Early Deficits in Speech Perception in Carriers of the p.Pro51Ser Variant in the COCH Gene: A Prospective Longitudinal Evaluation of Speech Perception in Quiet and Noise

The primary concern among adults regarding their hearing is the challenge of understanding speech, especially in noisy environments [1, 2]. In demanding listening situations, there is a need to use more cognitive resources, and listening effort, i.e., the mental exertion necessary to comprehend the auditory message [3], is substantially increased [4‒6]. This burden on listening effort induces frustration, fatigue, and decreased concentration ability [7]. Such difficulties may cause hearing-impaired individuals to cease attempts to engage in communicative situations, leading to social withdrawal [8]. As a consequence, hearing loss is correlated with an increase in social isolation, feelings of loneliness, and depression [9, 10]. Research indicates that difficulties in understanding speech and communication are often caused by a gradually developing high-frequency hearing loss. More precisely, the reduction and distortion of acoustic signals cause difficulties with speech discrimination [11, 12]. Many studies have emphasized the critical contribution of higher frequencies to understanding speech and unraveling speech from noise [13‒18].

One of the most frequent types of postlingual non-syndromic high-frequency hearing loss in Belgium and The Netherlands is DeaFNess Autosomal Dominant 9 (DFNA9) [19, 20]. Patients with DFNA9 in this region are most frequently carriers of the c.151C>T variant in the COCH gene, caused by a substitution of cytosine by thymine nucleotide of the 151th base pair in codon 51, p.Pro51Ser (P51S) [21‒23]. As a result of this mutation, patients experience a progressive deterioration of high-frequency hearing thresholds and loss of vestibular function with a mid-life onset. Their sensorineural hearing loss (SNHL) will eventually result in severe hearing impairment and a profound loss at an older age [24‒26]. Several studies, using both cross-sectional and longitudinal designs, have investigated the way in which SNHL develops during DFNA9 patients’ lifespan. Recent studies by our research group have identified the initial signs of hearing loss in both male and female carriers as early as 30 years of age. Additionally, a rapid decline in SNHL throughout all frequencies has been observed between the ages of 40 and 50 [27].

However, it should be considered that speech perception in patients with DNFA9 may evolve differently from liminal audiometry. Investigating speech perception in quiet (SPIQ) and noise (SPIN) in patients with the p.Pro51Ser mutation could provide useful information regarding the disease stage, and provide clues regarding its underlying pathophysiology. Additionally, comparing these results with those from a normal hearing group could provide more insight into the onset of hearing loss during the presymptomatic phase.

Furthermore, it can be stated that SPIQ and SPIN performance across the lifespan of carriers of the p.Pro51Ser mutation in the COCH gene remains largely unexplored. Literature review shows that only two cross-sectional phenotype studies have been performed up to now [28, 29]. The first objective of this study was to study in a prospective manner how speech perception, both in quiet and in noise, varies with respect to baseline age by means of a longitudinal design. Additionally, a comparison will be made between these speech perception scores and those of a normal hearing group. The second objective was to explore the correlation between pure-tone audiometry and speech perception, both in quiet and in noise.

This is a longitudinal, prospective study running from February 2019 to September 2022. It was led by the Department of Otorhinolaryngology at the Antwerp University Hospital (UZA) and, in cooperation, conducted at Jessa Hospital in Hasselt and the Leuven University Hospital. The study adhered to the Declaration of Helsinki (1996) and was approved by the Local Ethics Committee of the Antwerp University Hospital (B300202042807). ICH-GCP accredited researchers (masters in audiological sciences) performed a predetermined study protocol including pure-tone audiometry, SPIQ, and SPIN measurements. The tests were performed in a soundproof booth using a two-channel Interacoustics AC-40 audiometer and headphones. All tests were conducted in unaided situation. The protocol was repeated annually, under the same conditions. Informed consent was obtained from all participants prior to the start of the study. An online password-protected database, Research Electronic Data Capture (REDCAP) was used for data storage and processing [30].

For the longitudinal evaluation of speech perception, we included a total of 101 heterozygous carriers of the p.Pro51Ser variant in the COCH gene. In addition, we recruited a control group of individuals with normal hearing, according to their age (based on ISO-7029 [31]). Patients with DFNA9 in the study were individually matched for age and sex to control participants. Matching details of both study groups can be found in Table 1.

