Facial nerve stimulation in adult cochlear implant recipients with far advanced otosclerosis

Abstract

Objectives

The objective of this study was to predict occurrence of facial nerve stimulation (FNS) in cochlear implanted patients for far‐advanced otosclerosis (FAO) by correlating preoperative computed tomography (CT)‐scan data to FNS and to evaluate FNS impact on hearing outcomes.

Methods

Retrospective analysis on 91 ears (76 patients) implanted for FAO. Electrodes were straight (50%) or perimodiolar (50%). Demographic data, extension of otosclerosis on preoperative CT scan, occurrence of FNS, and speech performance were analyzed.

Results

Prevalence of FNS was 21% (19 ears). FNS appeared during the first month (21%), 1–6 months (26%), 6–12 months (21%), and over 1 year (32%) postimplantation. Cumulative incidence of FNS at 15 years was 33% (95% CI = [14–47%]). Extension of otosclerotic lesions on preimplantation CT‐scan was more severe in FNS ears compared to No‐FNS (p < .05): for Stage III, 13/19 (68%) and 18/72 (25%) ears for FNS and No‐FNS groups, respectively (p < .05). Location of otosclerotic lesions relative to the facial nerve canal was similar whatever the presence or not of FNS. Electrode array had no impact on FNS occurrence. At 1 year post‐implantation, duration of profound hearing loss (≥5 years) and previous stapedotomy were negatively associated with speech performance. FNS did not impact hearing outcomes, despite a lower percentage of activated electrodes (p < .01) in the FNS group. Nevertheless, FNS were associated with a decrease of speech performance both in quiet (p < .001) and in noise (p < .05).

Conclusion

Cochlear implanted patients for FAO are at greater risk of developing FNS affecting speech performance over time, probably due to a higher percentage of deactivated electrodes. High resolution CT‐scan is an essential tool allowing FNS prediction but not time of onset.

Level of evidence

2b, Laryngoscope Investigative Otolaryngology, 2022.

Prognostic factors for facial nerve stimulations on cochlear implanted patients for far advanced otosclerosis were studied. Radiological analysis using our classification allows prediction of stimulations. Short‐term hearing outcomes were not modified but long‐term hearing outcomes showed a tendency to decrease.

LAY SUMMARY

Otosclerosis is a common cause of conductive hearing loss. Patients have far‐advanced otosclerosis when appropriate hearing aids are not enough for appropriate speech discrimination. Cochlear implantation is a treatment option, but aberrant facial nerve stimulations may occur.

1. INTRODUCTION

Clinical otosclerosis is a relative common cause of hearing loss, particularly among 20 to 50 years old Caucasians, with a prevalence ranging from 0.1% to 2.1%. ¹ , ² Nevertheless, temporal bone studies show that histological prevalence is much higher, reaching 12%. ³ Sensorineural disease is described in 10% of the otosclerotic population and can lead to severe to profound hearing loss, but a purely sensorineural disease is rare. ⁴ Far‐advanced otosclerosis (FAO) was initially defined by specific audiometric criteria ⁵ : air conduction threshold above 85 dB, and unmeasurable bone conduction threshold. In the era of cochlear implantation, the diagnosis is made in case of poor speech discrimination with appropriate hearing aids. ⁵ , ⁶

To improve FAO patient intelligibility, two options are usually offered to patients: stapedotomy and/or cochlear implantation. Currently, there are no specific guidelines on the management of severe to profound hearing loss in unoperated patients. Stapedotomy is a simple and cost‐effective intervention to reduce or close the air‐bone gap, get a measurable air conduction threshold, and allow the use of hearing aids. In FAO with non‐measurable bone conduction thresholds, the benefit is variable and hard to predict. ⁷ A low risk of postoperative labyrinthitis exists, resulting in complete hearing loss, and sometimes rapid fibrosis of scala tympani with ossification, which jeopardizes future electrode insertion. Cochlear implantation is a more expensive and complex procedure with heavier patient care. ⁶ , ⁸ , ⁹ , ¹⁰ , ¹¹ Two meta‐analysis comparing hearing outcomes between primary cochlear implantation and stapes surgery conclude that cochlear implantation leads to a greater and consistent improvement in speech recognition scores, lower complication, and higher satisfaction rates. The choice between these two strategies takes into account the type, severity, and evolution of hearing loss in both ears, patient desires, cochlear implant cost, and surgical risks. ⁷ , ⁸ , ¹²

