CSF gushers in cochlear implantation: surgical planning and management

Abstract

Background

Cochlear implants (CIs) have made it possible to significantly improve hearing in people with profound hearing loss. Although, cochlear implants are considered a safe procedure, this minimally invasive surgery has an overall complication rate of 12.5%. With Gusher (cerebrospinal fluid outflow) considered a common intraoperative complication of cochlear implants.

Methods

In this retrospective study, clinical files of patients with severe to profound sensorineural hearing loss who had undergone cochlear implantation were retrospectively reviewed. We calculated the incidence and risk factors of gusher and management options used . Statistical analysis included non-parametric tests and multivariate ordinal logistic regression to explore predictors of CSF leak intensity.

Results

1050 patients with profound hearing loss who underwent CI, 21 of whom had an intraoperative cerebrospinal fluid (CSF) leak, i.e. 2%, with a mean age of 3.5 years, and a sex ratio of 0.62, i.e. 62% female and 38% male. 43% of patients with gusher had structural abnormalities on their CT scan. Dilatation of the vestibule and vestibular aqueduct, as well as Mondini dysplasia, were the most common anomalies in Gusher patients. Younger implantation age was the only factor associated with higher CSF leak intensity in univariate analysis, but no independent predictors were identified in multivariate analysis.

Conclusions

Advancements in surgical techniques, radiological assessments, and technological innovation have significantly reduced cochlear implant–related complications, leading to a decreased risk of cerebrospinal fluid leakage incidents.

Background

During a cochlear implantation surgery, an abundant flow of cerebrospinal fluid may occur during a cochleostomy and it is referred to as a cerebrospinal fluid (CSF) gusher and can be, either a light flow of clear fluid called ‘oozing’, or a heavy flow called ‘gushing’, ‘geyser’ or ‘gusher’ [1–3].

However, the leakage of CSF from the inner ear, known as gusher, is the result of abnormal communication between the CSF system and the perilymph of the cochlea occurring in the event of rupture or dehiscence of the membranes separating the perilymph cavity and the CSF [4, 5].

The risk of CSF leakage may increase in cases of abnormal development of the cochleo-vestibular apparatus, such as Mondini or common cavity malformations, also in cases of generalized vestibular-auricular dysplasia [6]. It is estimated that the occurrence of a CSF gusher affects around one cochlear implant patient in a 100, except in patients with inner ear malformations, in whom incidence rises significantly to 40–50% [3].

Both conservative and surgical treatments are used in the presence of a gusher. Although it can be managed by minor surgical intervention, it may lead to life-threatening complications like meningitis. Thus, the risk and management of gusher in CI should be considered [3].

Nevertheless, gusher management is essential to ensure patient safety and the success of the operation.

In this study, we aimed to review the surgical and medical management of CSF Gusher and study its association with inner ear malformations in our patients.

Materials and methods

This is a retrospective descriptive study of 1050 patients who underwent cochlear implantation for severe to profound sensorineural hearing loss between 2008 and 2024 at the ENT Department and Mohamed VI Hearing Center of Cheikh Khalifa International University Hospital and Mohamed VI International University Hospital.

Inclusion criteria

All cochlear implant patients in whom an intraoperative clear fluid leakage of cerebrospinal fluid (CSF) occurred after opening the inner ear through the round window or the cochleostomy site.

Exclusion criteria

Patients with incomplete medical files and missing data.

Preoperative evaluation

A systematic pre-operative evaluation of all patients was carried out, during which the following examinations were performed: Otological examination, objective audiological assessment (Auditory evoked potentials, Auditory Steady-State Response (ASSR)), high-resolution computed tomography (CT) of the temporal bone studying bony inner ear anatomy and cerebral Magnetic Resonance Imaging (MRI) with study of the internal auditory canal (IAC). A complete work up was performed in all cases prior to surgery to detect the various types of congenital anomalies.

