Long-Term Outcomes of Pediatric Cochlear Implantation for Single-Sided Deafness: A Western Australian Perspective

Abstract

Objectives:

To assess the outcomes of cochlear implantation (CI) for the management of pediatric single-sided deafness (SSD) in an Australian tertiary pediatric center.

Design:

We performed a retrospective review of data from the Western Australian childhood hearing implant program between 2014 and 2023. Patients with SSD age below 16 yr and undergoing unilateral CI at Perth Children’s Hospital were included. Data collected included demographics, history, pre-CI assessment, language ability, Bamford-Kowal-Bench (BKB)-Sentence in Noise (SIN) results, CI usage hours, localization, and Speech, Spatial and Qualities of Hearing Scale scores. Results were compared with and without CI where available.

Results:

There were 22 patients who underwent CI for SSD in the 10-yr period. A total of 54.5% were male and 45.5% were female. The average age at diagnosis was 4.2 yr (0.0 to 15.1, SD: 4.8). Six patients (27.3%) were classified as late-onset SSD and 16 (72.7%) as early SSD. The average time to implantation was 1.3 yr (0.1 to 3.9 yr, SD: 1.3). Eleven patients (50%) were classified as early CI and 11 patients as late CI. The average age at implantation was 5.5 yr (1.0 to 15.6 yr, SD: 4.7). The etiology of the SSD was unknown in 8 patients (36.7%). Three of our patients showed improvement in language ability 1 yr after implantation. No patients scored poorer compared with their pre-CI language assessment. Six of the 22 subjects (27.27%) underwent localization testing. There was no significant difference identified between CI and without CI for localization (20% versus 24%; p = 0.93). Thirteen of 22 (59.09%) completed BKB-speech-in-noise (SIN) testing. Five patients performed better in their BKB-SIN with their CI and 12 performed worse with their CI. Of these 5 patients, 3 were in the early-onset SSD group and 2 in the late-onset SSD group. Of the same 5 patients, 4 were classified as early CI and 1 was classified as late CI. The majority of patients who benefited with BKB-SIN (80%) and localization testing (75%) were classified as early CI, that is, within 1 yr of onset. The average SNR loss in the early SSD group was 2.91 (1.3 to 6) with CI and 2.84 (0.5 to 5.5) without CI (p = 0.462). For late-onset SSD, the average with CI was 3.72 (0.01 to 6) and 3.78 (0.1 to 7.7) without CI (p = 0.475). The overall average for all patients, with and without CI, was 3.28 versus 3.28. Eight patients (34.78%) completed the Speech, Spatial and Qualities of Hearing Scale. The average CI usage was 4 hr/d (0 to 12.9, SD: 3.7)

Conclusions:

Despite a well-established newborn hearing screening program and equitable access to specialized services, our patients have variable long-term outcomes. While several patients showed benefit in speech in noise, high usage rates, and improved language skills, the challenge remains in consistently predicting and rehabilitating a heterogenous population of pediatric SSD patients. Patients who were classified as early CI recipients performed better, regardless of the onset of SSD.

INTRODUCTION

Unilateral hearing loss (UHL) is distinct from bilateral hearing loss in its manifestation, impact, causes, and management. Historically, UHL was considered less significant, as patients could rely on their better-hearing ear. However, in the 1980s, Bess & Tharpe (1984) published a pivotal study revealing the considerable detriment UHL has on a child’s speech and language development. Landmark paper redirected attention to the rehabilitation of UHL, highlighting the importance of treating each ear individually (Bess & Tharpe 1984). Since research has increasingly highlighted the significant impact of UHL on speech and language development, balance, and social skills (Fitzpatrick et al. 2019; Benchetrit et al. 2021; Brown et al. 2022; Malhotra et al. 2022; Park et al. 2023a).

Congenital UHL has an incidence of 0.6 to 0.7 per 1000 live births (Hunter et al. 2022). In addition, a large proportion of pediatric UHL is acquired during early childhood (Benchetrit et al. 2021; Hunter et al. 2022). When this cohort of patients is included, the prevalence of UHL increases to 2.5 to 6% by the time the child reaches school age (Shargorodsky et al. 2010). The severity of UHL ranges from mild to profound; the latter is known more commonly as single-sided deafness (SSD). Untreated SSD can lead to cortical reorganization in favor of the better-hearing ear, and increase the risk of psychosocial, behavioral, cognitive, and functional issues compared with their peers with normal hearing (Arndt et al. 2015; Fitzpatrick et al. 2019; Benchetrit et al. 2021; Gordon et al. 2021; Selleck et al. 2021; Brown et al. 2022).

Children with SSD typically do not benefit from a conventional hearing aid due to the degree of hearing loss and poor speech discrimination levels (Bagatto et al. 2019; Selleck et al. 2021; Brown et al. 2022). Several alternative hearing rehabilitation options exist, including a contralateral routing of signal (CROS) hearing aid, bone conduction hearing aids, bone conduction implants, and cochlear implants (CI). CROS and bone conduction devices have inherent limitations, particularly in the pediatric setting (Hunter et al. 2022; Chandrasekar et al. 2023). CROS devices are associated with frequent reports of dissatisfaction and discomfort. Similarly, bone conduction devices are often attached to a soft band or tightly fitted ear mold. In many countries, children under the age of five are ineligible for implants and therefore rely on watchful waiting or nonsurgical devices until the age of at least seven or eight (Bagatto et al. 2019). These devices provide limited patient satisfaction and are unlikely to improve localization, as they do not facilitate binaural hearing (Bishop & Eby 2010; Bagatto et al. 2019; Brown et al. 2022; Hunter et al. 2022; Chandrasekar et al. 2023).

