Association of Hearing Loss and Otologic Outcomes With Fibrous Dysplasia

Key Points

Question  What are the potential mechanisms of hearing loss in individuals with fibrous dysplasia?

Findings  In this natural history study of 130 individuals with craniofacial fibrous dysplasia, conductive hearing loss was frequently associated with deformity of the epitympanum and rarely with external auditory canal stenosis, whereas sensorineural hearing loss was most often associated with elongation of the internal auditory canal and rarely with otic capsule involvement. Endocrine features, including growth hormone excess and neonatal hypercortisolism, were associated with hearing loss.

Meaning  Individuals with fibrous dysplasia should undergo clinical and radiologic evaluation to identify high-risk features for audio-otologic dysfunction.

Importance  Fibrous dysplasia (FD) and McCune-Albright syndrome (MAS) are rare bone and endocrine disorders in which expansile fibro-osseous lesions result in deformity, pain, and functional impairment. The effect of FD on hearing and otologic function has not been established.

Objectives  To characterize audiologic and otologic manifestations in a large cohort of individuals with FD/MAS and to investigate potential mechanisms of hearing loss.

Design, Setting, and Participants  In this natural history study, individuals with craniofacial FD seen at a clinical research center underwent clinical, biochemical, computed tomographic, audiologic, and otolaryngologic evaluations.

Main Outcomes and Measures  Clinical and radiologic features associated with hearing loss and otologic disease were evaluated. Conductive hearing loss was hypothesized to be associated with narrowing of the external auditory canal (EAC), FD involving the epitympanum, and FD crowding the ossicular chain. Sensorineural hearing loss was hypothesized to be associated with FD affecting the internal auditory canal (IAC) and otic capsule.

Results  Of the 130 study participants with craniofacial FD who were evaluated, 116 (89.2%) had FD that involved the temporal bone (median age, 19.6 years; range, 4.6-80.3 years; 64 female [55.2%]), whereas 14 (10.8%) had craniofacial FD that did not involve the temporal bone. Of the 183 ears with temporal bone FD, hearing loss was identified in 41 ears (22.4%) and was conductive in 27 (65.9%), sensorineural in 12 (29.3%), and mixed in 2 (4.9%). Hearing loss was mild and nonprogressive in most participants. Whereas EACs were narrower in ears with FD (mean difference [MD], 0.33 mm; 95% CI, 0.11-0.55 mm), this finding was associated with conductive hearing loss in only 4 participants. Fibrous dysplasia crowding of the ossicles was associated with conductive hearing loss (odds ratio [OR], 5.0; 95% CI, 2.1-11.6). The IAC length was not different between ears with and without FD (MD, −0.37; 95% CI, −0.95 to 0.211); however, canals were elongated in ears with sensorineural hearing loss (MD, −1.33; 95% CI, −2.60 to −0.07). Otic capsule involvement was noted in only 4 participants, 2 of whom had sensorineural hearing loss. Both MAS-associated growth hormone excess (OR, 3.1; 95% CI, 1.3-7.5) and neonatal hypercortisolism (OR, 11; 95% CI, 2.5-55) were associated with an increased risk of hearing loss .

Conclusions and Relevance  Hearing loss in craniofacial FD is common and mild to moderate in most individuals. It typically arises from FD crowding of the ossicular chain and elongation of the IAC, whereas EAC stenosis and otic capsule invasion are less common causes. Individuals with craniofacial FD should undergo otolaryngologic evaluation and monitoring, including assessment to identify those with high-risk features.

Fibrous dysplasia (FD) is an uncommon skeletal disorder in which normal bone and marrow are replaced with fibro-osseous tissue.¹ It arises from somatic mutations in GNAS (OMIM 139320) leading to constitutive activation of G_s G-coupled protein receptor signaling.² In the skeleton, these mutations impair differentiation of bone marrow stromal cells, resulting in cellular proliferation and formation of abnormal bone prone to expansion, deformity, and fracture.^1,3 Disease may occur in one bone (monostotic) or multiple bones (polyostotic) and may develop in isolation or in combination with café au lait skin macules and hyperfunctioning endocrinopathies, which include precocious puberty, hyperthyroidism, growth hormone excess, hypophosphatemia, and neonatal hypercortisolism.⁴ The association of FD with 1 or more of these extraskeletal features is termed McCune-Albright syndrome (MAS).³

