For infants and children with severe to profound sensorineural hearing loss, either congenital or acquired, cochlear implantation has become standard care. Literature shows that implantation in pediatric patients at early age is beneficial for auditory development, and minimizes language delays that result from hearing loss (1, 2). Bilateral cochlear implantation (binaural stimulation) in children provides even more benefits, leads to increased audiophysiological stimulation of the auditory cortex at an early age, and is therefore the mainstay of treatment in children that meet implantation criteria in Dutch healthcare (3, 4).

Cochlear implantation surgery has been proven to be a safe procedure, with low complication rates. Revision surgery rates vary between 4.6% and 8.7%, and are mostly due to device failure as a result of which re-implantation is necessary (5–8). Complications that are not due to device failure, such as migration or protrusion of the receiver/stimulator (R/S) device, wound infection with implant extrusion or electrode misplacement or migration can also occur (5, 9–12). The complication rate reported in the literature varies greatly between studies with reported rates of 0.6% to 30.9% (9, 11–14). Due to broadening of the indication criteria and expected improved functional outcome after bilateral cochlear implantation, more children are receiving a CI and at a younger age (3, 15).

Recent publications stress the importance of recognizing the challenges associated with operating on young children in order to prevent complications (2, 16). The standard surgical technique for cochlear implantation in our center is the mastoidectomy with posterior tympanotomy approach (MPTA). The alternative suprameatal approach (SMA) has also been used, although Bruijnzeel et al. (14) reported a higher (infectious) complication rate when using this technique. It would be informative to update and assess these data with a prolonged follow up. Another important step in the surgical procedure is the positioning and fixation of the R/S device which can be achieved by several surgical techniques. The most used bony well technique, requires drilling a recess in the temporal bone in which the implant will reside. Usually a canal, tunnel or overhang is made for protection of the electrode array. Some CI surgeons use additional sutures, screws or wires to secure the implant. The less invasive subperiosteal pocket technique uses the soft tissue of the pericranium-temporalis muscle to fixate the implant (17). Both techniques of CI fixation have been used in our academic medical center over the years since the start of pediatric cochlear implantation in 1996. However, our experience is that drilling a bony well to accommodate the implant is not always feasible in young children due to insufficient skull thickness. Cochlear implant manufacturers advise a bony recess depth of at least 1.0–3.0 mm for sufficient fixation of the device, depending on the implant model (18–20). In order to lower the profile of the housing, an even deeper recess is required. This is challenging in infants, with their immature skull thickness. Furthermore, the dimensions of a cochlear implant demand a bony recess with a width of at least 30 mm to house the case. The curvature and irregularity of the temporal bone make embedment of a flat surface such as a CI challenging. Additionally, attempting drilling a bony well without preoperative imaging data or planning to measure thickness, introduces possible risks to the patient. However, these attempts of drilling would be redundant if we knew that drilling a bony well under a certain age is not feasible or necessary.

Previous studies describe an adaptation of the fixation technique where (partial) exposure of the dura is necessary and a bony island is left in the center to function as resistant and protective layer (21–23). These studies demonstrate the difficulty of drilling in young infants and the risks involved. Possible complications associated with drilling are dural tears with subsequent cerebrospinal fluid leakage as a direct result of drilling close to the dura (10, 21, 24). Other complications that have been reported (but occur very rarely) and associated with the bony well technique are late onset hematomas, epi-/subdural hematoma, tentorial herniation, and cerebral infarction, as well as meningitis (24–29).

Therefore, in this study we aim to assess the feasibility of drilling a bony well adequate for CI embedment in different age groups. We also aim to review the pediatric cohort implanted in our institution, reporting the occurred complications, revision and device failure rate using different surgical techniques.

This mono-center, retrospective and exploratory study was conducted at the University Medical Centre (UMC) Utrecht The Netherlands, in compliance with the principles of the Declaration of Helsinki. Exemption was granted by the local ethical committee (Institutional Review Board of the UMC Utrecht) (METC protocol 22/560) as a non-WMO study where consent was provided. The exemption included the CT data as well as clinical cohort data. All data was pseudonymized, thus exempt from acquiring informed consent.

