Prevalence of Single-Sided Deafness in the United States

Abstract

Objective

The aim of this study was to obtain a reliable estimate of SSD prevalence in the adult United States population.

Methods

A cross-sectional national epidemiologic study was performed. Participants were included from the National Health and Nutrition Examination Survey (NHANES). Each cohort includes a nationally representative sample of approximately 5,000 noninstitutionalized civilians. Subjects 20 years-old and over with audiometric testing were included. SSD was defined as normal hearing (pure-tone average [PTA] of ≤25 dB) in one ear and severe or worse hearing (PTA > 70 dB) in the other, using a both a three and four-frequency PTA definition. Prevalence was measured as a raw number (n) and percentage (%) of the sample. Weighted estimates of prevalence were calculated based on the 2019 US population census.

Results

An estimated 345,064 Americans (estimated prevalence of 0.14%, 95% CI = 0.08–0.24) had SSD. SSD was more prevalent in individuals 60 to 79 years of age (estimated 155,917 US adults, prevalence of 0.25%). A higher prevalence of SSD was noted among women compared to men (215,430 US adult women, prevalence of 0.17% versus 131,726 US adult men, prevalence of 0.11%). Using a three-frequency PTA definition resulted in an estimated prevalence of 0.11%. Finally, 27% of adults with SSD reported having “good” or “excellent” hearing despite their hearing loss.

Conclusions

The prevalence of SSD in the United States is estimated at 0.11%–0.14% (271,122 to 345.064 adults), depending on PTA definition used. These individuals could potentially benefit from auditory rehabilitation, including cochlear implantation.

INTRODUCTION

Single-sided deafness (SSD) is defined as having normal hearing in one ear and severe hearing loss (or worse) in the other ear. Among adults, the most common cause of acquired SSD is idiopathic, with other potential causes being cholesteatoma and infections, tumors of the cerebellopontine angle, and more rarely head trauma, autoimmune disorders and Meniere’s disease¹. Current treatment options available for SSD include a contralateral routing of signal hearing aid (CROS-HA), a bone conduction device (BCD) and a cochlear implant (CI). A cochlear implant is the only intervention that may potentially restore bilateral or even binaural hearing, with the possible benefits of better speech understanding in quiet and in noise, improved sound localization and improved quality of life^2–9.

It is important to understand the incidence and prevalence of SSD to guide clinical evaluation and treatment, patient counseling, and efforts towards research. Several reports state that an incidence of 12 to 27 individuals per 100,000 adult population per year has been estimated in the UK, referencing a 2006 article by Baguley et al.¹⁰. Other reports extrapolate the same number to a lifetime prevalence of about 1%, based on the well documented fact that life expectancy is about 81 years in the UK. Indeed, multiplying the 12 to 27 per 100,000 figure by the number of years the average UK citizen lives as an adult (81 minus 18), the resulting lifetime prevalence of SSD would be 0.75%–1.7%.

Because this number is surprisingly high, we investigated the basis of this claim. Despite the frequent citations to this effect, Baguley et al. never claimed that 12 to 27 per 100,000 adults suffered from SSD. They simply indicated that the three most common conditions that may result in unilateral hearing loss are “vestibular schwannoma (VS) and associated surgery, Meniere’s disease, and sudden sensorineural hearing loss”. Then they provided two separate estimates of VS incidence (2 and 1.04 per 100,000 adults per year), one estimate of Meniere’s disease incidence (4.3 per 100,000) and one estimate of sudden unilateral sensorineural hearing loss (5–20 per 100,000). Adding these four numbers (mistakenly including the two separate estimates of VS incidence rather than their average) results in the incidence estimate of 12.34–27.34 per 100,000, or 12–27 per 100,000 in round numbers, that has been widely cited. This is problematic for several reasons. First, it is not true that all VS or Meniere’s patients have SSD. Second, even those who do have unilateral hearing loss may not necessarily have severe or profound hearing loss. Third, and most importantly, the incidence estimate for the larger group in the overall estimate (sudden unilateral sensorineural hearing loss) turned out to be groundless.