Cochlear implant recipients with FAO are more likely to experience facial nerve stimulation (FNS), ¹³ , ¹⁴ usually explained by a reduction of impedance and higher conductivity due to spongiotic bone properties. Incidence of FNS in otosclerotic patients widely varies between studies, ranging from 0% to 57%. ⁹ , ¹³ , ¹⁵ Patients experiencing FNS may require repeated settings adjustments or electrode deactivation, which can lead to poor performance. ¹¹ , ¹² Whereas several classifications for otosclerosis on computed tomography (CT)‐scans exist, ⁸ , ¹⁶ , ¹⁷ , ¹⁸ no studies focusing on the radiologic analysis of otosclerotic foci locations relative to the facial nerve have been published.

The objective of our study was, in cochlear implanted patients for FAO, to evaluate prognostic factors of FNS occurrence, especially the extension and location of otosclerotic lesions, and to assess the impact of FNS on short‐ and long‐term hearing outcomes. Indeed, a better prediction of FNS occurrence on preoperative CT scan and the knowledge of its influence on patient's performance could help in counseling patients.

2. MATERIALS AND METHODS

2.1. Subjects

A retrospective study was conducted in a tertiary referral center, in a cohort of adult patients implanted between 1991 and 2017 for FAO, that met the selection criteria for cochlear implantation recommended for adults (https://www.has-sante.fr): bilateral severe to profound sensorineural hearing loss with speech recognition scores ≤50% for open‐set disyllabic words presented at 60 dB sound pressure level (SPL) in quiet, in the best‐aided conditions after verification of optimal hearing aid fittings.

Patients were excluded if they presented:

• No preoperative CT‐scan.

• Other cause that could explain profound hearing loss.

• Follow‐up less than 1‐year post‐cochlear implantation.

• Poor fluency in French language affecting the speech recognition test results.

Patients were implanted with one of the four available cochlear‐implant brands (Advanced Bionics, Stäfa, Switzerland; Cochlear, Lane Cove, Australia; Med‐El, Innsbruck, Austria; Oticon Medical, Vallauris, France) with perimodiolar (Cochlear Contour Advance®; Advanced Bionics HiFocus MidScala®), and straight electrodes (all brands). The choice of implant brand and electrode design were not guided by extension of otosclerosis on preoperative CT‐scan but decided on a case‐by‐case basis. All patients provided written informed consent allowing analysis of their data and the study was approved by the National Commission for Data Protection and Liberties (CNIL‐France, approval no. 2040853). The study complies with the Declaration of Helsinki.

2.2. Data collection

Demographic data including sex and age at implantation, hearing loss onset, duration of profound hearing loss before implantation, and previous stapes surgery were collected. Implant data were collected for each patient and number of active electrodes at each control visit, and FNS occurrence at 1 month, 6 months, and yearly after implantation. The statistical unit of the study was the implanted ear, and in case of bilateral implantation, both ears of the same patient were included in the study.

2.3. Auditory evaluation

Pure‐tone audiometric thresholds were measured before implantation using headphones in a soundproof booth. Mean air and bone‐conduction thresholds were obtained by averaging 250, 500, 1000, 2000, and 4000 Hz frequency scores. Speech performance was evaluated yearly before and postimplantation using monosyllabic words (Lafon lists) in quiet with a 60 dB SPL front signal, and in noise using MBAA (Marginal Benefit from Acoustic Amplification) calibrated sentences at 60 dB SPL with a Speech‐to‐Noise‐Ratio (SNR) of 10 dB front signal.

2.4. Radiological classification

Presence of otosclerotic lesions was confirmed on preoperative temporal bone CT. We established a three‐stage classification of the otosclerotic extent adapted from previous publications ⁸ , ¹⁶ (Table 1, Figure 1). The location of otosclerotic foci relative to the facial nerve (four grades), in the region of the oval window was defined in four grades (Table 1, Figure 1).

TABLE 1.