The DFNA9 patients included in this study were genetically confirmed carriers of the p.Pro51Ser variant in the COCH gene. All patients presented at the outpatient clinic with a family history of the p.Pro51Ser pathogenic variant in the COCH gene. Single COCH gene testing was performed evaluating exons 4 and 5 of the COCH gene. All patients have European ancestry and have the Belgian or Dutch nationality. Moreover, the actual position of this identified variant is 14-30877640-C-T (GRCh38) and the reference sequence for c.151C>T is annotated on transcript NM_001135058.1 [32]. All carriers aged 18 years and older were eligible for the study. During the initial study visit, the patients were interviewed to determine eligibility. Participants were excluded if they had active middle ear conditions resulting in conductive hearing loss with an air-bone gap larger than 15 dB hearing Level (HL), or if they had SNHL or vestibular dysfunction due to reasons other than DFNA9, known neurological or central nervous system disorders, intracranial disease/tumor. Additionally, individuals who were unwilling or unable to attend the yearly study visits were also excluded. DFNA9 patients were recruited and tested at the Antwerp University Hospital, Jessa Hospital in Hasselt, and Leuven University Hospital from the beginning of 2019 onwards (as illustrated in online suppl. Digital Content Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000543263).

Pure-tone audiometry was conducted to evaluate unaided HLs. In a soundproof booth, air conduction thresholds were first determined using a two-channel Interacoustics AC-40 audiometer and headphones. HLs (in dB HL) at 125, 250, 500, 1,000, 2,000, 3,000, 4,000, 6,000, and 8,000 Hz were measured in both ears. Next, bone conduction thresholds were measured using a bone conductor (B71 transducer) within the frequency range of 250–4,000 Hz. If a participant was unable to detect a tone, the highest audiometer output level was documented as the threshold level (i.e., 120 dB HL). The pure-tone average (PTA) was determined by averaging pure-tone thresholds at 1, 2, and 4 kHz.

Speech audiometry was performed with the same equipment for each ear separately. The Dutch open-set NVA (Nederlandse Vereniging voor Audiology) lists developed by the Dutch Society for Audiology [33‒35] were used to evaluate SPIQ. For every list of CVC words, consisting of 12 words, the initial item was considered a training item and discarded for scoring. Correctly, reproduced consonants and vowels, 33 for every list, were scored to result in a percentage correct (phoneme) score. The evaluation of speech perception was carried out by measuring the speech reception threshold (SRT) and the Indice de Capacité Auditive (ICA). The SRT was calculated by identifying the intensity (in dB SPL) at which the patient correctly discriminates 50% of the phonemes. To obtain this 50%-point, all necessary presentation levels were used when dealing with individuals who have significantly varying degrees of hearing loss. The maximum intensity was set to 100 dB SPL, as it was the highest output level achievable with the audiometer. The ICA referred to the average percentage of speech intelligibility at intensity levels of 40, 55, and 70 dB SPL (in %). The test-retest reliability for CVC phonemes was 2.6 dB. This is in close agreement with the theoretical value [35].

SPIN was evaluated using the Leuven Intelligibility Sentence Test (LIST) with an adaptive procedure (van Wieringen and Wouters, 2008). In this procedure, the level of speech-weighted noise was fixed at 65 dB SPL, while the level of the speech signal was adjusted by the audiologist depending on the response of the patient. The starting level of the speech signal was 65 dB SPL. Every sentence had multiple keywords. When all keywords were repeated correctly, the level of the next sentence was decreased by 2 dB SPL. Otherwise, the level was increased by 2 dB SPL. Every list included 10 sentences. To determine the SRT, we calculated the mean level of the final five sentences together with the level of the imaginary 11th sentence of the list. The SRT represented a speech-in-noise ratio (dB SNR), where higher thresholds corresponded to a poorer ability to comprehend speech in noisy conditions. If the SRT exceeded 20 dB SNR, the SRT value was recorded at 20 dB SNR. The test-retest reliability for the LIST is estimated at 1.17 dB (and reference slope of 17.8%/dB). A sentence-in-noise test with a reliability around 1 dB and a slope above 15%/dB is generally considered an accurate test [34, 36, 37], suitable both in a clinical setting as well as for research.