In adults with SSD, CI has demonstrated superiority over CROS and bone conduction devices for hearing rehabilitation (Mertens et al. 2016; Buss et al. 2018; Sullivan et al. 2020; Peters et al. 2021). Unlike other devices which provide additional input to the contralateral ear, CI directly stimulates the auditory nerve in the deafened cochlea, offering the potential to restore true binaural hearing. Binaural hearing confers several critical advantages, including summation, squelch, binaural redundancy, and head shadow effects, which contribute to sound localization as well as improved speech understanding in both quiet and noisy environments. (Arndt et al. 2015; Benchetrit et al. 2021; Peters et al. 2021; Brown et al. 2022; Kubina et al. 2023; Park et al. 2023a). Early implantation in prelingual children with SSD is associated improved spatial awareness and speech and language outcomes (Arras et al. 2021; Malhotra et al. 2022). Timely intervention, during a period of high neuroplasticity, may prevent irreversible developmental deficits in the normal auditory pathways and reduce the risk of “aural preference syndrome” where processing shifts towards the normal-hearing ear (Gordon et al. 2013, 2021; Vanderauwera et al. 2020; Park et al. 2023a).

CI was only recently approved by the United States Food and Drug Administration for children with SSD aged 5 and older in 2019 (Benchetrit et al. 2021; Park et al. 2022). Australia, with a socialized healthcare model and universal newborn hearing screening, is a leader in access to pediatric CI (Sorkin & Buchman 2016). Careful assessment of candidacy and prudent resource allocation therefore becomes paramount. In addition, CI carries significant costs, both for the procedure and subsequent rehabilitation, which are often considered as better allocated to children with bilateral hearing loss. Systematic reviews of hearing rehabilitation options for children with SSD have not found one to be decisively superior (Kubina et al. 2023). Nevertheless, several publications of positive outcomes from international groups utilizing CI in a pediatric population suggest the need for a review of our own experience in Western Australia (Brown et al. 2022; Hicks et al. 2024).

OBJECTIVE

To report the long-term outcomes of CI for the management of children with SSD in an Australian tertiary pediatric center.

Methods and CI Protocol

A retrospective review of data from the Western Australian Children’s Hearing Implant program (CHIP) was performed to identify all pediatric patients with SSD who underwent CI over a 10-year period between 2014 and 2023. Our program serves as the main referral center for pediatric hearing loss and early CI surgery for Western Australia, a state with a population of approximately 3 million. Patients were included if they met the following criteria: congenital or acquired SSD up to the age of 16 yr old and had undergone CI surgery using any United States Food and Drug Administration–approved device. The age of inclusion for this study was selected based on the upper limit for specialist care at our tertiary pediatric institution. After 16 yr of age, patients are transitioned to respective adult health services. In our cohort, SSD was defined as unilateral severe-profound sensorineural hearing loss (unaided pure-tone average, i.e., ≥90 dB) and monosyllabic word recognition score of ≤40% in the worse hearing ear to be implanted, with normal hearing (pure-tone average ≤25 dB, word recognition score >90%) on the contralateral side.

Before surgery, all potential candidates underwent a rigorous workup and extended counseling. Patients were discussed and approved for CI at a multidisciplinary team meeting consisting of implant otolaryngologists, audiologists, speech and language pathologists, radiologists, geneticist, and a clinical nurse specialist. Preoperative measures included magnetic resonance imaging targeted to the internal acoustic canal, high-resolution computer tomography of temporal bones, language assessment via an age-appropriate tool, audiological thresholds via auditory brainstem response testing, otoacoustic emissions, tympanometry, otoscopy, and fitness for surgery and anesthetic assessment. Details of the tools used for age-appropriate language assessment are outlined in detail in Appendix 1, Supplemental Digital Content, https://links.lww.com/EANDH/B725. Intraoperative assessment of the CI position was conducted via either x ray, neural response telemetry (NRT), or both. Postimplantation, patients underwent intense auditory rehabilitation with a specialized pediatric team outlined in detail in Appendix 1, Supplemental Digital Content, https://links.lww.com/EANDH/B725. Full details can be found in the CHIP protocol (Department of Health. Government of Western Australia 2017). Patients followed up with their surgeon initially at 2 wk for a wound check, then at 2 mo following this.

A file review was performed to accurately identify patients who met the above criteria. The following outcomes were recorded: age at diagnosis, age at CI surgery, SSD duration (defined as the time between diagnosis and CI activation), etiology of hearing loss, follow-up duration, preoperative findings, intraoperative testing, type of implant, audiological and speech outcomes, and device adherence. Device log usage was acquired via custom-designed script from Cochlear (Sydney, Australia) applied to Custom Sound Pro database for all Cochlear devices. In MED-EL (Innsbruck, Austria) devices, we obtained data log through the Maestro software. Usage was recorded at 6 mo, 2 yr, and at the most recent follow-up appointment.