Surgical technique

The surgical procedures were performed using a minimally invasive technique via a posterior approach, with a minimal incision in the retroauricular groove without extension, followed by elevation of periosteum over the mastoid bone without opening the tympanic cavity. Next, a retentive antro-atticotomy, then posterior tympanotomy under facial Nerve Integrity Monitoring (NIM) after milling of the implant bed. The round window (RW) region was then marked and the crista fenestra milled. The RW is opened at the tip and filled with hyaluronic acid to prevent air and bone dust from entering the cochlea. The cochlear implant is then inserted and objective measurements verified. Finally, two planes closure were performed.

Round window exposure classification

The St. Thomas’ Hospital (STH) classification was referenced to assess round window exposure, which may influence the surgical approach. Its inclusion supports anatomical clarity (Fig. 1). According to this classification Type I round window membrane describes the 100% exposure of the round window membrane such that it is observed in its entirety. Type II describes partial exposure and is sub-divided into Type IIa and IIb; in Type IIa, more than 50% but less than 100% of the RWM is exposed and in Type IIb, the exposure is less than 50% but more than 0%. In Type III, the RWM cannot be identified at all even after best surgical effort. Type III occurs when the RWM lies more posteriorly and therefore closer to the facial nerve, making it much more challenging to perform a round window approach [7, 8].

Fig. 1.

St Thomas' Hospital (STH) classification of approachability of the RWM. Type I full exposure of the RWM is achieved, Type IIa more than 50% of the RWM, Type IIb less than 50% of the RWM, and Type III none of the RWM is visible

CSF leakage assessment

Based on its abundance, it is considered a “Gusher” if the flow is profuse and rapid needing 15–20 min to stop only after the packing of the round window site, while an “Oozing” is a gentle flow with spontaneous resolution within 5–10 min [9, 10].

CSF leakage management

Non surgical techniques

Placing the patient in reverse Trendelenburg position to slow CSF flow, in addition to intraoperative pCO2 lowering and mannitol administration in adults only, 30° head elevation and never leaving the operating room until the leakage stops. Afterwards, the patient is admitted for two to three days with prolonged antibiotic prophylaxis.

Surgical plugging techniques

When the gusher occurred, the first step was to wait for the flow to stop, then insert the electrode into the window to further slow-down the flow of liquid. We then proceeded to plug the hole by inserting periosteum wrapped around the electrode, packing it to prevent any leakage.

Statistical analysis

Descriptive statistics were performed to summarize the study variables, with quantitative data expressed as means and qualitative data as percentages. Statistical analyses were carried out using R software (version 4.3.1), with the significance level set at p < 0.05. Associations between continuous variables (age at implantation and sex) and leak intensity were evaluated using Spearman’s rank correlation and the Wilcoxon rank-sum test. The correlation between consanguinity and leak intensity was assessed using Spearman’s rank correlation. Kruskal Wallis was performed to determine if certain malformation types are associated with a higher or lower CSF gusher intensity during cochlear implantation. Multivariate logistic regression including age at implantation, sex, consanguinity, and inner ear malformation was performed to identify independent predictors of higher leak intensity.

Results

Among the 1,050 cochlear implant cases in our series, cerebrospinal fluid (CSF) gusher was observed in 21 patients, representing an incidence of 2%. This group included 2 adults and 19 children, ranging in age from 1 to 26 years. The mean age among the pediatric patients was 3.5 years. The sex ratio was 0.62, corresponding to 62% female and 38% male.

Deafness was congenital in 86% of cases and acquired in 14%. Of the total, 18 patients were pre-lingually deaf, 2 were post-lingual, and 1 was peri-lingual. Although the proportion of pre-lingual deafness was high, the distribution of deafness types did not significantly differ from the expected characteristics of the cochlear implant population.

Consanguinity was present in 34% of patients with CSF gusher, including 5 cases of first-degree and 2 cases of second-degree consanguinity. No patients had known risk factors for deafness or relevant medical history, except for 2 individuals diagnosed with tubular acidosis (see Table 1).

Table 1.