Radiological extension of otosclerosis on CT scan

Extension of otosclerotic lesions to the otic capsule
Stage I	Disease limited to both oval and round windows	28 (31%)
Stage II	Otic capsule lesions without endosteum involvement	32 (35%)
Stage III	Otic capsule lesions with endosteum involvement and loss of cochlear architecture	31 (34%)
Location of otosclerotic foci relative to the facial nerve
Grade a	No otosclerotic foci near the facial nerve	22 (24%)
Grade b	Disease reaching the 2nd portion of the facial nerve, posterior to the oval window region	16 (18%)
Grade c	Disease reaching the 1st and/or 2nd portion of the facial nerve, anterior to and at the level of the oval window region	42 (46%)
Grade d	Disease reaching all portions of the facial nerve	11 (12%)

Note: Values are presented as n (%).

Abbreviation: CT, computed tomography.

FIGURE 1.

Radiological classifications of otosclerosis on preoperative CT‐scan. (A) Stage I: unique foci on the fissula ante fenestram (black arrow). (B) Stage II: pericochlear foci (black arrow) without endosteum involvement. (C) Stage III: endosteum involvement with loss of cochlear architecture. (D) Grade c: otosclerotic foci relative to the 1st and 2nd portion of the facial nerve. (E) Grade d: otosclerotic foci relative to the three portions of the facial nerve (arrow: 3rd portion of the nerve). CT, computed tomography

2.5. Statistical analysis

Statistical analyses were computed using R software 4.1.1 (R Development Core Team, 2011). Quantitative variables were expressed as mean ± standard deviation or median [minimum–maximum] and compared using Student‐test or nonparametric Wilcoxon‐rank sum test. Qualitative variables were expressed in % and compared using Chi‐square test or Fisher's exact test, as appropriate. For overall description purpose, cumulative incidence (CI) of FNS and its 95% CI was estimated using the Kaplan–Meier approach; this analysis did not consider correlation of ears by patient. Linear regression models were used to study determinants of hearing outcomes at 1‐year postimplantation. Determinants of the evolution of hearing outcomes over time were explored using linear mixed models with random intercept and slope, and patient‐implant effects. All tests were two‐sided and used a level of significance of 5%.

3. RESULTS

3.1. Population

Among 2270 adult patients undergoing cochlear implantation between 1991 and 2017, 100 patients (4.4%) presented FAO. Considering the inclusion criteria, 24 patients were excluded. Seventy‐six patients were included, 61 (80%) and 15 (20%) were unilaterally and bilaterally implanted, respectively, with a perimodiolar electrode in 46 ears and a straight electrode in 45 ears. Demographic and implant data are presented in Table 2. The median follow‐up duration was 6 [4–10] years. Two patients were reimplanted on one side, one, 1‐year postimplantation due to device failure, and the other 10‐years postimplantation because of electrode array extrusion (cholesterol granuloma). The same type of electrode (one perimodiolar and one straight) was reintroduced in both cases without surgical difficulty. Speech performance remained stable after reimplantation.

TABLE 2.

Population characteristics (91 ears/76 patients)

Gender: female/male	45 (59%)/31 (41%)
Age at cochlear implantation (years)	61 ± 10.7, 61 (91) [21–87]
Duration of profound hearing loss (years)	6 ± 7.0, 2.5 (91) [0–47]
Cases with previous stapedotomy	54 (59%)
Post‐operative hearing loss	13 (25%)
Sudden hearing loss	2 (4%)
Progressive hearing loss	39 (71%)
Cases without previous stapedotomy	37 (41%)
Sudden hearing loss	2 (5%)
Progressive hearing loss	35 (95%)
Preimplantation pure‐tone average a
Air conduction thresholds (dB HL)
Group with previous stapedotomy
Missing data	1 (2%)
Not measurable	11 (22%)
Measurable	42 (76%)
Thresholds (dB HL)	101 ± 12, 104 (42) [75–115]
Group without previous stapedotomy
Missing data	1 (3%)
Not measurable	5 (13%)
Measurable	31 (84%)
Thresholds (dB HL)	104 ± 13, 109 (31) [70–115]
Bone conduction thresholds (dB HL)
Group with previous stapedotomy
Missing data	1 (2%)
Not measurable	29 (54%)
Measurable	24 (44%)
Thresholds (dB HL)	79 ± 18, 72 (24) [60–115]
Group without previous stapedotomy
Missing data	2 (5%)
Not measurable	19 (51%)
Measurable	16 (44%)
Thresholds (dB HL)	74 ± 16, 67 (16) [50–100]
Cochlear implantation
Unilateral	77 (85%)
Bilateral sequential	10 (11%)
Bilateral simultaneous	4 (4%)
Implant brand b
Advanced Bionics	16 (17%)
Cochlear	46 (50%)
Med‐El	10 (12%)
Oticon Medical	19 (21%)
Electrode array
Straight	45 (50%)
Perimodiolar	46 (50%)
Follow‐up duration [Q1–Q3] (years)	6 [4–10] (91)
Duration follow‐up >3 years	72 (79%)

Note: Values are mean ± SD, median (n) [min–max] or n (%).