Baguley et al.’s estimate of 5–20 cases of sudden sensorineural hearing loss per 100,000 adults per year is based on a 1984 publication by Byl, who actually put forth those numbers as an estimate of all such cases, not just the unilateral ones¹¹. Thus, even if Byl’s estimate was precise (which, as we shall see, is not) the yearly incidence of unilateral hearing loss, let alone SSD, would be an undetermined small fraction of the estimate. Byl’s estimate is based on three sources of data: “Jaffe estimated the incidence to be 5 per 100,000 per year in the United States¹². Van Caneghem estimated the incidence in Belgium to be about 10 per 100,000 per year¹³. Simmons presumed that the true incidence is probably higher because those who experience SHL and recover quickly do not seek medical care¹⁴.” It seems that Byl took the Jaffe estimate as the minimum and multiplied the Van Caneghem estimate by a factor of 2 to account for Simmons’ presumption.

Unfortunately, Byl’s two incidence estimates are completely arbitrary at best. Jaffe’s estimate is based on a “personal communication” made to Byl in 1974, it does not seem to be based on any published study. The Van Caneghem reference, a monograph on sudden deafness published in 1958, simply does not provide any estimates of SSD incidence. Dr. Van Caneghem reports that “after becoming interested in the matter of sudden deafness”, he “had the opportunity to examine 64 patients affected by labyrinthine deafness, who said they became suddenly deaf on one or both sides”¹³. He then calculated what fraction of the population of the West Flanders province these patients represented, by age group. The fractions ranged from 4.7 per 100,000 for 20–30-year-olds to 15.8 per 100,000 for 50–60-year-olds. A rough average of 10 per 100,000 was likely picked up by Byl¹¹, propagated to Baguley et al.¹⁰, and then to a large number of contemporary references. Of course, this number cannot be reasonably considered as an annual incidence estimate of SSD for several reasons. First, these were patients seen by Dr. Van Caneghem over the course of many years rather than a single year. Second, we have no way to ascertain that the patients seen by Dr. Van Caneghem were all the relevant cases in the province rather than a subset. Finally, there is no way to determine how many of the 64 patients had SSD or even asymmetric hearing loss.

In conclusion, it is fair to say that the incidence or prevalence of SSD in adults is “not well established”¹⁵. To the best of our knowledge, the only information available in the adult population is the prevalence of unilateral hearing loss, estimated at 7.2%¹⁶. This encompasses all severities of unilateral loss and is not specific to SSD. Therefore, the primary aim of the current study was to obtain a reliable estimate of SSD prevalence in the adult population using a precise definition and a large epidemiologic cohort. The secondary objective was to stratify the prevalence estimate by key demographic variables, including age, sex, and race/ethnicity.

MATERIALS AND METHODS

Study Participants

Participants were included from the National Health and Nutrition Examination Survey (NHANES). NHANES is a cross-sectional study conducted every two years that samples children and adults from across the United States. Each year, the study includes a nationally representative sample of approximately 5,000 noninstitutionalized civilian residents¹⁷. In addition to interviews and physical examinations, participants in certain cycles undergo audiometry and a hearing evaluation. The sampling includes noninstitutionalized civilian US residents, with oversampling of certain groups based on race/ethnicity and socioeconomic status. Deidentified data is made publicly available on the CDC website.

Participants were included in this study if they were ≥20 years old and had undergone audiometry during any of the NHANES cycles in the past 22 years. This resulted in the inclusion of participants from the 1999–2000, 2001–02, 2003–04, 2005–06, 2009–10, 2011–12, 2015–16 cycles (n=15,035).

Demographics

Demographic information was collected for each participant, including age, sex, and race/ethnicity. Age was categorized in approximate twenty-year intervals: 20–39, 40–59, 60–79, and ≥80 years old. Sex was categorized by NHANES as male or female. Race/ethnicity was categorized by NHANES as Mexican American, other Hispanic, Black non-Hispanic, White non-Hispanic, or Other/Multiple (which includes Asian non-Hispanic individuals).

