Association of Alopecia Areata with Sensorineural Hearing Loss: A Systematic Review and Meta-analysis Open Access

Alopecia areata (AA) is a prevalent immune-mediated hair loss characterized by well-circumscribed patches of baldness on the scalp [1]. Accompanying hair loss of other body sites, for example, the eyelashes, eyebrows, axillary, and pubic hairs may be present. Nail pitting and trachyonychia may appear as well [2]. The incidence of AA is about 20.2 per 100,000 person-years with an estimated lifetime risk of 2% in the general population [3, 4]. Several etiologic factors including genetics, immunology, and oxidative stress have been proposed [5‒9]. Among these, disruption of hair follicular immune privilege is considered as the key pathogenesis with melanocyte-associated autoantigens being the major immune target [9‒11].

Melanocytes not only play a crucial role in maintaining endocochlear potential and regulating inner hair cell function but also help resist free radical damage [12‒14]. Similar to hearing loss associated with vitiligo which may be related to melanocyte destruction [15, 16], immune-mediated melanocyte-related pathogenesis in AA may cause sensorineural hearing loss (SNHL). A few studies have reported hearing loss among patients with AA [17‒20]. Therefore, we aimed to systematically examine the evidence on the association of AA with SNHL.

We conducted this systematic review and meta-analysis and followed the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) guidelines [21]. The protocol was determined a priori and has been registered with PROSPERO (CRD42021284274). The evidence search, study selection, data extraction, and risk of bias assessment were independently performed by two investigators (K.L. and T.G.), with discrepancies resolved by discussion with the supervisor (C.C.) until a consensus was reached.

We searched MEDLINE and Embase databases for relevant articles from their respective inception to July 25, 2022. As shown in the search strategy, the search terms included “alopecia areata,” “ophiasis,” “hearing loss,” “hearing,” “hypoacusis,” “anacusis,” and “audiometry.” No restrictions on language nor geographic regions were applied. The titles and abstracts were screened, and the full text of potential studies was examined for relevance. Also, the references of included studies were scanned to identify additional relevant articles. Studies were included if they met the following criteria: (1) observational studies examining the association of AA with hearing loss, including cross-sectional, case-control, or cohort studies; (2) the non-AA comparison group was composed of age-matched controls; and (3) the study subjects excluded patients with prior ototoxic drug use, otologic surgery, chronic noise exposure, otologic pathology, head trauma, or any other neurological or vascular disease. The exclusion criteria were case report, case series, or animal studies.

From included studies, we extracted the following data: first author, publication year, country, study design, number of study subjects in the case and control groups, risk estimates presenting as odds ratio (OR) with 95% confidence interval (CI), and patient characteristics including sex, age, mean duration and severity of AA, frequency-specific pure-tone thresholds (PTTs), and the number of subjects with SNHL. Additional methods for hearing evaluation such as otoacoustic emissions were not consistently reported and were therefore not included in the meta-analysis.

The risk of bias of included studies was evaluated by using the Newcastle-Ottawa Scale [22]. Specifically, in case-control studies, for the “adequacy of case definition” domain, a low risk of bias was denoted if the diagnosis of AA was confirmed by dermatologists, an unclear risk of bias if AA was identified by using relevant diagnostic codes in databases, and a high risk of bias if AA was not explicitly defined. As to the “comparability of cases and controls” domain, a low risk of bias was rated if controls were matched by both age and sex, an unclear risk of bias if controls were matched only by age, and a high risk of bias if no matching was performed.

For studies that reported respective PTT for right and left ears, we calculated the mean difference (MD) and combined standard deviation (SD) of PTT by using the formula of averaging SD [23]. For the study only reported median and range of PTTs, we used formulas to estimate the mean and SD [24]. We performed a meta-analysis to obtain the MD with 95% CI in frequency-specific PTTs between patients with AA and age-matched healthy controls. For all included studies which provided the respective number of AA and control patients, we performed a meta-analysis to obtain a pooled OR with 95% CI for SNHL in relation to AA. Moreover, we conducted a sensitivity analysis by including only case-control studies. The random-effects model was implemented due to anticipated clinical heterogeneity across included studies [25]. The statistical heterogeneity was evaluated by using the I² statistic, with an I² of >50% representing at least moderate heterogeneity [26]. We originally planned to use funnel plot to assess publication bias but did not execute because the number of included studies was <10. All statistical analyses were conducted by Review Manager 5.4 (The Cochrane Collaboration, 2020).

As shown in the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) study flowchart (see Fig. 1), our search yielded 63 articles. One additional study was identified from the references of the included studies. After deleting duplicates, the abstracts and titles of 52 articles were assessed. The full text of 11 articles was examined for eligibility, with 6 studies (including 5 case-control studies and 1 cohort study) with a total of 55,419 subjects assessing hearing loss among patients with AA included for analysis [17‒20, 27, 28].