^a The pure‐tone average was calculated using the thresholds at 0.25, 0.5, 1, 2, and 4 kHz for both ears.

^b Advanced Bionics (Stäfa, Switzerland), Cochlear (Lane Cove, Australia), Med‐El (Innsbruck, Austria), and Oticon Medical (Vallauris, France).

3.2. Preimplantation CT‐scan assessment and analysis

The round window was totally, partially, and not ossified in 25 (28%), 15 (16%), and 51 (56%) cases, respectively. For the extension of otosclerotic lesions to the otic capsule (Table 1), each stage represented about 1/3 of the patients. Considering the location of otosclerotic foci relative to the facial nerve, only 24% of ears did not present any foci near the nerve (Table 1). The distribution of these three radiological parameters taken together is reported in Figure 2. It could be noticed that in stage II and III, round window ossification was not observed in 65% and 10% of cases respectively (Figure 2).

FIGURE 2.

Otosclerosis extension stages relative to the round window ossification (A) and to foci around the facial nerve (B)

3.3. Occurrence of FNS

FNS occurred in 19 ears (21%) in median after 5 months (Q1–Q3 = [1.6–67]) postimplantation. In 13 ears (68%), FNS appeared during the first year after cochlear implantation (Early‐FNS): the first month in four cases (21%), between 1 and 6 months in five cases (26%) and between 6 and 12 months in four cases (21%). In six ears (32%), FNS occurred between 19 and 309 months after cochlear implantation (Late‐FNS). FNS occurred on both ears in 3 out of 15 bilaterally implanted patients (20%), and in one ear in 2 patients (13%). For the two patients who underwent reimplantation, FNS was observed before reimplantation in one of them, and no FNS occurred after reimplantation. Cumulative incidence of FNS, 15 years after implantation, was 33% (95% CI = [14–47]) (Figure 3).

FIGURE 3.

Cumulative incidence of facial nerve stimulation (Kaplan–Meier analysis)

Age at implantation, previous stapes surgery, duration of profound hearing loss, preimplantation air, and bone‐conduction thresholds were similar between FNS and No‐FNS groups. The type of electrode array, straight or perimodiolar, was also similar between the two groups (data not shown).

3.4. Relationship between preimplantation radiological data and FNS occurrence

Considering the analysis of preimplantation CT‐scans and the occurrence of FNS (Table 3), ossification of the round window was more frequently observed in the FNS group (14/19, 74%) than in the No‐FNS group (27/72, 38%) (p < .001). Extension of otosclerotic lesions on preimplantation CT‐scan was more severe in FNS ears compared to No‐FNS (p < .05): for Stage III, 13/19 ears (68%) and 18/72 (25%) for FNS and No‐FNS groups, respectively (p < .05). Surprisingly, location of otosclerotic lesions relative to the facial nerve canal was similar whatever the presence or not of FNS. No difference was observed between Early‐ and Late‐FNS. Nevertheless, it should be noticed that all Late‐FNS (n = 6) presented extended otosclerotic lesions (stage II and III) with grade c and b facial nerve location.

TABLE 3.

Radiologic analysis on CT scan of otosclerotic lesions before cochlear implantation in patients without, with early (1‐year), and with late facial nerve stimulation after cochlear implantation

n = 91	No‐FNS, 72 (79%)	Early‐FNS, 13 (14%)	Late‐FNS, 6 (7%)
Round window ossification
Absent	45 (63%)	3 (23%)	2 (33%)
Partial	9 (12%)	6 (46%)	0
Complete	18 (25%)	4 (31%)	4 (67%)
Extension of otosclerotic lesions
Stage I	27 (37.5%)	1 (8%)	0
Stage II	27 (37.5%)	3 (23%)	2 (33%)
Stage III	18 (25%)	9 (69%)	4 (67%)
Location of otosclerotic foci relative to the facial nerve
Grade a	19 (26%)	2 (15%)	0
Grade b	15 (21%)	1 (8%)	0
Grade c	28 (39%)	9 (69%)	5 (83%)
Grade d	10 (14%)	1 (8%)	1 (17%)

Note: Values are presented as n (%). Percentages were calculated reported to the total number of patients in the group No‐, Early‐, and Late‐FNS.