Hearing

NHANES performs pure-tone air-conduction audiometry across octave frequencies 500 to 4,000 Hz in sound-isolated rooms by trained examiners. Pure-tone averages (PTA) were calculated for each ear using the mean dB hearing threshold at 500, 1,000, 2,000, and 4,000 Hz. Pure-tone averages in each ear were categorized by severity: mild (PTA 26–40 dB), moderate (41–55 dB), moderately severe (56–70 dB), severe (71–90 dB), or profound (≥90 dB). Because only 1 participant had profound unilateral hearing loss, the severe and profound categories were combined (>70 dB) for all analyses. SSD was defined as normal hearing (PTA of ≤25 dB) in one ear and severe or worse hearing loss (PTA >70 dB) in the other ear. Unilateral hearing loss was defined as normal hearing (PTA of ≤25 dB) in one ear and mild or greater loss (PTA >25 dB) in the other ear. We also performed two sensitivity analyses using modified definitions of the hearing variable. The first used a 3-frequency definition for PTA (500, 1,000, 2,000 Hz) to define normal hearing and unilateral hearing loss. The second sensitivity analysis defined normal hearing as hearing thresholds ≤25 dB for all individual frequencies, rather than using the PTA ≤ 25 dB cutoff. A final analysis was performed using the criteria from the Global Burden of Disease Expert Group on Hearing Loss^{18, 19}. In this severity classification system, normal hearing is defined as ≤19.9 dB hearing level, mild hearing loss is 20.0–34.9 dB hearing level, moderate is 35.0–49.9 dB hearing level, moderately severe is 50.0–64.9 dB hearing level, severe is 65.0–79.9 dB hearing level, profound is 80.0–94.9 dB hearing level, complete or total hearing loss is ≥95.0 dB hearing level.

Additionally, participants were excluded if they had tympanometric or otoscopic abnormalities suggestive of conductive hearing loss (n=467), as opposed to sensorineural loss. These abnormalities included a flat (compliance < 0.2) or negative peak (< −150) on tympanometry, cerumen impaction, the presence of a tympanostomy tube, or any other otoscopic abnormalities. We acknowledge that this exclusion criteria serves as an imperfect proxy for a diagnosis of conductive hearing loss, as the type of hearing loss was not formally evaluated in NHANES.

Statistical Analysis

Prevalence of hearing loss was first measured as a raw number (n) and percentage (%) of the sample from NHANES. Weighted estimates of US population prevalence were then calculated based on the 2019 US population census for adults ≥20 years old. Weighting was made possible due to the unique sampling design of NHANES²⁰. Data analysis was performed using Stata 16.1 (StataCorp LLC, College Station, TX). The svy:prop function was used to calculate point estimates and 95% confidence intervals. This study was exempt from review from our institutional review board as it uses de-identified, publicly available data.

RESULTS

The overall prevalence of unilateral hearing loss among US adult civilians was 5.20% (95% confidence interval [CI] = 4.72–5.72) (Table 1). This represents 12,816,676 individuals based on 2019 US census counts. The prevalence of severe/profound unilateral hearing loss, or SSD, was 0.14% (95% CI = 0.08–0.24), equating to 345,064 individuals.

Table 1.

Prevalence of all forms of unilateral hearing loss

Type of hearing loss	Number of US adults with unilateral hearing loss	Prevalence of unilateral hearing loss in US population (%)	95% Confidence interval
Mild	10,992,764	4.46	4.03–4.93
Moderate	1,158,430	0.47	0.34–0.65
Moderately severe	320,417	0.13	0.08–0.23
Severe and profound	345,064	0.14	0.08–0.24
Total	12,816,676	5.20	4.72–5.72

Using a 3-frequency PTA definition (500, 1000, 2000 Hz), the prevalence of unilateral hearing loss and SSD was expectedly lower at 3.95% (95% CI = 3.56–4.38) and 0.11% (95% CI = 0.06–0.20), respectively (Supplementary Table 1). When normal hearing was even more strictly defined as ≤25 dB for all hearing thresholds, the prevalence of unilateral hearing loss and SSD was 0.72% (95% CI = 0.56–0.92) and 0.04% (95% CI = 0.02–0.10), respectively (Supplementary Table 2). When using the Global Burden of Disease Expert Group on Hearing Loss criteria, the prevalence of SSD was slightly higher, at 0.17% (Supplementary Table 3).