Patients were divided into two groups based on the criteria from Lee et al. (2020), and in line with previous studies from our institution (Garcia-Matte et al. 2024). Onset of SSD was determined by medical history and defined as occurring before 5 yr of age (early-onset SSD) before or at 5 yr of age, or older (late-onset SSD). Patients were further distinguished based on the timing of intervention, with those having a CI within 12 mo of diagnosis defined as early-CI surgery, and those after 12 mo as late-CI surgery.

Before audiological testing, all sound processors underwent a maintenance check and calibration to ensure they are in proper working order. This included a change of microphone covers, a device listening check, and electrical impedance measurement. In addition, all participants underwent otoscopy and tympanometry before testing to control for any middle ear dysfunction.

The seven-speaker layout used for the localization task is shown in Figure 1. A random number generator was used to create a series of 20 stimulus presentations from each of the 7 speakers and the participants used a pointing device (a toy wand) to indicate from which speaker they thought the stimulus (a 1 kHz pure tone) originated. A test assistant was made available to each participant to keep them on task and assist if required. Each response (speaker number) was depicted on a scatterplot for each participant relative to the correct origin of each stimulus for each of the 20 trials, with and without the use of their CI.

Fig. 1.

The speakers are number 1 to 7 in a clockwise fashion from the patients bearing. Speaker 1 is at 0° and the remaining six speakers are divided equally around the patient. The number inside the box represents the speaker number, while the number outside the box represents the degrees azimuth.

Comprehension amidst background noise was assessed via the Bamford-Kowal-Bench speech-in-noise (BKB-SIN) test. A 0:0 configuration was used as it was the only configuration with SNR loss norms. Two lists were administered if the patient would comply. Significance was defined as a p value of <0.05, determined by two-tailed t test. The critical difference was also calculated to determine the significance of an individual’s results based on their age and number of lists tested (Interacoustics 2024). In children who were unable to perform these due to age or compliance, surrogate measures, including aided thresholds and Northwestern University-Children’s perception of speech test (NUCHIPS) were used. Results were compared with and without the use of their CI where relevant. Patients and their caregivers were also given the Speech, Spatial and Qualities of Hearing Scale (SSQ) pediatric version to complete.

Ethical Considerations

Institutional ethics review was approved at Perth Children’s Hospital.

RESULTS

Twenty-two patients underwent CI surgery for SSD in the 10-y period. A total of 54.5% were male and 45.5% were female. The average age at diagnosis was 4.2 yr (range: 0.0 to 15.1, SD: 4.8). The average time to implantation was 1.3 yr (0.1 to 3.9 yr, SD: 1.3). Eleven (50%) patients were classified as early CI and 11 patients as late CI. The average age of patients at time of implantation was 5.5 yr (1.0 to 15.6 yr, SD: 4.7). Preoperative workup is summarized in Table 1. Patients had their diagnosis confirmed via auditory brainstem response or distortion product otoacoustic emissions. Magnetic resonance imaging was performed in 20 (90.1%) patients and computer tomography in 14 (63.6%). Twenty-one children (95.5%) were implanted with Cochlear devices, and one (4.6%) with a MED-EL implant. Seven (31.8%) underwent intraoperative x ray to assess electrode position, 19 (86.4%) underwent NRT with only 1 of these showing an inadequate response—a patient with known intracochlear fibrosis.

TABLE 1.

Demographics, preoperative and intraoperative assessment of cohort (separated into early and late SSD)