Abbreviations: CT, computed tomography; FNS, facial nerve stimulation.

3.5. FNS and fitting strategies/electrode deactivations

FNS can be resolved by changing the fitting strategy and by deactivating the responsible electrodes. At the last available fitting session, 17 out of the 19 FNS ears (89%) had deactivated electrodes versus 28/72 (39%) of No‐FNS ears (p < .0001). The mean percentage of deactivated electrodes was higher in the FNS group (23 ± 17.4%) compared to No‐FNS (5 ± 7.7%, p < .0001).

To assess the location of deactivated electrodes, each array was divided in thirds (apical/middle/basal). In FNS ears, 44 ± 34.6% of deactivated electrodes were basal, 40 ± 31.2% middle (closest to the facial nerve), and 16 ± 23.4% apical compared to 90 ± 47.0%, 10 ± 17.1% and 0%, for No‐FNS ears. The repartition of deactivated electrodes varied between the two groups. It should be noted that no patient was explanted because of FNS.

3.6. Prognostic factors of hearing outcomes

Speech performance at 12 ± 3 months postimplantation were available for 64/91 ears; eight of them (12.5%) had FNS (Table 4). No difference in speech performance scores in quiet and noisy environment was observed between Early‐FNS and No‐FNS groups, although the mean percentage of active electrodes was lower in patients with FNS (81 ± 18% vs. 94 ± 8% for No‐FNS, p < .01). In cases with profound hearing loss duration over 5 years, poorer performance in quiet was observed compared to shorter duration (43 ± 24% vs. 59 ± 27%, p < .05). The existence of previous stapedotomy also affected performance: in quiet (46 ± 27% vs. 62 ± 24%, p < .05) and noisy environment (41 ± 30% vs. 61 ± 28%, p < .05). No difference in patient's age or duration of profound hearing loss was observed between patients having or not stapedotomy. Measurable air or bone‐conduction thresholds, age at implantation, hearing loss onset, and preimplantation CT‐scan lesion analysis, had no impact on hearing outcomes. Multivariate analysis confirmed these two prognostic factors.

TABLE 4.

Hearing outcomes at 1‐year postimplantation relative to potential prognostic factors

Facial nerve stimulation	No‐FNS	FNS
Monosyllabic words in quiet (%)	51 ± 27.4 (53) [6–100]	60 ± 13.9 (63) [37–74]
Sentences in noise (%)	48 ± 31.1 (53) [0–100]	55 ± 28.0 (63.5) [14–93]
Duration of profound hearing loss	≤5 years	>5 years
Monosyllabic words in quiet (%)	59 ± 26.5 (56) [6–100]	43 ± 23.8 (35) [12–100]*
Sentences in noise (%)	53 ± 28.9 (53) [0–100]	41 ± 32.1 (40) [0–93]
Previous stapedotomy	No	Yes
Monosyllabic words in quiet (%)	61 ± 23.3 (63) [15–100]	46 ± 26.8 (44) [6–100]*
Sentences in noise (%)	61 ± 27.7 (60) [0–100]	41 ± 30.3 (40) [0–93]*

Note: Values are presented as mean ± SD (median) [min–max]; n = 64.

Abbreviation: FNS, facial nerve stimulation.

^* p < .05.

Regarding the evolution of speech performance over time (mixed linear models), FNS was associated with a decrease of speech performance both in quiet (p < .001) and in noisy environments (p < .05). A higher percentage of active electrodes was positively associated with better speech performance over time in noisy environments (p = .015). Previous stapedotomy was also associated with better speech performance in noise over time (p < .05). No other factor was found to have an influence on hearing outcomes over time.