Statistical analyses were executed using the statistical software package R, version 4.1.2 (R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2021). The data assembled in this study constitutes of a cross-sectional and longitudinal component, covering a wide age range in a short study duration. More specifically, baseline age of the participants ranged from 18 to 75 years and patients contributed with a variable length of follow-up, with a maximum of four consecutive years. Therefore, individual responses of speech perception scores as a function of age were visualized using spaghetti plots, created with ggplot2 (H. Wickham. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2016). In addition, a lowess line representing the group mean rates of change for both males and females was fitted to the data. In order to investigate the PTA-SRT relationship, a plot of SRT values as a function of pure-tone-averages of 1, 2, and 4 kHz was drawn.

Table 2 shows the demographic details of age and sex in the total study population of 101 patients at baseline. Consequently, audiological thresholds and evaluation of SPIQ and SPIN scores were reported for 202 ears. The mean age in the DFNA9 population was 43.83 years (SD = 12.03 years), with a minimum of 18 years and a maximum of 67 years. Within this study group, 53% were female, and 47% were male. The average number of years patients were tested in this prospective longitudinal study was 2.45 years with a minimum of 1 year and a maximum of 4 years.

First, the longitudinal trajectories of an individual patient’s SRT value within the follow-up window were visualized by individual lines for both male and female carriers (Fig. 1). Data are shown in the age window from 18 to 70 years. For both males and females, a superimposed lowess curve was determined as a nonparametric estimator of the mean change over time.

As shown in Figure 1, a (near) normal SRT for both male and female carriers between the ages of 18 and 29 years (second and third decade) is found. Specifically, as noted in Table 3, the mean SRT for this age-group is 28.18 dB SPL (±3.67) and 29.29 dB SPL (±3.01) for males and females, respectively. From the age of 35 years onwards, a steep increase in SRTs is evident, resulting in a maximum SRT of 100 dB SPL around the age of 70 years for both groups. However, between the ages of 50 and 59 years, there seems to be a sex difference, with female carriers performing worse than male carriers. More specifically, the average SRT for females in the sixth decade is 77.00 dB SPL (±22.88) compared to an SRT of 62.63 dB SPL (±19.03) for males. However, a large variability between the patients can be noticed, especially in the 40–59 age-group, where the standard deviation varies from 17.74 to 22.88 dB SPL, although this could be attributed to the large variation in PTA thresholds. For male and female carriers older than 60 years, the mean SRT is 97.05 dB SPL (±4.96) and 88.93 dB SPL (±17.19), respectively.

To compare these speech perception results with a matched group of normal hearing individuals, the mean SRT values of the control participants are added to Table 3. In this group, the SRT values for SPIQ show an increase from 25.41 dB SPL in the age range of 18–29 years to 34.56 dB SPL in the seventh decade for male participants. For female participants, the SRT values evolve from 23.89 dB SPL to 30.5 dB SPL. However, this increase is remarkably less than seen in the DFNA9 cohort. Differences between the DFNA9 and control group become apparent from the third decade, as visually shown in the figure, and become more pronounced in the fourth decade. In this decade, male carriers exhibit a SRT of 31.11 dB SPL, in comparison to 28.17 dB SPL in the control group. Female carriers have a SRT of 33.06 dB SPL, compared to 27.22 dB SPL in the control group. Furthermore, the mean SRT for male and female carriers in the fifth decade is 56.53 dB SPL and 53.54 dB SPL, in comparison with 29.38 dB SPL for male and 28.87 dB SPL for female control participants. During the sixth decade, male and female carriers have a mean SRT of 62.63 dB SPL and 77 dB SPL, respectively, compared to 30.42 dB SPL for male and 29.81 dB SPL for female controls. The most significant disparity is observed in the seventh decade, with male carriers obtaining a SRT of 97.05 dB SPL compared to 34.56 dB SPL in the control group. Female carriers, on the other hand, obtain a SRT of 88.93 dB SPL in comparison to 30.50 dB SPL for female control participants. To further investigate SPIQ in patients with DFNA9, Figure 2 shows the ICA score (in %) as a function of age for both male and female carriers.