No.	Gender	Age at Dx	Etiology	Time to Implantation	Age at Implantation	Otoscopy	Tympanometry	Audiology Test	MRI	CT	Intraop test	Device
Early-onset SSD
1	M	0.00	Unknown	1.02	1.02	NAD	A	ABR	NAD	NAD	X ray, NRT	Cochlear CI 522 slim
2	F	0.00	CMV	1.06	1.06	NAD	A	DPOAE	Temporal pole cysts b/l (CMV)	Not performed	NRT	Cochlear CI 512
3	M	0.00	CMV	1.52	1.52	NAD	A	DPOAE	Large b/l caudothalamic groove cyst (?CMV)	Not performed	X ray, NRT	Cochlear CI 632
4	M	0.00	CMV	1.84	1.84	Not recorded	A	ABR	NAD	Not performed	NRT	Cochlear CI532
5	M	0.00	Unknown	2.59	2.59	NAD	A	DPOAE	NAD	NAD	NRT	Coclear CI24RE
6	F	0.00	CMV	3.40	3.41	NAD	C	VRA	NAD	Not performed	NRT	Cochlear CI 632
7	F	0.37	Unknown	0.69	1.06	NAD	A	ABR	NAD	NAD	X ray, NRT	Cochlear CI 522 straight
8	M	0.68	CMV	0.55	1.23	NAD	A	ABR	Not performed	NAD	Not recorded	Cochlear
9	F	1.26	Unknown	0.59	1.86	NAD	C	VRA	NAD	Not performed	Not recorded	Cochlear CI 532
10	M	1.34	CMV	0.99	2.33	GISAP	B-high volume	ABR	NAD	NAD	X ray, NRT	Cochlear CI622
11	M	2.05	Trauma	0.92	2.98	Blind sac	N/A	ABR	NAD	CWD mastoidectomy and blind sac. Early labyrinthitis ossificans. Sleeper electrode in situ	X ray, NRT (partial response)	Med El Flex 28
12	F	3.72	Trauma	1.35	5.08	NAD	C	DPOAE	Resolution of previous subdural hemorrhage and dural venous thrombosis	NAD	NRT	Cochlear CI 512
13	F	4.10	Meningitis	0.13	4.24	NAD	A	ABR	Cortical injury at bilateral parietal cortices	Not performed	Not recorded	Cochlear CI (contour advance CP910)
14	F	4.12	Unknown	3.93	8.07	NAD	A	DPOAE	Mild asymmetry between vestibular aquaduct. Both within normal limits	NAD	NRT	Cochlear 622
15	M	4.38	SSNHL	1.01	5.40	NAD	A	DPOAE	Right SCC dehiscence possible. Probably left labyrinthitis ossificans.	Left labyrinthitis ossificans.	X ray, NRT	Cochlear CI 612
16	M	4.93	Unknown	0.50	5.45	NAD	A	DPOAE	Not performed	NAD	NRT	Cochlear CI24RE
Late-onset SSD
1	M	5.47	EVA	2.03	7.53	NAD	A	DPOAE	Right EVA.	Right EVA	NRT	Cochlear CI512
2	F	7.35	Trauma	0.10	7.46	NAD	A	OAE	PLF. Intralabyrinthine hemorrhage. Evolving granulation fibrosis	CT no labyrinthitis ossificans. NAD	NRT: all but 1 basal electrode	Cochlear CI622
3	F	10.84	SSNHL	2.74	13.58	NAD	A	PTA	NAD	Not performed	NRT	Cochlear CI632
4	M	12.51	Trauma	0.68	13.23	NAD	A	PTA	Obliteration of scala vestibuli of basal turn of cochlea. Fibrosis of labyrith.	(Later than MRI) No ossification of cochlear	X ray, NRT present but not in all electrodes 2nd to fibrosis	Cochlear CI612
5	M	13.25	EVA	0.56	13.85	NAD	A	PTA	Endolymphatic sac dilatation. Mild cochlear dysplasia. CN 8 normal	Not performed	NRT	Cochlear Cl24RE
6	F	15.05	Unknown	0.52	15.62	Tympanosclerosis	A	DPOAE	NAD	NAD	NRT	Cochlear CI512

ABR, auditory brainstem response; CMV, cytomegaly virus; CT, computer tomography; DPOAE, distortion product otoacoustic emissions; EVA, enlarged vestibular aqueduct; GISAP, grommet in situ and patent; MRI, magnetic resonance imaging; NAD, no abnormality detected; NRT, neural response telemetry; SSD, single-sided deafness; SSNHL, sudden sensorineural hearing loss; PTA, pure-tone audiometry; VRA, visual reinforced audiometry.

Six patients (27.3%) were classified as late-onset SSD and 16 (72.7%) as early SSD. Late-onset patients had passed newborn hearing screens and subsequently developed SSD due to a clear etiology, such as trauma or infection. It is crucial to clarify that these cases are not early-onset SSD with a delayed diagnosis, but represent a separate subgroup with distinct onset patterns, which has implications for analysis, management, and expected outcomes. The etiology of early-onset SSD was unknown in 8 patients (36.7%). Five (22.7%) had congenital cytomegalovirus and 2 (9.1%) had enlarged vestibular aqueduct syndrome. For late-onset SSD, 4 (18.2%) were sequelae of trauma, 2 (9.1%) were due to sudden sensorineural hearing loss, and 1 (4.6%) was related to meningitis.

The average length of follow-up was 5 yr (range: 1.4 to 9.5). Six patients are no longer being actively reviewed in the program (1 patient switched to a CROS aid, 1 patient was a nonuser declining further follow-up, 1 patient relocated and was uncontactable, 2 patients transitioned to adult services, and 1 had been lost to follow-up). The average CI usage was 4 hr/d (range: 0 to 12.9, SD: 3.7). The average use in the early phase (3 to 6 mo post-CI) was 5.5 hr/d. Table 2 outlines the follow-up duration and CI usage for our cohort.

TABLE 2.

Duration of follow-up and CI usage

No.	Ongoing Follow-Up?	Duration of Follow-Up	CI Usage Early	CI Usage Current	Comment
Early-onset SSD
1	Y	6.30	-	3.00	Mother has expressed concerns regarding if CI is helpful or not
2	Y	7.35	-	0.50	Lives rurally. Was using CI 5 hr/d at 4 yr follow-up
3	Y	1.40	6.00	7.00
4	N	0.35	0.00	0.00	Multiple missed appointments and missed operation date once before subsequently being implanted. Only attended 2 appointments post-CI. Has since moved rurally
5	Y	8.96	3.50	1.30	Lives rurally
6	Y	4.05	5.00	3.80
7	N	4.90	8.00	3.50
8	N	9.52	-	0.00	Lives rurally
9	Y	4.69	4.00	2.20	Lives rurally
10	Y	2.73	8.00	8.80	Implant extrusion starting with 1 electrode in 3 mo post-CI. Most recent NRT shows only 6 electrodes in situ
11	Y	4.04	5.00	7.60
12	Y	7.19	-	2.00
13	Y	7.45	8.00	10.90
14	Y	3.63	6.00	3.90
15	Y	4.53	6.00	6.20
16	N	3.00	4.00	0.00	Pt prefers using CROS aid and has since been explanted surgically
Late-onset SSD
1	Y	7.92	9.00	0.00	Was using CI for 5 yr postop until patient decided to be nonuser
2	Y	4.15	7.00	2.00
3	Y	1.98	4.70	3.60
4	Y	1.62	2.00	2.00	States low use at school due to social pressure
5	N	4.15	-	12.90	Transitioned to adult services
6	N	5.36	8.00	6.00	Transitioned to adult services

CI, cochlear implantation; CROS, contralateral routing of signal; NRT, neural response telemetry; SSD, single-sided deafness.

