moistness and/or granulation around the titanium skin-penetrating implant in the medical record, this was still interpreted as presence of a soft tissue reaction. A soft tissue reaction was considered not present in case of a Holgers score 0 or no notation of inflammation of the skin in the notes of the physical examination. A lack of description about the tissue surrounding the implant in the notes of a follow-up contact was considered as missing data.

Therapeutic interventions for skin problems were recorded per visit. Conforming to the general protocol in our hospital, all patients received topical therapy with hydrocortison/oxytetracycline/polymyxine B during the first weeks post-surgery. This topical therapy was therefore not considered as a therapeutic intervention in our study. Alternative therapeutic interventions were distinguished: topical antibiotic ointment, healing cap replacement and revision surgery (change of the abutment and/or soft tissue revision). Implant loss was registered as well.

Finally, the background characteristics mental retardation, dermatological disease and diabetes mellitus were registered, because recent studies focus on identification of these comorbidities as possible risk factors in the context of soft tissue reactions after BAHI surgery [7, 18‐20]. If there were no notes for these conditions in the medical chart or any correspondence within the chart of the patient, this comorbidity factor was considered absent. End of the follow-up was defined as the last visit before March 2015.

A total of 202 patients and 211 implants could be included in the cohort in the period from 1 January 2010 until 31 January 2014. Surgeon A placed the implants consistently outside the line of incision and surgeon B placed the implants consistently inside the line of incision. Three exceptions were retrieved, in which the purpose always was to place the implant inside the line of incision. However, after closure, due to anatomical variation, it turned out the implant was outside the line of incision. From all implants, 96 BAHIs were placed outside the line of incision. The mean age in this group was 55 years (range 18–85 years, SD ± 16) and the median follow-up was 653 days per implant [interquartile range (IQR) 337–1058 days]. There were 115 implants placed inside the line of incision. The mean age was 53 years (range 18–83 years, SD ± 15) and the median follow-up was 548 days per implant (IQR 353–1046 days). A number of 81 of 202 patients participated in a clinical trial with a more extensive (standard) follow-up protocol, similarly distributed over the two different cohorts. All the baseline characteristics of the patient population are shown in Table 1. No significant differences in these baseline characteristics between both groups were noticed. The use of longer abutments was equally distributed between the study groups, however, slightly more previous generation implants and abutments were used in the inside group. In Table 2 the other surgical characteristics are summarised. The length of all implants was 4 mm.

Three implants were lost during complete follow-up (1.4 %). All these implants were placed outside the line of incision. One implant was lost 3 days after surgery. The medical chart reported a poor quality of the temporal bone. The other implants were lost after 46 days and after more than 3 years (this patient suffered from recurrent infections with peri-implantitis in the period prior to implant loss).

During the complete follow-up, in 18 of 211 implants, one or multiple revisions were performed (8.5 %). In the group with the implant outside the line of incision, revision surgery was performed in 5 of 96 implants (5.2 %). In the set of implants placed inside the line of incision, revision surgery was undertaken for 13 of 115 implants (11.3 %). This difference in performed revision surgery between both groups, as calculated with a log-rank test, was not significant (p = 0.129). An overview of the revision surgery and other therapeutic interventions in both groups is given in Table 3.

The outcome was divided in the presence of a soft tissue reaction (i.e. Holgers grade 1 or higher) and the presence of an adverse soft tissue reaction (i.e. Holgers grade 2 or higher). In 6.7 % of the follow-up contacts a notation of a soft tissue reaction was available but no Holgers classification was given, and in 3.7 % of the follow-up contacts a description about the tissue surrounding the implant was missing. A soft tissue reaction Holgers ≥1 was noticed in 47 implants (49.0 %) when the implant was placed outside the line of incision compared to 70 implants (60.9 %) which were placed inside the line of incision. The median time until the first soft tissue reaction was 90 days (IQR 21–366 days) and 95 days (IQR 44–344 days), respectively.

An adverse soft tissue reaction, Holgers grade ≥2, was registered in 17 implants (17.7 %) when the implant was placed outside the line of incision. In the group of implants placed inside the line of incision, 20 implants (17.4 %) presented with a Holgers ≥2. The median time until the first adverse soft tissue reaction was 363 days (IQR 127–675 days) and 183 days (IQR 112–370 days), respectively.

For both outcome measures a survival curve was calculated by the Kaplan–Meier method; the Kaplan–Meier curves and survival tables are shown in Figs. 2 and 3 and Tables 4 and 5. The Kaplan–Meier curves show the probability of surviving, i.e. not encountering an (adverse) soft tissue reaction, in a given length of time. The corresponding survival tables provide additional information about the cumulative events (CE), remaining cases (RC) and cumulative proportion surviving (CPS) at given points in the time during the follow-up. In these tables, the cumulative events are defined as the number of implants with (adverse) soft tissue reactions and the remaining cases are the implants still in the follow-up without soft tissue problems. The term cumulative proportion surviving can be explained as a statistical representation of the proportion of implants that have not reached the terminal event (i.e. skin reaction) by the end of an interval. A log-rank test was executed to compare the survival curves of the two surgical techniques. No significant differences were found between the two groups for both the presence of soft tissue reactions (p = 0.322) and a Holgers score of 2 or higher (p = 0.951) during the follow-up.