4. DISCUSSION

4.1. Occurrence of FNS

FNS is the most frequent complication reported in cochlear implant recipients with FAO. In our study, which included a large number of patients with a longer mean follow‐up compared to literature, the rate of FNS was 21%, with a CI of 33% at 15 years. In Table 5, we summarized 22 articles reporting FNS after cochlear implantation for a FAO, published between 2000 and 2022. FNS incidence varied highly among studies, ranging from 0% incidence ¹⁹ , ²⁰ up to 57%. ¹³ , ¹⁴ The high heterogeneity among studies regarding inclusion criteria, radiological extension of otosclerosis, number of patients, and cochlear implant technology may explain this variability. ²¹ , ²² In a meta‐analysis on factors influencing FNS after cochlear implantation, FNS incidence was higher in the group with otosclerosis (28%) compared to patients with other etiologies (3.5%), with an odds ratio of 13.73 predicting FNS in this population. ¹³

TABLE 5.

Summary of literature related to FNS incidence in cochlear implant recipients with far‐advanced otosclerosis

Authors	Study design	Study date	Implanted ears (n)	FNS incidence (%)	Onset of FNS
Broomfield et al., 2000 37	Retrospective Case–control	1988–1998	12	50	NR
Rayner et al., 2003 15	Retrospective Case–control	1986–2001	14	57	NR
Rotteveel et al., 2004 16	Retrospective	1990–2002	53	38	NR
Marshall et al., 2005 26	Retrospective	NR	30	17	<12 months in all cases
Quaranta et al., 2005 38	Retrospective Case–control	1995–2003	9	33	NR
Rama‐López et al., 2006 39	Retrospective	NR	30	0	‐
Matterson et al., 2007 32	Retrospective	1986–2004	59	23	NR
Mosnier et al., 2007 33	Retrospective	1991–2003	16	6	NR
Psillas et al., 2007 40	Retrospective Case–control	1997–2007	5	20	NR
Sainz et al., 2009 41	Prospective	NR	15	13	At 2 years in 1 case At 5 years in 1 case
Berrettini et al., 2011 42	Retrospective	1999–2007	6	33	Immediate in 1 case Delayed in 1 case.
Semaan et al., 2012 30	Retrospective Case–control	2003–2011	34	0	‐
Seyyedi et al., 2013 31	Retrospective	NR	13	31	NR
Castillo et al., 2014 43	Retrospective	NR	17	12	NR
Kabbara et al., 2015 8	Retrospective	1993–2013	34	17	NR
Vashishth et al., 2017 20	Retrospective	NR	38	5	NR
Calvino et al., 2018 24	Retrospective Case–control	1992–2017	22	18	NR
Dumas et al., 2018 9	Retrospective Case–control	NR	35	9	NR
Atanasova‐Koch & Issing, 2021 44	Retrospective Case–control	2004–2020	22	14	NR
Van Horn et al., 2020 13	Systematic review and meta‐analysis		74	28	NR
Assiri et al., 2021 45	Systematic review		351	11	NR
Kondo et al., 2022 14	Systematic review and meta‐analysis		474	18	NR

Note: Only studies published between 2000 and 2021were included. Case–control: comparison of FNS incidence between cochlear implant recipients with and without otosclerosis.

Abbreviations: FNS, facial nerve stimulation; NR, not reported.

Several mechanisms have been proposed to explain the higher proportion of FNS in patients with otosclerosis. A reduction in bony resistance due to spongiotic bone, resulting in electrical current shunts between the electrode and the facial nerve, especially for mid‐array electrodes, could be the main explanation. ¹⁶ , ²³ , ²⁴ , ²⁵ A computer modeling study confirmed increased bone conductivity in otosclerosis and showed increased excitation levels of the auditory nerve. This could explain that, in some regions of the cochlea, facial nerve excitation thresholds were lower than those of auditory nerve, ²² leading to aberrant stimulations. These theories suggest that analysis not only of the extension of the disease, but also of the location of otosclerotic foci relative to the facial nerve canal on preoperative CT‐scans, could predict FNS occurrence. Several classifications rate otosclerosis extension disease in patients. ⁸ , ¹⁶ We created a modified version to include description of otosclerotic foci around the facial nerve canal. However, if extension of the otosclerotic foci was associated with a higher incidence of FNS, there was no impact of the location of otosclerotic foci relative to the facial nerve in our cohort. This confirms that FNS occurrence could be predicted radiologically on preoperative CT‐scans based only on extension of the disease as suggested in previous studies. ⁸ , ¹⁶

In the literature (Table 5), the onset of FNS is very rarely mentioned. In our study, FNS occurred in 75% of cases during the first‐year postimplantation, a value slightly lower than that reported in literature (92% to 100%) in patients with various etiologies. ²⁶ , ²⁷ In our study, in six patients, FNS occurred later. In the few studies describing Late‐FNS, it was suggested that progression of otosclerosis with changes of tissue impedance and/or erosion of the thin bony layer between the scala tympani and the facial nerve was greater than in patients without otosclerosis. ²⁷ , ²⁸ Surprisingly, in our study, all the six patients with Late‐FNS already had extensive otosclerotic lesions on preoperative CT‐scan. A progression of the extension of the disease over time cannot be ruled out but is difficult to evaluate due to artifacts caused by the electrode array.