Figure 2 shows a mean value of more than 90% in the age range 18–29 years for both male and female carriers. More specifically, as noted in Table 4, we measure an ICA of 96.36% (±2.51) for male carriers and 96.83% (±2.08) for female carriers. A 100% score at 70 dB SPL is obtained for all carriers in the second and third decades. In the fourth decade, the total ICA score is 93.38% (±4.11) and 90.41% (±4.79) for male and female carriers, respectively. In this decade, all scores at 55 and 70 dB SPL are larger than 95%. There is a small difference between the groups for the ICA score at 40 dB SPL, with male carriers achieving a score of 83.68% (±8.29) and female carriers achieving a score of 75.05% (±11.70). From the age of 40 years onwards, a steep decline in scores can be observed as the ICA score drops to 47.17% (±26.62) for male carriers and to 55.35% (±36.94) for female carriers. As this score is worse for male carriers compared to female carriers, the decline is even steeper for female carriers. More specifically, the deterioration in ICA scores continues, reaching 35.78% (±31.85) and 23.00% (±29.98) for male and female carriers, respectively, in the sixth decade. In this age decade, the deterioration is more pronounced in female carriers than in the previous one. Subsequently, at the age of 60 years, a score of 0% is reached for both groups at intensity levels of 40 and 55 dB SPL. Even at 70 dB SPL, all patients over the age of 60, except one female, have no speech understanding in quiet. In addition, there is a large variability between subjects, both male and female, in the fifth and sixth decade. Online supplementary Digital Content Figure 2 illustrates the longitudinal trajectories of an individual score (expressed as a percentage) at 40, 55, and 70 dB SPL as a function of age.

The natural evolution of SPIN is illustrated in Figure 3. A normative SRT value can be achieved up to the age of 35 years. Table 5 displays the specific values where the mean SRT is −5.86 dB SNR (±0.99) for male carriers and −4.84 dB SNR (±2.16) for female carriers in the second/third decade of life. An SNR value of 0 dB is obtained between the ages of 40 and 45 years. From this age onward, there is a steep increase in thresholds that results in SRTs of 4.96 dB SNR (±7.84) and 4.03 dB SNR (±9.88) in the fifth decade, and 8.83 dB SNR (±8.34) and 12.67 dB SNR (±8.52) in the sixth decade for male and female carriers, respectively. This will eventually result in an SRTs larger than 20 dB at the age of 65. Male carriers in their seventh decade have a mean SRT of 19.50 dB SNR (±1.66), whereas female carriers have an SRT of 18.31 dB SNR (±2.78), on average. However, there is large variability among patients, particularly after the age of 45. For instance, at the age of 50, some patients still demonstrate SRTs with a negative SNR, whereas others have SRTs greater than 20 dB SNR. Furthermore, the differences in SPIN between male and female carriers appear to be rather limited.

Regarding the cross-sectional comparison, the SPIN results of the matched group of normal hearing individuals are added to Table 5. This table illustrates that the SRTs for male control participants changes from −5.50 to −3.28 dB SNR, and for females from −6.43 to −4.58 dB SNR, from the second/third to the seventh decade. This contrasts with the DFNA9 population, as the discrepancy between the two groups is already present as soon as the second/third decade. In this age-group, female carriers achieve a SRT of −4.84 dB SNR compared to −6.43 dB SNR for the control group. For male carriers, a discernible difference becomes evident starting from the fourth decade as the average SRT equals −5.18 dB SNR compared to −6.06 dB SNR for the control group. Furthermore, a large discrepancy between male and female carriers can be observed in the fifth decade, with SRTs of 4.96 dB SNR and 4.03 dB SNR, respectively, compared to the control group’s SRTs of −6.52 dB SNR and −5.61 dB SNR. This difference further increases in the sixth and seventh decades as male and female carriers in their sixth decade obtain scores of 8.83 dB SNR and 12.67 dB SNR, on average, whereas scores of 19.50 dB SNR and 18.31 dB SNR are recorded in their seventh decade.