Clinical and demographic characteristics of patients who experienced intraoperative CSF leak

Patients characteristics	N=21	Frequency (%)
Mean implantation age	3.5
Males	8	38.10%
Females	13	61.90%
Sex Ratio	0.62	2.95%
Consanguinity
1 st degree Consanguinity	5	23.81%
2nd degree Consanguinity	2	9.52%
No consanguinity	14	66.67%
Deafness type
Post lingual Deafness	2	9.52%
Prelingual Deafness	18	85.71%
Perilingual Deafness	1	4.76%
Medical History
Tubular acidosis	2	9.52%
None	19	90.48%
IEM
Vestibule dilation	1	4.76%
Mondini	7	33.33%
Canal Cochlear Dilation	1	4.76%
None	12	57.14%
Leak Intensity
Gusher Leak Intensity	5	23.81%
Oozing Leak Intensity	16	76.19%

Spearman’s correlation analysis revealed a significant association between age and leak intensity (p = 0.035), indicating that younger patients tended to exhibit more severe leaks. No significant difference in leak intensity was found between male and female patients (p = 0.97). Similarly, consanguinity was not significantly correlated with leak intensity (p = 0.19), nor was the type of inner ear malformation (IEM).

Multivariate logistic regression analysisincluding variables such as age at implantation, gender, consanguinity, and IEMdid not identify any independent predictors of CSF leak intensity, with all p-values exceeding 0.05.

Inner ear malformations were found in 43% patients. When compared to the total CI cohort assuming IEM prevalence < 10% p < 0.001 : Highly significant association. The most frequent cochlear malformation found was Mondini (Fig. 2) in 33% of cases followed by cochlear canal dilation in 11% and vestibule dilation in 11% of cases (Fig. 3) all summarized in Table 2, whereas all the other CT scans were normal. All our cases were treated using both medical and surgical approaches.

Fig. 2.

Patient A, MRI- FIESTA, vestibular dilation. Patient B, CT-Scan of the inner ear showing vestibular dilation

Fig. 3.

Coronal FIESTA MRI showing Mondini dysplasia of the right cochlea. (A) One and a half turn of coil with absence of the modiolus. (B) Dilated vestibular aqueduct.

Table 2.

Inner ear malformations associated with intraoperative (CSF) gusher in our cohort

Malformation type :	Patients	Percentage
Cochlear canal dilation	1	11,11%
Vestibule dilation	1	11,11%
Mondini	7	77,78%
Total	9	100,00%

The round window exposure in our series was of type IIa in 15 cases, type IIb in 5 cases and type III in one case (according to STH classification [8]). We successfully introduced the electrode in all the cases. through cochleostomy in 3 patients, and from the round window approach in 18 cases.

When the gusher occurred, the first step was to wait for the flow to stop, all patients were placed in reverse Trendelenburg position to slow CSF flow, in addition to intraoperative pCO2 lowering and 30° head elevation. In all cases the plugging technique consisting of the insertion of the electrode and periosteum reinforced by muscle wrapped around the electrode and packing it into the window to slow down the flow of liquid was successfully done.

No cases of delayed postoperative CSF leakage were recorded. If postoperative leakage occurs, management includes head elevation, lumbar drainage, prophylactic antibiotics, and possible surgical re-packing under sterile conditions.

All implants were successfully activated, with no need for reoperation. Only transient vertigo was observed in 2 patients (10%), resolving spontaneously within 48 h.

Auditory thresholds improved postoperatively, with notable gains in speech perception even among patients who experienced intraoperative gusher. They were evaluated using the APCEI [11] (Acceptation, Perception, Comprehension, Expression, Intelligibility) score and free-field audiometry from one month after surgery. Long-term follow-up (6–24 months) revealed stable implant performance and auditory integration in all cases, with APCEI scores ranging between 21 and 24.

Discussion

The term ‘gusher’ is generally used in the literature to describe the egress of profuse clear fluid upon making an opening into the inner ear [12]. There are two different types of CSF outflow upon opening the cochlea. A gentle flow of clear fluid is called ‘oozing’ and a profuse flow is termed ‘gusher’. Oozing is an intermittent flow of CSF in small quantities which usually stops after a few minutes, whereas gusher flow is generally more profuse and lasts longer.

The incidence of CSF gusher in cochlear implantation is reported between 1% and 5% in large series (Wootten et al., 2006 [13]; Brito et al., 2012 [14]; Ding et al., 2009 [15]; Kim et al., 2004 [16]). According to Papsin (2005) [17], reports that include minor leaks may lead to an overestimate of the number of CSF gushers and may help explain the wide range in incidence of leaks in reports [18].