The characteristics of included studies are summarized in Table 1. All the included studies employed sex- and age-matched controls. As presented in Figure 2, the risk of bias of included studies was generally low. The disease duration ranged from 2.9 ± 4.6 to 18.7 ± 39.5 months. The disease severity varied in the included studies with localized to extensive AA and was classified by different tools such as the Severity of the Alopecia Tool (SALT) score, SALT II, or number of AA lesions. Three case-control studies which provide odds calculation used different definitions of hearing loss [19, 27, 28]. The Ertugrul 2021 study adopted the 1964 International Organization for Standardization hearing threshold parameter (IOSH) which defined mild hearing loss as 21–40 dB and moderate hearing loss as 41–55 dB [28]. The Ucak 2014 study defined minimal hearing loss if the PTT was >30 dB [19]. Another study defined hearing loss as average of 500, 1,000, and 2,000 Hz ≥26 dB [27]. The one included cohort study using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code to identify AA and hearing loss [20].

Four case-control studies provided PTT data at specific frequencies from 250 to 8,000 Hz [17, 18, 27, 28]. Two studies provided extended high frequencies (>8,000 Hz) [17, 18]. The frequency-specific PTTs available for meta-analysis were 250, 500, 1,000, 2,000, 4,000, 8,000, and 12,000–12,500 Hz, in which 250–8,000 Hz were also the most frequently used in clinical practice.

As shown in Figure 3, the meta-analysis showed that the PTTs of the AA and control groups did not significantly differ at 250 Hz (MD: 1.23; 95% CI: −1.03–3.49; I² = 67%), at 500 Hz (MD: 1.59; 95% CI: −0.73–3.92; I² = 58%), at 1,000 Hz (MD: 1.03; 95% CI: −0.20–2.25; I² = 0%), at 2,000 Hz (MD: 1.08; 95% CI: −0.43–2.60; I² = 0%), and at 8,000 Hz (MD: 1.53; 95% CI: −0.96–4.01; I² = 0%), respectively. By contrast, the PTTs of the AA and control groups significantly differed at 4,000 Hz (MD: 2.96; 95% CI: 0.63–5.29; I² = 0%) and 12,000–12,500 Hz (MD: 10.26; 95% CI: 5.41–15.12; I² = 0%).

As shown in Figure 4, the meta-analysis showed increased odds for SNHL among patients with AA (OR: 3.18; 95% CI: 2.06–4.89; I² = 0%) [19, 20, 27, 28]. The sensitivity test by including only three case-control studies with 131 AA patients and 125 controls still found significantly increased odds of SNHL among patients with AA (OR: 2.55; 95% CI: 1.40–4.66; I² = 0%; See Fig. 5) [19, 27, 28].

To the best of our knowledge, this is the first systematic review and meta-analysis to investigate the association between AA and SNHL. We found 3.18-fold odds for SNHL among patients with AA when compared with controls. Generally, all PTTs were elevated in the AA group. The MD in PTTs ranged from 1.03 to 10.26 dB at frequencies from 250 to 12,500 Hz. Significant elevations of PTTs in AA patients were only found at higher frequencies including 4,000 Hz and 12,000–12,500 Hz. The hearing loss was most marked at 12,000–12,500 Hz. However, among 250–8,000 Hz, the range which most sounds of daily conversations occur, only 4,000 Hz showed significant difference in our meta-analysis. This result mimicked the 4,000-notch presentation in noise-induced hearing loss and could be explained by the outer hair cells area corresponding to the 4,000 Hz over the basilar membrane of the human cochlea which is the most vulnerable [29‒31]. Besides, among 250 Hz–8,000 Hz, human ear is most sensitive to sounds around 4,000 Hz due to the amplifying mechanism of the ear canal resonating anatomical shape [29].

Extended high-frequency audiometry (>8,000 Hz) is regarded as a highly sensitive screening test in young adults and helps early detect high-frequency hearing loss [32, 33]. This may explain the larger extent of hearing loss at 12,000–12,500 Hz in our meta-analysis and give clinicians clues to potential progressive SNHL among AA patients. Besides, high-frequency hearing loss has been regarded as the initial presentation of damaged hearing ability and though some without significant manifestations but some will present with tinnitus which causes the daily hassles [34‒36].