The temporal bone is frequently affected by FD and has been associated with a variety of otologic and audiologic conditions, including hearing loss, pain, auditory canal stenosis, and cholesteotoma.^5-7 However, the prevalence, spectrum, and natural history of ear-related disease have not been well characterized, and the mechanisms of hearing loss have not been established. A transgenic FD mouse model demonstrated severe and progressive hearing loss attributable to bony overgrowth around the ossicles and otic capsule⁸; however, it is not known whether this model replicates human disease. The purposes of this study are to characterize audiologic and otologic manifestations in a large cohort of individuals with FD/MAS and to investigate potential mechanisms of hearing loss.

Individuals with FD/MAS were evaluated as part of a long-standing natural history study at the National Institutes of Health.⁹ All participants underwent evaluation at the National Institutes of Health Clinical Center, including history and physical examination, biochemical testing, skeletal imaging, and medical treatment for MAS-associated endocrinopathies. Participants were diagnosed with FD/MAS based on previously established clinical guidelines.³ The protocol was approved by the institutional review board of the National Institute of Dental and Craniofacial Research, and all participants and/or their guardians gave written informed consent or assent.

Participants with craniofacial FD underwent comprehensive otolaryngologic and audiologic evaluation. Standard audiometric measures, including air- and bone-conduction pure-tone thresholds for 250 to 8000 Hz and 226-Hz tympanometry, were conducted. Clinically significant hearing loss was determined using established definitions¹⁰ by a 4-frequency pure-tone average (0.5/1/2/4 kHz) greater than 20 dB hearing level (HL), and degree of hearing loss was further categorized as mild (21-40 dB HL), moderate (41-70 dB HL), severe (71-90 dB HL), and profound (>90 dB HL). Type of hearing loss was determined using a 3-frequency pure-tone average (0.5/1/2 kHz) and was classified as conductive (difference between 3-frequency pure-tone average by air and bone conduction >10 dB and normal hearing for bone conduction and hearing loss by air conduction), sensorineural (difference between 3-frequency pure-tone average by air and bone conduction <10 dB and hearing loss by air and bone conduction), or mixed (difference between 3-frequency pure-tone average by air and bone conduction >10 dB and hearing loss by both air and bone conduction). In addition, ears with normal hearing by air conduction were classified as subclinical conductive when there was a mean air bone gap greater than 10 dB. For participants with multiple audiograms, the most recent, most complete audiogram was used for evaluation of cross-sectional data.

Head computed tomographic scans with a section width of 3 mm or smaller were evaluated in the axial and coronal reconstructed planes. Tomographs were evaluated for factors selected a priori as potential causes of hearing loss. Conductive hearing loss was hypothesized as potentially associated with deformities of the outer and middle ears, including narrowing of the external auditory canal (EAC), FD involvement of the epitympanum, and FD crowding the ossicular chain. Sensorineural hearing loss was hypothesized as potentially associated with FD that affected inner ear structures, including the internal auditory canal (IAC) and the otic capsule. Dimensions of the IACs and EACs were recorded by a single reader (A.M.B.), as were specific areas of FD involvement within the temporal bone (H.J.K.). Readers were masked to auditory status at the time of computed tomography evaluation.

Comparisons were made between ears affected and unaffected by temporal bone FD and between ears affected and unaffected by hearing loss, as indicated. For participants followed up longitudinally, clinical and radiologic data from the initial and most recent evaluations were analyzed for progression. Statistics and figures were prepared using GraphPad Prism 6 for Windows, version 6.02 (GraphPad Software Inc). Comparisons between groups were made using effect size metrics and 95% CIs. Data are presented as mean, SD, and SE.

A total of 130 individuals with craniofacial FD were identified. Of these, 116 (89.2%) had FD that involved the temporal bone (median age, 19.6 years; range, 4.6-80.3 years; 64 female [55.2%]), whereas 14 (10.8%) had craniofacial FD that did not involve the temporal bone. Temporal bone FD was bilateral in 67 individuals and unilateral in 49 individuals, affecting 183 total ears. A total of 77 ears were unaffected by temporal bone FD. Two individuals (1 with unilateral temporal bone FD and 1 without temporal bone FD) were eliminated from the analyses because of the presence of sinusitis at the time of evaluation.