Speech assessment was performed preoperatively in 17 of 22 patients via a variety of tools depending on patient factors, time of assessment, and speech pathologist preference. Seven patients were assessed with clinical evaluation of language fundamentals, 4 with the Receptive Expressive Emergent Language Test, 4 with the Preschool Language Scale, and 2 with the Rossetti Infant Toddler Language Scale. Twelve patients were retested 1 yr post-CI activation. In the postoperative setting, 9 patients were assessed with clinical evaluation of language fundamentals, 2 with Preschool Language Scale, and 1 with the Parents Evaluation of Aural/Oral Performance. Table 3 shows the pre- and post-CI language outcomes for our cohort. Three of our patients showed improvement in language ability 1 yr after implantation. No patients scored poorer compared with their pre-CI language assessment. Fourteen patients did not complete both preoperative and/or postoperative language assessment, thus, no conclusions can be drawn on their change in language ability.

TABLE 3.

Pre and post CI language assessment

No.	Assessment Tool (Pre-CI)	Core Language Percentile	Receptive Language Percentile	Expressive Language Percentile	Assessment Tool (1 yr Post-CI)	Core Language Percentile	Receptive Language Percentile	Expressive Language Percentile	Change
Early onset
1	Rossetti	Age appropriate			CELF	70	42	84	Remains age appropriate
2	Nonet				PEACH	86			Age appropriate
3	REEL 4	39	45	37	PLS		45	19	Remains average
4	Rossetti	Age appropriate			None
5	PLS5	37			CELF	75	30	61	Remains average
6	REEL	82	89	63	CELF	27	3	12	Remains below average but core language ranks in the average zone
7	REEL	-	9	42	PLS	37	50	25	Improved from below average to average
8	None	-			None
9	PLS 5	-	39	34	None				Average
10	PLS 5	75	73	75	None				Average
11	PLS5	-	12	2	None				Below average
12	CELF	95	95	95	CELF	98	98	98	Improved from above average to well above average
13	REEL 3		39	25	None				Below average
14	CELF	9	13	4	None				Below average
15	None	-			CELF	12	5	19	Below average
16	CELF	79	95	68	None				Average
Late onset
1	None				None
2	None				CELF	73	61	79	Average
3	CELF	55	81	32	None				Average
4	CELF	16	5	23	None				Below average
5	CELF	87	86	86	CELF	87	96	75	Remains above average
6	CELF	12	32	9	CELF	53			Improved from below average to average

CELF, clinical evaluation of language fundamentals; CI, cochlear implantation; PEACH, Parents Evaluation of Aural/Oral Performance; PLS, preschool language scale; REEL, receptive expressive emergent language test.

Six of the 22 patients (27.27%) underwent localization testing as this method was introduced later in the 10-yr interval. The remaining patients could not attend or did not respond to contact for retesting. These 6 patients were no better at identifying the correct speaker with their CIs than without (p = 0.93) as shown in Figures 2 and 3. The mean absolute difference is the average difference in the number of consecutive speakers between the actual stimulus and the perceived stimulus. The mean absolute difference with CI was 1.88 speakers, the mean absolute difference without CI was 1.90 speakers.

Fig. 2.

Response positions (degrees azimuth) are plotted against stimulus presentation positions (degrees azimuth) for n = 6 participants who performed this testing. No significant relationship was found when patients were tested: (A) with their CI (R² = 0.03, p > 0.05) or (B) without their CI (R² = 0.03, p > 0.05). CI indicates cochlear implantation.

Fig. 3.

The difference from the target is plotted against the stimulus speaker position with the CI on (A) and without the CI (B) with 0 being no difference (i.e., correctly identifying the source of stimulus). A positive number indicates the number of consecutive speakers in an anticlockwise direction (without crossing 0°) that the patient perceived the stimulus to be coming from. Conversely, a negative number indicates the number of consecutive speakers in a clockwise direction (without crossing 0°), that the patient perceived the stimulus to be coming from. There was no significant differences in mean absolute differences between CI on and off conditions (R = 0.20, p > 0.05). CI indicates cochlear implantation.

Thirteen of 22 (59.09%) patients completed BKB-SIN testing. Of the remaining 9, 4 completed other tests which were age appropriate (e.g., NUCHIPS). Five patients performed better in their BKB-SIN with their CI (average of 1.98 with CI versus 4.36 without CI) and 8 performed worse with their CI (average of 4.1 with CI versus 2.6 without CI). Of the 5 children who performed better in BKB-SIN testing with CI, 4 had early CI surgery: 2 with early-onset SSD and 2 with late-onset SSD. Conversely, among the 8 who performed worse with CI, 4 had early-onset SSD and all but 1 were late implantations. 95% critical difference thresholds were met by 3 of 5 who improved versus 1 of 8 who worsened.