Older individuals who were 60–79 years of age had a higher prevalence of SSD (0.25%) compared to younger individuals (0.11% in ages 20–39 and 0.11% in ages 40–59). Of note, SSD was not reported among those 80 years of age and older (Table 2). Thus, we are unable to calculate SSD prevalence in this age group.

Table 2.

Prevalence of single-sided deafness (SSD) by age, sex and ethnicity

Description	Number of US adults with SSD	Prevalence of SSD within each age category (%)	95% Confidence interval
Age group
20–39	97,762	0.11	0.05–0.28
40–59	90,798	0.11	0.04–0.29
60–79	155,917	0.25	0.09–0.70
Sex
Male	131,726	0.11	0.05–0.22
Female	215,430	0.17	0.08–0.35
Ethnicity
Mexican American	25,995	0.11	0.03–0.36
Other Hispanic	25,852	0.16	0.04–0.67
White (Non-Hispanic)	233,596	0.15	0.07–0.30
Black (Non-Hispanic)	17,782	0.06	0.01–0.26
Other/multi	38,371	0.18	0.06–0.53

^*No NHANES participants >=80 years old had SSD.

Females had a higher prevalence of SSD (0.17%, 95 CI = 0.08–0.35, 215,430 individuals) than males (0.11%, 95% CI = 0.05–0.22, 131,726 individuals) (Table 2). However, given the relatively wide confidence intervals, this difference may be due to sampling error. Stratifying by ethnicity demonstrated the highest prevalence of SSD among non-Mexican Hispanic adults (0.16%) and Other/Multi ethnic category (0.18%). However, looking at the absolute number of US adults affected, most individuals were non-Hispanic white adults (233,596) or within the Other/Multi ethnic category (38,371) (Table 2).

Subjective hearing level was assessed among those with SSD. Approximately 27% of individuals with SSD reported “excellent” or “good” hearing, 38% reported “a little trouble”, and 35% reported “moderate trouble” or worse (Table 3). No individual with SSD reported wearing a hearing aid or a history of wearing one.

Table 3.

Subjective hearing level according to individuals with single-sided deafness (SSD)

Subjective hearing level	Number of US adults with SSD	Prevalence of subjective hearing category among those with SSD (%)	95% Confidence interval
Excellent/good	92,408	26.8	0.03–99.8
A little trouble	131,193	38.0	0.05–99.9
Moderate trouble or worse	121,497	35.2	0.02–99.9

DISCUSSION

In the current study, we used NHANES to estimate the number of individuals affected by SSD in the United States. We found that the prevalence of SSD among a group of adult civilians in the United States was about 0.14% (95% CI 0.08–0.24), for an estimated 345,064 individuals. This represents a small proportion of the estimated 13 million Americans with unilateral hearing impairment, and it is almost an order of magnitude less that the estimates cited in current literature.

The prevalence of SSD has not been clearly established in the current literature. This may be due to the lack of a consistent definition and confusion about nomenclature, with single-sided deafness sometimes being interchanged with the terms “asymmetric hearing loss” and “unilateral hearing loss”. We used a clear definition in this study based on pure-tone averages (PTA). Although we could not definitively distinguish between conductive and sensorineural hearing loss with the NHANES data, we used strict criteria and excluded any participant who could potentially have conductive hearing loss based on abnormal tympanograms, cerumen impaction, the presence of a tympanostomy tube, or any other otoscopic abnormalities. This improved our chances of capturing individuals with sensorineural hearing loss only. In addition, we conducted a sensitivity analyses, changing the definition of SSD from a 4-frequency PTA to a stricter 3-frequency PTA had an impact, which reduced the estimated prevalence from 0.14% to 0.11%. To allow comparability between studies, SSD should be clearly defined when performing any type of research.