Participant characteristics and clinical symptoms are given in the Table. Most participants had polyostotic FD and MAS-associated endocrinopathies. The most common concern was skull pain, whereas otologic symptoms were uncommon.

Hearing loss was identified in 41 of 183 ears (22.4%) with temporal bone FD. Conductive hearing loss was most frequent, affecting 27 ears (65.9%). Of these, there was a subclinical conductive component in 13, and the hearing loss was mild in 11, moderate in 2, and profound in 1 ear. Sensorineural hearing loss affected 12 ears (29.3%), which was mild in 10 and moderate in 2 ears. Mixed hearing loss occurred in 2 ears, including 1 with a moderate and 1 with a severe degree of hearing loss.

Longitudinal audiologic data were available for 72 participants, with 112 ears affected by temporal bone FD for a mean period of 6.2 years (SD, 4.4 years; SE, 0.5 years; range, 0.9-15.2 years). The categorical degree of hearing loss worsened in 13 ears, improved in 14 ears, and remained unchanged in 85 ears during the follow-up period.

Hearing loss was detected in 7 of 77 ears (9.1%) in participants with craniofacial FD without temporal bone involvement, which was significantly less prevalent compared with ears with temporal bone FD (41 of 183 ears [22.4%]) (odds ratio [OR], 3.2; 95% CI, 1.4-7.7). This hearing loss included mild unilateral conductive hearing loss of unclear origin in an 11-year-old participant. Sensorineural hearing loss occurred in 6 ears: 3 in elderly participants (>80 years of age) with moderate to severe hearing loss and 3 in participants aged 53, 22, and 19 years with mild hearing loss of unclear origin.

The EAC diameters were compared between the 183 ears affected by temporal bone FD and the 77 ears that were unaffected by temporal bone FD. Ears affected by temporal bone FD were significantly narrower (mean, 4.49 mm [SD, 10.9 mm; SE, 0.08 mm] vs 4.82 mm [SD, 0.69 mm; SE, 0.08 mm]; mean difference [MD], 0.33 mm; 95% CI, 0.11-0.55 mm) and variable among participants (Figure 1). No difference was found between EAC diameter and the presence of conductive hearing loss at any frequency (mean EAC diameter, 4.69 mm [SD, 1.13 mm; SE, 0.21 mm] for participants with conductive hearing loss vs 4.52 mm [SD, 1.03 mm; SE, 0.08 mm] for participants without conductive hearing loss; MD, 0.17 mm; 95% CI, −0.63 to 0.30 mm). On clinical evaluation, severe EAC stenosis (Figure 2) was believed to be directly contributory to conductive hearing loss in 4 participants, all of whom underwent canalplasty. Two of these participants had improvement in hearing at 4 years postoperatively and no recurrence of EAC stenosis at 11 years postoperatively. A third participant underwent canalplasty with removal of cholesteotoma (Figure 2B). Postoperatively, his conductive hearing loss improved from severe to mild and remained stable after 3 years. One participant underwent another operation 2 years after her initial canalplasty because of postoperative FD regrowth. She continues to have normal hearing 5 years after her second operation.

Fibrous dysplasia involvement of the epitympanum was common, affecting 150 of 183 ears (82.0%) with temporal bone FD. In 86 (57.3%) of those ears, FD was limited to the area surrounding the ossicles, whereas 64 (42.7%) had crowding of the ossicular chain (Figure 3). No correlation was found between the presence of FD surrounding the epitympanum and the presence of hearing loss; however, ossicular crowding was associated with conductive hearing loss (OR, 5.0; 95% CI, 2.1-11.6). There was no association between ossicular crowding and sensorineural hearing loss (OR, 2.1; 95% CI, 0.61-7.1). Although the presence of sensorineural hearing loss was nearly double among those with ossicular crowding and the true difference could be as big as 7 times, the precision of the estimate was low and the lower bound crossed the null effect value. Differences in air conduction pure-tone thresholds were observed at both low and high frequencies (eFigure, A in the Supplement), whereas there were no differences in bone conduction thresholds between participants with and without ossicular crowding (eFigure, B in the Supplement).