To examine the correlation between SPIQ and pure-tone audiometry, a graph is plotted that displays SRT values as a function of the PTA of 1, 2, and 4 kHz (Fig. 4). This graph indicates that those with a PTA under 20 dB HL have an average SRT of 30 dB SPL. As hearing function declines, a decrease in speech discrimination is noticeable. In addition, the graph shows that a hearing loss of 40 dB HL corresponds to an SRT value of 50 dB SPL, while a hearing loss of 70 dB HL corresponds to an SRT value of 90 dB SPL. The correlation between SPIQ and pure-tone audiometry is similar in male and female carriers. The correlation between the FI and mean SRT is 0.98 for male participants in year 0, while the corresponding correlation for female participants is 0.96. For males, the correlation in the subsequent years (Y1, Y2, and Y3) stands at 0.98, 0.97, and 0.98, respectively. Conversely, females achieved correlations of 0.96, 0.97, and 0.97 in years 1, 2, and 3.

Figure 5 shows the correlation between the SRTs in noise and the high Fletcher Index (mean of 1, 2, and 4 kHz). A normal hearing function (Fletcher index <20 dB HL) corresponds with a SPIN SRTs of −5 dB SNR. When hearing function degrades from normal to a mild hearing loss of 40 dB HL, the average SRT ranges between 0 and 5 dB SNR. If an individual has a moderately severe hearing loss of 70 dB HL, their SPIN drops to a level between 15 and 20 dB SNR. Once the FI level reaches 75 dB HL, it is not possible to establish the speech perception values. The correlation between the FI and mean SPIN for male participants (in year 0) is 0.96, as is the case for female participants (in year 0). For males, the correlation in the subsequent years (Y1, Y2, and Y3) is 0.94, 0.96, and 0.97, respectively. Meanwhile, females obtained correlations of 0.96, 0.95, and 0.96 in years 1, 2, and 3, respectively.

The present study explored the natural evolution over time of speech perception in both quiet and background noise among 101 carriers of the P51S mutation. To the best of our knowledge, our study comprises the largest database of DFNA9 patients in the literature, providing valuable and instructive insights. First, our data show near normal SRTs in quiet in the second and third decades of life. With increasing age, a steep decrease was noticed, and no speech discrimination ability in quiet remained for carriers in their seventh decade. Differences between carriers and control participants seem evident in the third decade of life and become increasingly pronounced in the decades that follow. As such, in the sixth and seventh decade, the differences evolve to more than 30 dB and more than 55 dB respectively. Second, for patients with DFNA9, SRTs in noise demonstrated a similar trend, varying from −5 dB SNR in the youngest age-group (18–29 years), to no speech-in-noise thresholds in patients above the age of 60 years. The discrepancy between the DFNA9 and control group is already present as soon as the second/third decade. From this age-group onwards, the difference increases steadily and becomes more and more pronounced. In the fifth decade, the difference between the two groups equals −11.48 dB for male carriers and −9.64 dB for female carriers. In the sixth and seventh decade, the discrepancy increases to more than 13 dB and more than 20 dB, respectively.

Despite substantial research on hearing function in carriers of the p.Pro51Ser mutation in the COCH gene, limited attention has been given to their SPIQ and noise. Only two cross-sectional phenotype studies have been carried out on this topic. Bischoff et al. [28] first analyzed SPIQ in 74 carriers of the p.Pro51Ser variant. Speech recognition scores were studied as a function of age, and pure-tone thresholds. In this study, nonlinear regression analysis was used to fit sigmoidal dose-response curves with a variable slope. The results indicated that the onset age for 90% speech understanding was 49 years and the recognition score deteriorated by an average of 2.3% per year. Eventually, this decline led to a speech understanding of 50% when the patients reached 66 years of age. Similar findings were reported by Bom et al. [29] (2001) regarding speech recognition scores in DFNA9 patients, with speech recognition score of 90% observed at age 43 which declined to 50% at age 65. Moreover, DFNA9 patients experienced a yearly average deterioration rate of speech perception of 1.8%.