As to the disease severity and hearing loss, there were various classifications of AA severity in these studies such as SALT, SALT II, or number of AA lesions. In the Ucak 2014 study, hearing loss occurred more frequently when AA involved >25% of the scalp [19]. The Shaheen 2015 study found AA severity correlated with SNHL at 8,000 and 12,000 Hz [17]. Besides, higher recurrence of AA was associated with SNHL at 12,000 Hz. Three studies found no association between AA severity and SNHL, but the severity of their study subjects was generally low [18, 27, 28]. In the Koçak 2018 study [18], only 3.9% AA patients presented with ≥3 involved areas; in the Erdoğan 2018 study [27], most AA patients presented with a single area involvement (60.4% scalp only and 22.9% beard only); in the Ertugrul 2021 study, only 15.6% AA patients had involvement of more than one location [28]. The various classifications of AA severity and relatively low severity of most study subjects may account for the lack of associations between severity and SNHL in these three studies.

The Shaheen 2015 study found correlations between SNHL at 8,000 Hz, 12,000 Hz, and the disease duration of AA [17]. By contrast, the other 4 case-control studies found no association of the duration of AA with hearing loss. The relative short mean duration of AA in all included studies (ranged from 2.9 to 18.7 months) may limit the ability to detect the correlation between disease duration and hearing loss.

AA occurs when the immune privilege of the hair follicles breaks down and then autoantigen-dependent CD8⁺ T cells are triggered to attack hair follicles [10, 37]. Besides, predisposing factors such as environmental stress, viral infections, or drugs may produce reactive oxygen species (ROS) and lead to accumulation of natural killer cells, production of interferon and tumor necrosis factor, and major histocompatibility complex presentation [38‒40]. Furthermore, various antigens are targeted, mainly melanocyte-associated molecules of hair follicles [41]. Pro-inflammatory cytokines including interferon-γ, substance P, and tumor necrosis factor are released and the hair follicles are attacked [10, 42‒45].

Melanocytes reside in the spiral ligament and stria vascularis in the inner ear [13]. These melanocytes play a crucial role in regulating potassium channels of endocochlear potential which is essential to maintain auditory signal transmission [46, 47]. In addition, melanin helps neutralize ROS by acting as a Ca²⁺ chelator [48, 49]. Therefore, both immune-mediated inflammatory reactions and oxidative stress can directly damage melanocytes in the cochlea and lead to SNHL. Previous studies also revealed that damage to melanocyte by autoimmune diseases, such as vitiligo or Vogt-Koyanagi-Harada (VKH) disease, may deteriorate hearing status [15, 50]. The VKH disease affects tissues containing melanin and severely dysfunctions of melanocytes which causes severe anterior uveitis, alopecia, poliosis, vitiligo, and also tinnitus or even hearing loss especially at high frequencies [51, 52]. These findings are consistent to our results. Several studies already reported that VKH patients may experience SNHL and suggested and to undergo a review of systems for auditory examinations [53]. Overall, the shared pathogenesis including the ROS- and autoantigen-triggered immune/inflammatory cascade in melanocyte-associated molecules of hair follicles and inner ear cells may account for the association between AA and SNHL.

There are some limitations in this study. Some recent studies have used evoked otoacoustic emissions to evaluate hearing in AA patients which were not calculated in our study due to lack of complete and standardized data [18, 28, 54]. Second, considering the shared pathogenesis of AA and the above autoimmune diseases mentioned, the confounders may exist. Therefore, all of our included case-control studies have excluded other autoimmune diseases including vitiligo except one study which included AA patients accompanying autoimmune thyroiditis which might have been a confounder [27]. However, a sensitivity analysis excluding this study did not change the results (OR: 2.69; 95% CI: 1.28–5.65; I² = 12%). Third, although some included studies classified the AA patients into different severity groups, no consistent criteria were applied such as SALT and SALT II [17, 28]. This made it difficult to figure out the association between the degree of hearing loss and disease severity.

This study demonstrates an association between AA and SNHL especially at high frequencies. High-frequency SNHL is often neglected and may initially present with only tinnitus [34, 55]. Consultation with otolaryngologists may be needed when AA patients complain of hearing loss or tinnitus. Further investigation especially direct evidence of AA influencing inner ear is still needed to support our findings.

AA is associated with an increase of SNHL, especially at high frequencies.

No ethics approval was needed because data were retrieved from previously published studies in which informed consent was obtained by the primary investigators.

Kuang-Hsu, Lien, Tzong-Yun Ger, and Ching-Chi Chi have no conflicts of interest directly relevant to the content of this article.

No funding was received for the conduct of this study or the preparation of this article.

Conceptualization: Kuang-Hsu Lien and Ching-Chi Chi; writing – original draft preparation: Kuang-Hsu Lien.; writing – review, editing: Tzong-Yun Ger and Ching-Chi Chi.; supervision: Ching-Chi Chi. All authors have read and agreed to the published version of the manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.