Demographic differences in prevalence

The prevalence of SSD was higher in individuals between 60 and 79 years of age, although with a wide confidence interval it is difficult to make definitive conclusions. There were no individuals in the included NHANES cycles over the age of 80 with SSD. One explanation is that only two NHANES cycles performed audiograms in this age group. Second, it is likely that most subjects over the age of 80 will not meet criteria for SSD, as their other ear is likely to have hearing loss greater than 25 dB due to bilateral age-related hearing loss.

We found a higher prevalence of SSD in females. Although males in the cohort had a higher prevalence of unilateral hearing loss overall, when broken down by severity, the prevalence of SSD in women was nearly twice as high, although once again the confidence interval was wide. Over 50% of cases of adult SSD are idiopathic, with other causes including tumors of the cerebellopontine angle, cholesteatoma, Meniere’s disease, and trauma¹. Among all the disease entities listed, none are known to have a female predominance²¹.

The prevalence of SSD was comparable between ethnicities. Non-Mexican Hispanic adults had a higher prevalence of SSD within their ethnic category than non-Hispanic white adults. Interestingly, previous evidence suggests that unilateral hearing loss in general may be more prevalent among non-Hispanic white individuals¹⁶. It is possible that our results reflect a higher prevalence of more severe unilateral hearing loss among United States ethnic minorities.

Potential for intervention

Although the prevalence of SSD is overall low at 0.14%, it should be noted that an additional 0.13% of participants in this cohort had unilateral moderately severe hearing loss. In total, this study identifies over 650,000 individuals who could potentially benefit from hearing rehabilitation, including a cochlear implant, for SSD or for unilateral moderately severe hearing loss.

The ability to hear with both ears has several advantages. Individuals with single-sided deafness experience difficulties hearing in background noise, issues with localizing sound and often report significant tinnitus. Furthermore, Wie et al. examined the social impact of SSD on adults and found that 93% of affected individuals believed that their hearing loss had a negative effect on interactions with others⁶. Rehabilitation options include CROS aids, bone-anchored devices and cochlear implants. Cochlear implantation is the only option that can potentially restore elements of binaural hearing. Patients who undergo cochlear implantation for SSD report better speech understanding in quiet and in noise, improved localization of sound and even improved quality of life^{2, 6, 9, 22, 23}. For this reason, knowing the prevalence of individuals with SSD who could potentially benefit from treatment is important.

Finally, approximately 27% of participants with SSD subjectively rated their hearing as “excellent” or “good”. Indeed, a certain group of individuals with SSD chose not to undergo any form of rehabilitation. The implications of SSD in adults are less studied than those in children. Among children, all forms of unilateral hearing loss have been associated with worse performance in school²⁴. Among adults, the long-term effects are less well understood. While bilateral age-related hearing loss has been linked to cognitive impairment and incident dementia in previous studies^{25, 26}, to our knowledge, there has not been any reported association between SSD and these conditions. Nonetheless, it has been described that adults with SSD experience reduced well-being in social settings, may avoid social gatherings in certain conditions and overall have difficulties in multi-speaker environments⁶. This highlights the importance of diagnosing SSD and considering intervention.

Limitations

There are limitations to the current study. The first is our use of a dataset from the United States, thus we may not be presenting a globally representative estimate. Furthermore, the most recent NHANES cycle available is six years old, although this is unlikely to strongly impact results given hearing loss prevalence is not expected to change quickly over time. NHANES data only comprises air-conduction hearing thresholds, which makes it impossible to distinguish sensorineural versus conductive hearing loss. Given that several forms of conductive hearing loss may be transient (e.g., middle ear effusion, cerumen impaction), we attempted to exclude those at risk of conductive hearing loss based on tympanometric and otoscopic findings.

CONCLUSIONS

The estimated prevalence of SSD among a group of adult civilians in the United States was 0.14%, approximately 345,064 individuals. This is much smaller than the prevailing estimate of 0.75%–1.7% described in the introduction. Having more precise estimates of SSD prevalence may be useful to identify patients that could potentially benefit from auditory rehabilitation, including cochlear implants, and to help guide decision making in the public health system as well as the private sector.