The average SNR loss in the early SSD group was 2.91 (range: −1.3 to 6) with CI and 2.84 (range: −0.5 to 5.5) without CI (p = 0.462). For late-onset SSD, the average with CI was 3.72 (range: −0.01 to 6) and 3.78 (range: −0.1 to 7.7) without CI (p = 0.475). The overall average for all patients, with and without CI, was 3.28 versus 3.28. Table 4 highlights the BKB-SNR results as well as the other age-appropriate tests.

TABLE 4.

Speech in noise results (BKB-SIN, NUCHIPS, and PLOTT)

No.	SNR Loss With CI	SNR Loss Without CI	Comment
Early SSD
1	−1.3	1.7	Meets 95% critical difference threshold
2	-	-	Nonuser
3	PLOTT score 3/12	-	Unable to perform BKB, thus PLOTT test utilized
4	-		Did not attend
5	1.7	−0.3
6	-	-	Minimal detection of sound, even at levels well above comfort. Recent NRT shows a flat profile
7	6	5
8	NUCHIPS 70% phenome		NUCHIPS
9	PLOTT score 11/12		Unable to perform BKB, thus PLOTT test utilized
10	-
11	-		ADHD
12	1.5	−0.5
13	4	5.5
14	NUCHIPS 66% phenome, 47% word		NUCHIPS
15	6	5
16	2.5	3.5
Average	2.91	2.84
Late SSD
1	4.7	1.7	Meets 95% critical difference threshold
2	4.7	7.7	Meets 95% critical difference threshold
3	1.4	−0.1
4	−0.01	3.4	Meets 95% critical difference threshold
5	6	5
6	5.5	5
Average	3.72	3.78
Combine early- and late-SSD average	3.28	3.28

PLOTT: a monosyllabic word speech perception test (see Appendix 2 in Supplemental Digital Content, https://links.lww.com/EANDH/B729). SNR loss: the increase in SNR that is required by the listener to obtain 50% of key words correctly as compared with normative data from matched peers of the same age with normal hearing.

BKB-SIN, Bamford-Kowal-Bench Speech in nose; CI, cochlear implantation; NUCHIPS, Northwestern University-Children’s perception of speech test; SSD; ADHD, attention-deficity hyperactivity disorder.

Eight patients (34.78%) completed the SSQ. Figure 4 shows the average ratings for these patients. The SSQ in children is subdivided into 3 sections comprising 33 questions total. Like the SSQ for adults, the questions address the child’s ability to listen to speech, spatial hearing ability, and qualities of hearing. Scores are graded on a visual analog scale 0 to 10, with 10 being the maximum score and 0 the minimum. The average overall score for all patients was 6.9. The average score for all patients by section was 7.2 for speech, 5.9 for spatial hearing, and 7.8 for quality of hearing. Figure 4 shows the average scores of patients, separated by individual sections. A pre-CI SSQ was not performed and thus no comparison can be made before and after CI based on SSQ scores.

Fig. 4.

Average SSQ scores per patient, by section. Blue: Speech; orange: spatial hearing; green: qualities of hearing. SSQ indicates Speech, Spatial and Qualities of Hearing Scale.

A subgroup analysis was performed to identify patients who would benefit more from CI than others. Patients in the early- and late-onset SSD were further stratified into early-onset SSD + early-CI surgery, early-onset SSD + late-CI surgery, late-onset SSD + early-CI surgery, and late-onset SSD + late-CI surgery. Out of the 5 patients who benefited with CI in BKB-SIN, 4 (80%) had early CI (2 early-onset SSD and 2 late-onset SSD). The last patient who benefited was classified as a late CI but had only marginally missed the 1-yr cut-off (implant at 1.01 yr from diagnosis). Of those who benefited with CI for localization testing, 3 (75%) were in the early CI subgroup (2 early-onset SDD, 1 late-onset SSD). Both patients who did not benefit from CI for localization testing were in the early-onset + late-CI subgroup.

DISCUSSION

Current evidence supports the use of CI for the treatment of pediatric SSD in selected cohorts. (Benchetrit et al. 2021; Deep et al. 2021; Selleck et al. 2021).The only meta-analysis to date evaluating CI for pediatric SSD was conducted by Benchetrit et al. (2021) comprising 12 studies, and found significant improvements in audiological and patient-reported outcomes, including speech perception in noise.

Since then, several studies have reported generally positive results. Park et al. (2023a) demonstrated improvements in binaural hearing of 18 children. Arras et al. (2021) reported that 15 prelingual children with SSD who received a CI performed comparably with their normal-hearing peers in grammar testing. Brown et al. (2022) similarly reported significant improvements in BKB-SIN, SSQ, and localization outcomes in 20 patients.