In our series, out of 1050 cochlear implantations including children and adults, 21 (2%) cases had CSF leak from which 5 cases of gusher and 16 of oozing. Univariate analysis showed that younger age at the time of implantation may be associated with higher CSF leakage intensity, suggesting potential vulnerability in pediatric patients. However, no independent predictive factors were identified in the multivariate analysis.

Most of the literature addresses CSF gushers in the pediatric population with inner ear malformations (Wootten et al., 2006 [13]; Sennaroglu, 2010 [12]; Papsin, 2005 [17]). There are few reports of CSF gusher during cochlear implantation in adults (Wootten et al., 2006 [13]; Marks, 2004). Wootten et al. (2006) have reported six adults. One of them had radiographic features consistent with the X-linked progressive mixed deafness (‘gusher’) syndrome and the remaining five patients were pre-lingually deafened. Wootten did not describe any reason for CSF gusher occurrence in these five patients. Marks (2004) [18] reported CSF gusher in one adult. His patient had progressive deafness over many years of unknown origin with normal anatomy of the cochlea in its pre-operative CT scan with no mentioning of the specific reason for CSF occurrence. In our series, we had one post-lingually deafened adult who suffered from progressive deafness treated at first by hearing aids. Her pre-operative imaging was unremarkable. We do not have any explanation for the CSF gusher in her case.

The key point in the management of CSF gusher is to create a watertight seal around the entrance of cochlear implant array. Various types of materials for sealing are suggested in the literature. Achiques et al. (2010) [19] use muscle. Loundon et al. (2008) [20] use fragments of muscle with bioglue. Mylanus et al. (2004) [21] use periosteum. Hoffman et al. (1997) [22] use fascia. Wootten et al. (2006) [13] use pieces of pressed fascia or periosteum. They believe that these are easier to manipulate in this small area and has less tendency to atrophy than muscle. Papsin (2005) [17] packs the cochleostomy with four or five small pieces of temporalis fascia more tightly than usual and also considers bolstering his pack with Tisseel fibrin glue. Marks (2004) [23] uses bone-waxed silk suture material. He cuts it into ∼1 cm segments. Then, using a Rosen needle, the segments of waxed silk are carefully packed in a circular fashion around the electrode array in the cochleostomy. Sennaroglu (2010) [12] although using pieces of the temporalis muscle, has also advocated a custom-made electrode array with a ‘cork’ stopper.

We harvest a small round piece of fresh temporalis muscle, about 4–5 mm in diameter. When the electrode is inserted, first, we pack tightly the electrode array entrance (whether it is the round window or the cochleostomy) with three to four pieces of muscle then, the muscle is advanced to the level of the entrance around the electrode and is packed all around it. This completely surrounds the electrode at the level of the entrance. Despite literature concerns about muscle atrophy, we used fresh temporalis muscle reinforced with periosteum and packing techniques to ensure a robust, watertight seal. Our results showed no postoperative complications related to electrode stability.

Inner-ear malformation is considered a predisposing factor for CSF gusher. Merhy et al. [24] previously have shown that the incidence of CSF gusher increases in the presence of IEM. Hashemi et al. [25] statistically proved the presence of a correlation of IEM with the CSF gusher. In our study, gusher rate of patients who had IEM was 43% in parallel with the literature. It can be said IEMs increase the risk of gusher in the CI [3].

In our series, there were 12 patients (57%) without IEM. It has been previously reported that gusher was found in patients with bone defects at the fundus of the IAC or at cochlear basal turn. We evaluated the radiologic examinations of our 12 patients again and no cochlear basal turn or IAC defect was identified.