4.2. Deactivated electrodes and hearing outcomes

To manage stimulations, different strategies exist such as changing the programming strategy, reduction of stimulation levels, or selective deactivation of affected electrodes. ²⁰ , ²³ , ²⁷ In refractory cases, explantation with or without reimplantation has been reported in very rare cases. ²⁹ Changes in program strategy could not be analyzed in our cohort due to the retrospective design of the study, the long follow‐up period, and differences in implants/speech processor generations. As described in previous studies, the number of active electrodes was significantly lower in FNS ears. ¹⁹ , ²⁹ , ³⁰ , ³¹ Deactivated electrodes in this group were more often located at the middle of the electrode carrier, corresponding to the upper basal turn of the cochlea, close to the facial nerve.

Considering the design of the electrode array, some studies report that using perimodiolar electrodes, further from the facial nerve, could limit facial stimulations compared to straight electrode arrays. ¹³ , ²⁰ , ³² , ³³ , ³⁴ A recent meta‐analysis found a 3.92 odds ratio predicting FNS in patients implanted with straight arrays in cochlear implant recipients of various etiologies. ¹³ In our cohort, no significant difference in FNS was observed between straight (68%) and perimodiolar (44%) electrode arrays.

In this series, speech performance in quiet at 1‐year postimplantation was in accordance with that reported in a national register including 3176 adult patients with various etiologies. ¹¹ These scores remained stable long term with an average follow‐up of 14 years. Even if the percentage of deactivated electrodes was higher in the FNS group, this did not negatively affect speech performance. At 1‐year, only two factors contributed to performance variability: the duration of profound hearing loss >5 years, a parameter that is a well‐known factor affecting speech performance, ³⁵ and previous stapes surgery. Whereas the presence of FNS at 1‐year was not a prognostic factor, long‐term analysis of the entire cohort showed a negative effect of FNS on speech performance in both quiet and noise, possibly due to a higher percentage of deactivated electrodes. Regarding the effect of FNS on speech performance, conflicting results are found in literature ⁹ , ¹⁴ , ¹⁶ , ²⁴ , ²⁵ , ³⁶ in patients with FNS. Considering previous stapes surgery, some studies ⁷ , ⁸ , ²⁶ did not find negative impact of previous stapedotomy on speech performance whereas others observed poorer performance, as in our study. ⁹

4.3. Strengths and limitations

This study concerned the largest cohort of patients with FAO to date, with a long follow‐up duration compared to previous published studies. However, the patient population was heterogenous in disease presentation, design of electrode, brand, and generation of cochlear implants, which introduced bias. The retrospective study design may underestimate late FNS due to loss of follow‐up and a small rate of FNS.

In our population, otosclerotic foci were limited to the windows on CT scans in 1/3 of cases, suggesting no endosteum damage. Although patients with other identified etiologies explaining the profound hearing loss were excluded from the study, another cause of deafness cannot be ruled out, and the otosclerotic foci could be an incidental finding on the preoperative CT scan. Indeed, subclinical otosclerotic foci have been described in 12% of temporal bone analysis. ³

5. CONCLUSION

To conclude in FAO, the radiologic extent of otosclerosis foci before cochlear implantation was a pertinent prognostic FNS occurrence but did not predict time of onset. Over time, FNS had a negative impact on speech performance in both quiet and noise, probably due to deactivated electrodes. A precise evaluation of the extension of the pathology can therefore allow better counseling and monitoring of patients.

FUNDING INFORMATION

This research received no external funding.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

Tuset M‐P, Baptiste A, Cyna Gorse F, et al. Facial nerve stimulation in adult cochlear implant recipients with far advanced otosclerosis. Laryngoscope Investigative Otolaryngology. 2023;8(1):220‐229. doi: 10.1002/lio2.984