In contrast, our study demonstrated more variable audiological outcomes. As a group, no statistically significant improvements in localization or speech in noise performance were observed. SSQ results were similarly inconsistent, and only some patients showed gains in language assessments. We observed a high proportion of nonusers (18.18%) and limited users (22.73% using ≤2 hr/d). In our experience, compliance is a great challenge in managing pediatric SSD. Potential explanations include lack of perceived benefit, unpleasant stimulation, social pressures, or increasing autonomy as the child becomes older. Garcia-Matte et al. (2024) similarly noted a decline in use, particularly after 2 yr, and Zhan et al. (2020) reported a high rate of reduced or discontinued CI usage (84.7%) with prolonged follow-up. Thomas et al. (2017) also found that longer follow-up was associated with higher rates of nonuse, with 3 out of 5 patients who were followed to 3 yr being limited or nonusers. Studies which observed a shorter follow-up period (Polonenko et al. 2017; Sangen et al. 2019; Ganek et al. 2020; Arras et al. 2022) or relied on self-reporting via questionnaires (Thomas et al. 2017; Rauch et al. 2021) reported higher device usage.

Device usage, captured via datalogging, is an important objective measure which correlates with better auditory and speech outcomes (Easwar et al. 2018; Busch et al. 2020; Arras et al. 2022, 2024). However, this relationship is likely bidirectional: greater use may promote better outcomes, while better performance may encourage use (Sangen et al. 2019; Arras et al. 2024). Children with SSD may find it difficult to appreciate the benefit of a CI, as even optimal function cannot fully replicate natural hearing (Arras et al. 2024). Other reported hindrances to SSD patients include lack of perceived benefit, discomfort, social stigma, difficulty managing the device, and fear of damage or loss (Polonenko et al. 2017; Thomas et al. 2017; Benchetrit et al. 2021; Park et al. 2023b; Arras et al. 2024; Garcia-Matte et al. 2024). These issues explain declining use over time, as observed in our study and others. Longer-term studies like ours are crucial to see if these positive outcomes of CI in SSD stand the test of time.

These concerns are particularly relevant in the Australian context, which may explain why our results differ from those of other international centers. In Australian schooling environments, where physical activities—particularly water-based ones—are strongly promoted, wearing a device may be impractical, interfere with peer integration, or lack active support from educators. Compared with other studies, ours captured more data during the primary school years, a period when device use appears to be more inconsistent. Their time spent in different acoustic environments may also play a role as suggested by some authors; however, this was not explored in our datalogging (Polonenko et al. 2017; Ganek et al. 2020). In addition, patients in Western Australia may travel great distances from rural and remote communities, and the logistical barriers associated may discourage both consistent CI wear and attendance at rehabilitation sessions.

Furthermore, our cohort benefits from government-funded CIs. Although promoting equitable access, this may inadvertently reduce the perceived value of the device and lower motivation for consistent use (Raine et al. 2008; Promberger & Marteau 2013). Lipschitz et al. (2020) found that private health insurance or higher socioeconomic status were associated with better usage, although significance was lost on multivariate analysis. Lim et al. (2022) identified poor family commitment—due to factors such as travel burden, family dynamics, and household income—as a key predictor of poor CI outcomes. In our cohort, 2 patients opted for private funding; 1 became a nonuser after 9 yr despite initial parental persistence, suggesting that financial investment may influence short-term adherence but is not protective against eventual nonuse.

Delayed intervention and longer auditory deprivation have been implicated in poorer outcomes (Benchetrit et al. 2021; Rauch et al. 2021; Park et al. 2022). Emerging data suggest better outcomes with earlier implantation—before 3 to 4 yr for congenital SSD and within 7 yr for acquired SSD (Benchetrit et al. 2021; Deep et al. 2021; Rauch et al. 2021; Spitzer et al. 2024). Our cohort was well within these benchmarks, with a mean time to CI of 1.38 yr (early SSD) and 1.11 yr (late SSD). The age at implantation may also be relevant. Arras et al. (2024) reported high device usage; however, all their patients had prelingual SSD and were implanted before 2.5 yr of age. Other studies suggest better outcomes in those implanted before age four (Deep et al. 2021; Rauch et al. 2021). While young children rely on parental engagement; older children may show autonomy and independent consistent use (Ganek et al. 2020; Deep et al. 2021; Park et al. 2023b; Spitzer et al. 2024). Garcia-Matte et al. (2024) found longer use in those diagnosed later in life, while Polonenko et al. (2017) noted that their most consistent users were teenagers with sudden SSD. Conversely, other studies have reported lower usage in teenagers with SSD, potentially reflecting psychosocial or developmental challenges unique to this age group (Lee & Sim 2020; Gordon et al. 2023).

Audiological test methods may have influenced outcomes. Due to equipment limitations, a pure-tone stimulus was used for localization testing. Although calibrated to minimize standing wave concerns, this may have impacted results (Institute for Cochlear Implant Training 2021). Similarly, with speech in noise testing, we have only reported on the 0:0 orientation of BKB-SND testing which is not in line with the Minimum Speech Test Battery 3 (MSTB-3) (Institute for Cochlear Implant Training 2021). While some children were compliant with testing in additional orientations, the majority of our cohort’s attention span was consumed with two lists in the 0:0 orientation. Some adult studies did not find any difference in the 0:0 orientation for speech in noise testing for SSD patients with CI but reported benefit in other orientations. Sladen et al. (2024) reported no significant difference in results. Vlastarakos et al. (2014) similarly found inconclusive results when both speech and noise were presented from the front.