There is limited data regarding the incidence of gusher in IEM subtypes in the literature. It has been reported that the gusher rate was 39–45.9.9% in incomplete partition type 1 (IP1), 8–15% in IP2, 100% in IP3, 27% in cochlear hypoplasia (CH), 23–27% in Common cavity (CC), and 0–11% in Enlarged Vestibular (EVA) groups [3].But no statistical analyses were performed between the groups. Sennaroglu et al. [12] reported in their review that gusher rates were 100% in IP3, 39% in IP1, 15% in IP2 and there was not any gusher in CC and EVA groups. A study from Tokyo [26] found that there was no significant difference in the frequency of CSF gusher among the major malformation types, however, the frequency of CSF gusher was significantly larger in cases with poorly formed modiolus. and the horizontal width of vestibular aqueduct (VA) and the horizontal modiolus base were significantly wider and larger in cases with CSF gusher compared with those without it. These findings suggest that the shape of the modiolus and VA can be important factors to predict intraoperative CSF gusher. These results suggest that intracranial pressure may affect the inner ear if the modiolus does not exist and VA also may participate in pressure regulation in the cochlea [26]. In our study, Mondini dysplasia was found to be the most frequent malformation in patients with CSF leak i.e. 78% followed by vestibular dilation i.e. 11% and cochlear canal dilation i.e. 11%.

Several studies showed that children with a CSF gusher at cochlear implant surgery were more likely to have a partial electrode insertion and fewer active electrodes [27, 28]. We successfully and fully introduced the electrode in all the cases. All the electrodes were active.

Study limitations and future directions

Limitations include the retrospective design, absence of uniform imaging protocols, and lack of long-term neurodevelopmental outcomes. Future research should investigate imaging biomarkers for gusher prediction and post-CI auditory performance in gusher cases.

Conclusion

Structural abnormalities of the cochlea, as well as other patient-related factors, influence the incidence of gusher in CI recipients. Although CSF leaks are more common in the presence of cochleovestibular anomalies such as common cavity deformities and enlarged vestibular aqueducts, they may also occur in radiographically normal ears. Most CSF gushers can be managed conservatively, but CI surgeons should always be prepared for this risk and know the appropriate intraoperative strategies. Careful preoperative planning, considering the type of malformation and the cochlear structures involved, can help improve surgical outcomes. Cochlear implantation remains a reasonable recommendation for most patients with IEMs, given the potential audiometric benefits and the role of thorough perioperative evaluation in minimizing complications. However, it is important to manage family expectations, particularly for patients with more extensive deformities. Our findings, in line with previous studies, suggest that younger age at implantation may be associated with higher CSF leak intensity, although no independent predictive factor was identified in multivariate analysis.

Abbreviations

ASSR

Auditory Steady: State Response

APCEI

Acceptation, Perception, Comprehension, Expression, Intelligibility

CI

Cochlear implant

CT

Computed tomography

CSF

Cerebrospinal fluid

ENT

Ear, nose, and throat

EVA

Enlarged vestibular aqueduct

IAC

Internal auditory canal

IEM

Inner ear malformations

MRI

Magnetic Resonance Imaging

NIM

Nerve Integrity Monitoring

pCO2

Partial Pressure of Carbon Dioxide

RWM

Round window membrane

RW

Round window

STH

St. Thomas’ Hospital

VA

Vestibular aqueduct

Authors’ contributions

**Amal Hajjij** (*equal contribution*): Designed the study, drafted the manuscript, and discussed the literature review.**Hiba El Hani** (*equal contribution*): Collected clinical data, performed the analysis, conducted the literature review, and drafted the manuscript.**Abderrahim Bourial** : Collected radiological data, drafted the radiological findings, prepared figures, and radiological images.**Khadija El Bouhmadi** : Drafted the manuscript and summarized the literature review.**Sara Benchikh** : Performed data analysis and formatted the manuscript to meet the journal’s criteria.**Said Anajar** : Revised the manuscript.**Amal Rami** : Assisted in collecting radiological data and revised the manuscript.**Nadia Moussali** : Analyzed the radiological findings, interpreted the patient images, and revised the manuscript.**Khalid Snoussi** : Designed the study and revised the manuscript.

Funding

No specific funding was received for this study.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki. Formal ethics committee approval was not required, as only retrospective and anonymized patient data were analyzed. Written informed consent for surgery and data use was obtained from all participants or their legal guardians.

Consent for publication

Written consent for publication was obtained from the patients (or their parents/guardians) as required, and their personal information has been kept confidential in accordance with ethical guidelines.

Competing interests

The authors declare no competing interests.