Tympanometry demonstrated that ears with epitympanic FD had stiffened middle ear systems, as evidenced by lower peak admittance levels (median, 0.4 cm³ [95% CI, 0.4-0.5 cm³] for participants with epitympanic FD vs 0.8 cm³ [95% CI, 0.6-0.8 cm³] for participants without epitympanic FD; absolute median difference, 0.4 cm³; 95% CI of difference, 0.2-0.4 cm³). Peak admittance data were not included for ears with middle ear effusion (n = 2), pressure equalization tubes (n = 6), or tympanic membrane perforation (n = 1).

No difference was found in IAC length from the fundus to the porous in ears with and without temporal bone FD (mean, 11.05 mm [SD, 2.32 mm; SE, 0.17 mm] for ears with temporal bone FD vs 10.68 mm [SD, 1.77 mm; SE, 0.20 mm] for ears without temporal bone FD; MD, −0.37 mm; 95% CI, −0.95 to 0.21 mm). When ears affected by sensorineural hearing loss were analyzed separately, these IACs were found to be elongated compared with ears with temporal bone FD and normal hearing or conductive hearing loss (mean, 12.28 mm [SD, 2.88 mm; SE, 0.77 mm] for ears with sensorineural hearing loss vs 10.91 mm [SD, 2.25 mm; SD, 0.71 mm] for ears without sensorineural hearing loss; MD, −1.33 mm; 95% CI, −2.60 to −0.07 mm) (Figure 4). No difference was found in IAC width in ears with and without FD (mean, 5.31 mm [SD, 1.42 mm; SE, 0.10 mm] vs 5.29 mm [SD, 1.11 mm; SE, 0.12 mm]; MD, −0.02 mm; 95% CI, −0.37 to 0.33 mm) or those with and without hearing loss (mean, 5.34 mm [SD, 0.96 mm; SE, 0.21 mm] vs 5.30 mm [SD, 1.35 mm; SE, 0.09 mm]; MD, −0.04 mm; 95% CI, −0.65 to 0.56 mm).

The area surrounding the otic capsule was a frequent site for FD involvement, affecting 120 of 183 ears (65.6%) with temporal bone FD. No association was found between sensorineural hearing loss and the presence of FD in this area (OR, 0.79; 95% CI, 0.32-1.89). Extension of FD to the membranous labyrinth, such as the semicircular canals and cochlea, was rare, occurring in only 4 ears; however, 2 of these had sensorineural hearing loss and 2 had normal hearing.

No statistical difference in age was found between participants with and without hearing loss (median, 19.3 years [95% CI, 15.7-22.8 years] vs 20.6 years [95% CI, 16.8-21.1 years]; actual median difference, −0.2 years; 95% CI of median difference, −5.1 to 1.8 years). Hearing loss was associated with MAS-associated growth hormone excess, which affected 12 of 29 participants (41.4%) with hearing loss and 16 of 87 participants (18.4%) without hearing loss (OR, 3.1; 95% CI, 1.3-7.5). Hearing loss was also associated with a history of MAS-associated neonatal hypercortisolism, which affected 6 of 29 participants (20.7%) with hearing loss and 2 of 87 participants (2.3%) without hearing loss (OR, 11; 95% CI, 2.5-55). No important associations were found between hearing loss and other MAS-associated endocrinopathies.

Data from this largest series of individuals with FD to date demonstrate that the causes of audio-otologic dysfunction are multifactorial and largely determined by the extent and location of skeletal involvement. Conductive hearing loss was most commonly associated with FD that involves the bony epitympanum, leading to crowding of the ossicular chain. This finding is supported functionally by tympanometry studies, which demonstrated decreased tympanic membrane mobility in ears with epitympanic involvement. Stenosis of the EAC was a less common cause of conductive hearing loss, accounting for only 15% of cases and affecting only 2% of the total cohort. Sensorineural hearing loss was most commonly associated with elongation of the IAC and rarely with invasion of the otic capsule and membranous labyrinth. Understanding the association between these radiographic features and hearing loss will allow clinicians to more accurately identify at-risk patients and ensure monitoring in those with higher-risk features.