Contrary to this, other studies focusing on pediatric SSD have reported significant findings in the 0:0 orientation. Brown et al. (2022) found significant improvement in the 0:0 orientation along with other orientations; Benchetrit et al. (2021) reported a 79.6% improvement in speech in noise measured in the 0:0 orientation. These suggest that although some benefit of CI in SSD may be missed by not testing various spatial conditions, the 0:0 orientation, which we believe best reflects the real-world situation, is likely sufficient to identify children who benefit from CI compared with those who do not. Importantly, we calculated 95% critical differences to identify meaningful change.

Despite some studies demonstrating improvements in spatial hearing following CI (Benchetrit et al. 2021; Brown et al. 2022), our cohort showed no significant gains in localization accuracy, with a mean absolute difference of 1.88 speakers with CI versus 1.90 without. This overall poor localization performance suggests that CI may offer limited benefit in spatial hearing for some children with SSD, potentially due to delayed implantation, cortical reorganization, or inconsistent device use. These findings align with emerging data suggesting that true binaural integration may remain challenging to achieve in this population, particularly without early intervention.

Publication bias may contribute to overly favorable impressions of CI success in this population. Many studies report small cohorts, short follow-up, and selective outcomes, potentially failing to identify patients who ultimately become nonusers beyond the study period. Some studies rely on strict selection criteria or focus on outcomes, such as speech in quiet, that may naturally yield better outcomes rather than including all pediatric CI recipients. Collectively, these factors suggest that early reported benefits may not be sustained over time or may be more subtle than initially perceived. Longer-term studies, such as ours, are crucial to determine whether initial positive outcomes persist and to better understand the factors influencing sustained success.

While overall results were variable, some individuals clearly benefited: 5 of 13 performed better in BKB-SIN, 3 of 8 improved in language, and 7 of 22 used their CI >6 hr/d. Notably, our results did not show a significant difference in outcomes between early- and late-onset SSD groups.

Subanalysis suggests that earlier implantation, regardless of age of SSD onset, may yield superior outcomes. Our cohort consisted of 50% early CI surgeries and 50% late-CI surgeries; yet 80% of those with improved speech-in-noise performance and 75% of those with improved sound localization were from the early CI surgery group. The only 2 patients in the late-CI surgery group demonstrated benefit—1 in BKB-SIN (implanted 1.02 yr after diagnosis) and 1 in localization (implanted 1.01 yr after diagnosis)—were implanted just beyond the 12-mo threshold and are functionally comparable to the early CI group. Although we are mindful of the small numbers involved, findings suggest an emerging pattern: for both early- and late-onset SSD, earlier implantation may be associated with better long-term objective outcomes.

Finally, our cohort was largely implanted with cochlear-branded implants which generally manufacture modiolar electrodes. One hypothesis is that an individualized straight electrode that conforms around the cochlear better may be important in early-onset pediatric SSD. Whether alternative arrays better suited to pediatric anatomy influence outcomes warrants future investigation. While long-term large trials would be ideal, there are significant practical barriers in conducting such studies for this population, rendering treatment decisions challenging.

LIMITATIONS

Despite spanning a decade, the cohort remains small due to the rarity of pediatric SSD and strict CI candidacy criteria. This sample size, while comparable to other published studies, restricts the statistical power and the generalizability of subgroup findings.

Second, cohort heterogeneity complicates interpretation and necessitates a personalized rehabilitation process for each patient. Every child is managed by their individual audiologist and speech pathologist, each with slight variations in their approach and training. The retrospective nature of this study further limits control of confounders. While some prospective studies with stricter candidacy criteria and standardized rehabilitation protocols have reported more favorable short-term outcomes, longer-term results appear less predictable, as reflected in our cohort with extended follow-up.

Finally, not all our patients completed localization and the SSQ due to the later implementation of these measures and patient compliance. Our center halted routine audiological operations to set up the department for the specific localization testing in November 2021. Furthermore, most of our cases were implanted before 2021 with a lull period during the COVID-19 pandemic. Only 3 SSD patients were implanted after November 2021; thus, it was not feasible to offer localization testing (or repeat testing) to all our patients. After recruiting more patients, our institution will investigate retesting to increase the accuracy of our results. While measures such as the 95% critical difference help interpret test significance, consistent testing and long-term follow-up in pediatric populations can be difficult to achieve.

CONCLUSION

Despite a well-established newborn hearing screening program and equitable access to specialized services, long-term outcomes in our cohort were variable. While several patients showed benefit in speech in noise, high usage rates, and improved language skills, the challenge remains on how to both predict and rehabilitate a heterogenous population of pediatric SSD patients consistently. Our findings support earlier implantation as a potentially modifiable factor linked to better outcomes.

This study contributes valuable long-term data to the evolving field of cochlear implantation in children with SSD. Larger, prospective studies are needed to build on these findings, though challenges in recruitment and retention for this population must be recognized. In the interim, pediatric otology centers should prioritize careful patient selection, timely implantation, and consistent rehabilitation to support meaningful benefits.

ACKNOWLEDGMENTS

The authors thank Elizabeth Park, senior audiologist, who collected the initial data and performed much of the testing; Raymundo Garcia, otolaryngology fellow, who helped with data collection; Pia Leeming, senior audiologist, who contributed to the planning of this study and performed testing; Tooey Tran, CHIP coordinator, who facilitated the organization and care of children for testing and data recruitment.

Supplementary Material

aud-47-041-s001.pdf ^{(218.7KB, pdf)}

aud-47-041-s002.pdf ^{(536KB, pdf)}