The analysis of CT data of 192 ears of pediatric patients aged 0–17 years, showed that in the majority of the cases (79.7%, n = 153) it was not feasible to drill a bony well deep enough to lower the profile of the housing. The temporal bone thickness has been studied previously for the safety of implanting various bone-anchoring devices. In most cases a thickness of at least 3 mm was found in patients of five years and older. Below the age of five, several patients had a thickness of less than 3 mm. However, these studies used either a fixed location on the skull where the measurement took place (32, 33) or searched randomly within the segmented temporal bone (34). In our study, the search for an optimal location was systematic and the ROI was defined based on surgical practices for cochlear implantation. Our analysis of the most optimal location in the ROI showed that the mean minimum bone thickness for the age group 0–4 years was 1.84 mm with a range of 0.07 mm to 2.93 mm (Figure 3). These data confirm the difficulty and even impossibility in this age group, of drilling a bony well that complies with the advised dimensions of CI manufacturers. This is due not only to the depth of the recess but also the surface area that needs to be drilled out in order to accommodate the implant housing. This surface area is larger for the current R/S devices than previous generations (28). The curvature of the skull and irregularity of the surface increase the probability of exposing the dura mater.

In the retrospective review of our pediatric cohort data, 344 implants were placed with a complication rate of 8.1% (n = 29) major and 48.3% (n = 227) minor complications. The tight pocket technique was more frequently applied in the major complication group of which origin is questionably relate to the specific implant fixation technique. The most frequent major complication was infection of the implant site (20.7%, n = 6). Also no apparent difference was found in the fixation subgroups regarding device failure.

Previous studies on CI implantation in infants and small children have advised a limited bony recess due to the thin cranial bone (11, 35). To avoid risks such as dura exposure, especially in very young children, alternative fixation techniques have been introduced. In 2009 Balkany et al. (28) first reported the minimally invasive subperiosteal tight pocket technique. Variations of this technique have since been applied in pediatric and adult cohorts reporting a low major complication rate of 0%–5.2% (11, 35–37). Jethanamest et al. (38) reported no device migration or any complications related to device migration using the subperiosteal tight pocket technique. Some surgeons prefer to create a shallow well to fixate the implant (11, 35).Our clinical data on the complication rates of the different fixation techniques were inconclusive. Although the tight pocket technique was used more frequently is the major complications subgroup than the general cohort, there was no apparent difference in the rate of R/S device related issues between fixation technique groups, such as R/S device migration, infection of the implant or electrode array migration or extrusion. Furthermore, all tight pocket cases in the major complications subgroup were also operated with the SMA technique. The sample size of the tight pocket subgroup was too small to perform a statistical significance analysis. A previous review on R/S device complications in adults reported no evidence of a difference for the different fixation techniques (39). To fill this knowledge gap we are doing more research on R/S device related complications by directly comparing the two fixation techniques (bony well vs. tight pocket) in a prospective, randomized controlled study design (40).

The differences found in our retrospective study regarding the cochlear approach subgroups, were noteworthy. The most frequently used technique was the MPTA technique (74.4%, n = 256) (Table 4). However, in the major complications group, the most frequently applied surgical technique was the SMA technique (58.6%, n = 17), contrary to the general cohort. These findings are in line with an older study that included part of our cohort (14). The minor complication rate has increased over the years which could be explained by the increase of children operated under 12 months age. We included 102 children (29.7%) operated under the age of 12 months, vs. 17.7% (n = 33) that were included previously. The high number of young children could also explain the high rate of minor complications in our cohort of 48.3% (n = 166), compared to the literature, that reports rates of 1.8%–16% (10, 11, 41–43). Infectious (minor) complications such as acute otitis media and mastoiditis are known to occur more frequently in children under the age of 12 months, and comprised 30.9% (n = 86) of the minor complications in our cohort (44, 45). The higher rate could also be due to a difference in classification of complications, or potential bias such as information bias or selection bias (30).

This study is also at risk of beforementioned biases due to the retrospective design. Chart reviews are often incomplete, as was the case in our study. Due to the retrospective, non-standardized design we miss data that the observer/surgeon/doctor did not report. There could be variability in identification of complications. Moreover, the majority of CI's implanted in our study were of the brand Cochlear (n = 307, 89.2%) and most were of the CI400 series or older. The older CI models have different dimensions (thicker profile) making the comparison of R/S device related complications between subgroups difficult. Limitations are also introduced by the use of an in-house designed algorithm. However the effect of these limitations are minimized thanks to the large population size, detailed follow up and clinical application of the used algorithm. It should be noted that the algorithm searched the most optimal location within a predetermined ROI, based on expert opinion which could vary depending on the CI surgeon.

Based on the results of this study we would recommend the MPTA surgical technique over the SMA technique for cochlear approach. The results concerning the fixation techniques for the R/S device were inconclusive, but there is reason to question the current practices in pediatric patients of drilling out a bony well, especially in the 0–4 years age group. There is currently no evidence of a difference of the two surgical techniques regarding R/S migration and electrode array migration in adults (39). Further research is needed to validate complication differences in light of patients experiences (46). These outcomes are investigated in our ongoing randomized controlled trial, the results of which will be published in a peer-reviewed journal (40).