These findings provide insight into potential mechanisms of hearing loss in individuals with FD. Differences in hearing sensitivity were observed for the low and high frequencies by air conduction but not by bone conduction in participants with and without ossicular crowding in the epitympanum. This finding suggests a potential mechanism in which stiffening of the ossicles may lead to low-frequency hearing loss, whereas high-frequency hearing loss may result from mass effect on the ossicles related to the surrounding FD. The association of sensorineural hearing loss with IAC length (but not width) suggests that stretching of its contents may be a potential mechanism of hearing loss in FD. Of interest, this mechanism is analogous to the development of optic neuropathy, which occurs rarely in patients with craniofacial FD, resulting from elongation of the optic canal with traction on the optic nerve.¹¹ Future investigations into the anatomical and functional effects of FD on the IAC and its contents could include advanced imaging techniques with 3-dimensional reconstruction and auditory brainstem response testing.

Our findings are consistent with those of the ColI(2,3)+/Rs1+ mouse model, in which invasive FD formation developed. In this transgenic mouse model, observed progressive hearing loss was attributable to FD-like lesions that surrounded the ossicular chain and obliterated the oval and round window of the cochlea. Because the organ of Corti showed no abnormality in histologic and immunocytochemical findings, the progressive hearing loss in this model was conductive in nature rather than sensorineural.⁸

The extensive phenotyping performed in this FD/MAS natural history study also offers an opportunity to identify clinical features associated with audio-otologic disease. This study was the first, to our knowledge, to demonstrate an increased risk of hearing loss in patients with MAS-associated growth hormone excess. Overproduction of growth hormone is presumed to drive expansion of craniofacial FD and has also been linked with optic neuropathy^12,13 and postsurgical regrowth after craniofacial procedures.¹⁴ The correlation between hearing loss and neonatal hypercortisolism is another novel finding; however, the cause of this association is unclear. Of interest, a history of neonatal hypercortisolism has been linked to developmental abnormalities in patients with MAS and, in particular, with disorders of speech and language.¹⁵ Further investigation is needed to determine whether hearing deficits are a contributor to developmental delays in this population.

Findings from this study expand on the relatively limited audio-otologic literature in FD.^16-18 The largest previous series was a retrospective review of 66 patients referred for otolaryngologic evaluation at tertiary care centers that reported a higher prevalence of otologic symptoms and hearing loss compared with our series.⁵ Most in that series were managed nonsurgically, and those authors concluded that conservative management with serial evaluation and imaging review is warranted in most patients. Findings from our series expand on this approach by identifying specific radiologic and clinical features that place patients at increased risk for disease.

Strengths of this study include large participant numbers for this rare disease, making it the largest series in the literature to date. Participants underwent extensive clinical phenotyping and longitudinal follow-up as part of a longstanding natural history study. Because participants were evaluated systematically as part of a research protocol, the prevalence and spectrum of audio-otologic disease were less likely to be affected by referral bias compared with clinical series. Limitations include the inherent weakness in retrospective reviews. In addition, because of the paucity of procedures performed in our cohort, this series was unable to define surgical indications in patients with FD-related audio-otologic disease.

Hearing loss in craniofacial FD is common and typically mild and nonprogressive. The mechanisms of hearing loss are multifactorial based on the location and extent of FD lesions. Deformities of the epitympanum and IAC are most frequently associated with conductive and sensorineural hearing loss, respectively, whereas less common associations include EAC stenosis and otic capsule involvement. Patients with craniofacial FD should undergo evaluation and serial monitoring for audio-otologic disease, including clinical and radiologic evaluation to identify those with high-risk features.

Corresponding Author: Alison Boyce, MD, Section on Skeletal Disorders and Mineral Homeostasis, National Institute of Dental and Craniofacial Research, National Institutes of Health, 30 Convent Dr, Room 228, Mail Stop Code 4320, Bethesda, MD 20892 (boyceam@mail.nih.gov).

Accepted for Publication: September 21, 2017.

Published Online: November 30, 2017. doi:10.1001/jamaoto.2017.2407

Author Contributions: Dr Boyce had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Boyce, Brewer, DeKlotz, Zalewski, Collins, Kim.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Boyce, Brewer, Collins, Kim.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Boyce.

Obtained funding: Collins.

Administrative, technical, or material support: Boyce, Brewer, Zalewski, King.

Study supervision: Boyce, Brewer, Collins, Kim.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.

Funding/Support: This research was supported by the Intramural Research Programs of the National Institute of Dental and Craniofacial Research and the National Institute on Deafness and Other Communication Disorders.

Role of the Funder/Sponsor: The National Institutes of Health Intramural Program had a role in the design of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.