For future research on the natural evolution of hearing impairment, it will be essential to not only investigate pure-tone audiometry, but also SPIQ and noise. As HLs worsen progressively, patients with DFNA9 will require hearing aids to amplify sounds and improve speech perception due to the progressive nature of hearing loss. Current (partial) reimbursement criteria for hearing aids are most frequently based on PTA, i.e., averaged hearing thresholds should be equal to or larger than 40 dB HL. However, our research indicates poor performance on speech understanding in noise, even at a young age. Therefore, reimbursement for hearing aids based on a speech-in-noise tests could be more beneficial. For instance, in Belgium, the National Institute for Health and Disability Insurance (RIZIV) enables (partial) reimbursement until the age of 65 years if a patient scores 3 dB worse than the norm on a speech-in-noise test. The speech-in-noise test measures the SNR for 50% score, where speech and noise are presented to the same ear at a noise level of 60 dB SPL. The normative SRT for LIST sentences in noise is −7.6 dB SNR. A patient below 65 years of age can receive (partial) reimbursement for a hearing aid only if his or her SRT for LIST sentences is −4.6 dB SNR or poorer (less negative). Our data show that the youngest DFNA9 patient eligible for a right ear hearing aid was only 18 years old (SRT in noise = −3.67 dB SNR), the second youngest 20 (SRT = −4.33 dB SNR). The youngest person with DFNA9 eligible on the left ear is 20 years old, with a score of −4.33 dB SNR. However, based on PTA, these 2 patients would be classified as normal hearing. Furthermore, the results demonstrate that almost 60% of patients who fail to meet the eligibility criteria for hearing aids based on pure-tone thresholds (PTA <40 dB HL) would be eligible for hearing aids if reimbursement was based on SPIN.

Overall, analyzing the correlation between SPIN and pure-tone audiometry indicates that a high FI of 40 dB HL corresponds to a SPIN between 0 and +5 dB SNR. Consequently, research indicates that SPIN outcomes offer greater practicality in everyday situations compared to pure-tone audiometry, as listening to speech often occurs in background noise [38]. Moreover, it is crucial to acknowledge that also extended high-frequency (EFH) hearing loss may amplify challenges in comprehending speech amidst noisy surroundings. Given that our current clinical data only extends up to 8 kHz, future studies might explore the evaluation of EHF hearing. Integrating EHF assessment into routine clinical practice could enhance our understanding and management of hearing impairments within our population [39]. Additionally, we have to bear in mind that hearing aids will not halt the progression to the level of severe-to-profound hearing loss. At this stage, the amplification with the hearing aids is not resulting in adequate speech perception, and patients become cochlear implant candidates to directly stimulate spiral ganglion (SG) neurons in the cochlea [40].

Furthermore, the current study sheds new light on the so-called presymptomatic stage of the disease, i.e., the phase without hearing loss identified through pure-tone audiometry as the golden standard measurement. As demonstrated, even young patients diagnosed with DFNA9 show deviant SPIQ and SPIN scores. More specific, if patients have a normal hearing function (FI <25 dB HL), the speech perception scores in quiet will already develop to 40 dB SPL, and in noise, to −5 dB SNR. A potential explanation for the discrepancy between speech perception and pure-tone audiometry can be attributed to the degeneration of SG neurons, as evidenced by post-mortem evaluation of temporal bone samples of an individual from a large Dutch kindred segregating the P51S mutation [41], although this requires more future research. According to Keithley et al. [42], a loss of neurons leads to insufficient stimulus coding information, transmitted to the central nervous system, which can make it difficult to discriminate sounds, such as speech, particularly in noisy environments. In addition, it is widely recognized that understanding speech in the presence of background noise depends on high-level auditory processing. To further investigate the central role of speech-in-noise processes and cognitive overload, it may be useful to measure cortical auditory evoked potentials. Cortical auditory evoked potentials provide a precise and objective evaluation of these higher order auditory processing functions [43, 44]. This potential exhibits the presence of components P1, N1, P2, N2, and P3 (also known as P300), between 50 and 400 ms after an auditory stimulus. In particular, P1 and N1 components are recognized as sensory responses, indicating cortical activity in response to sound stimuli. More specific, the auditory brainstem response and envelope following response have shown potential to define biomarkers for non-invasive diagnosis of cochlear synaptopathy (CS) [45]. This is relevant as CS is defined by the loss of connection between the cochlear nerve fibers and inner hair cells located in the SG [46]. As indicated in the literature, the main clinical manifestations of CS are the difficulty in understanding speech in noisy environments in individuals with normal hearing. Recent studies in animal models suggest that a more insidious process is taking place, which permanently interrupts synaptic communication between sensory inner hair cells and subsets of cochlear nerve fibers, even before hearing loss can be observed [47].

Consequently, it is imperative to consider alternative factors that could potentially account for the observed deficits in speech understanding in noise despite clinically normal hearing thresholds. Existing literature offers several hypotheses to elucidate these phenomena. Notably, studies by Hunter et al. [48] and Polspoel et al. [49] suggest that hearing loss in EHFs may contribute to these deficits, as mentioned above. This implies that traditional audiometric assessments, which primarily focus on frequencies within the conventional hearing range, might overlook impairments in higher frequency regions critical for speech perception in noisy environments. Furthermore, the concept of auditory processing disorder, as explored by Ahmmed Ansar et al. [50], emerges as another potential explanation. Auditory processing disorder involves difficulties in the processing of auditory information by the central auditory nervous system, leading to challenges in understanding speech, particularly in challenging listening conditions like noise. Moreover, Saunders and Haggard [51] propose the notion of obscure auditory dysfunction as a plausible factor. This encompasses a range of subtle auditory abnormalities that may not be detected through routine audiological assessments but can significantly impact an individual’s ability to comprehend speech, especially in adverse listening situations.

As we acknowledge, the limitation in our retrospective study is the lack of a matched control group with similar age and similar hearing thresholds. However, it is important to note that achieving a control group with both similar hearing thresholds and age presents significant challenges. As DFNA9 patients exhibit significantly greater hearing loss compared to noninfected peers it is doubtful a control group can be matched according to hearing loss AND age as the amount of hearing loss would probably imply the presence of an underlying (other) cause of hearing degeneration and therefore could not be considered as a control (normal) group. If such a group were attainable, other underlying conditions would likely confound the results, rendering them less applicable to healthy controls. In addition, by opting to match solely based on hearing thresholds and not age, we risk overlooking the impact of age on speech comprehension in both quiet and noisy environments [52].

In conclusion, this study indicates that the impact of hearing loss emerges earlier among patients with DFNA9 based on SPIQ and noise, in comparison of a normal hearing group. Nonetheless, further research is imperative since the natural evolution of hearing function in DFNA9 has not been extensively studied. Obtaining additional data over a more extended follow-up period will enable us to generate long-term predictions and aid in the early identification of speech understanding deficits in presymptomatic carriers. The ultimate objective was to determine a therapeutic window that will help identify the onset of pathophysiology. In this regard, gene therapeutic approaches have the potential to prevent, stabilize, or slow down the process of hearing loss. To this end, the rational design of a genomically humanized mouse model for dominantly inherited hearing loss, DFNA9, is crucial. This model allows for the precise study of the disease’s progression and the testing of gene therapy strategies, advancing our understanding, and treatment of DFNA9 [53].

Written informed consent was obtained from the parent/legal guardian of participants prior to the study. The research adhered to the Declaration of Helsinki (1996) and was approved by the Local Ethics Committee of the Antwerp University Hospital (B300202042807).

The authors have no conflicts of interest to declare.

This work was supported by a FWO Grant No. 1901023N (special PhD).

V.V.R. and J.M.: conceptualization. J.M.: data curation, methodology, project administration, writing – original draft, and writing – review and editing. E.F.: formal analysis and software. V.V.R. and A.G.: supervision. J.M., H.G., G.M., M.L., O.V., S.J.V., N.V., S.D., R.B., R.P., E.F., V.V.R., and A.G. critically revised the article for important intellectual content. J.M., H.G., G.M., M.L., O.V., S.V., N.V., S.D., R.B., R.P., E.F., V.V.R., and A.G. approved the final version of the manuscript.

Vincent Van Rompaey and Annick Gilles contributed equally to this work.

The data supporting the findings of this study are not publicly available due to information that could compromise the privacy of research participants but are available from corresponding author J.M. upon reasonable request.