Craniofacial, Temporal Bone, and Audiologic Abnormalities in the Spectrum of Hemifacial Microsomia FREE

HEMIFACIAL microsomia (HFM) is a term coined by Gorlin and colleagues.^1- 2 This disorder has also been called "otomandibular dysostosis,"³ "first branchial arch syndrome,"⁴ "second branchial arch syndrome,"⁵ "oculoauriculovertebral sequence,"⁶ "Goldenhar syndrome,"⁷ "lateral facial dysplasia,"⁸ and "craniofacial microsomia."⁹ Hemifacial microsomia manifests in a highly variable phenotype. It is the second most common craniofacial malformation after cleft lip and cleft palate. Any structures derived from the first and second pharyngeal arches can be affected. Although unilateral presentation is preponderant, bilateral anomalies are also seen in 30% of these patients.¹⁰

This study analyzed the severity of the craniofacial features, temporal bone abnormalities, and audiologic findings in patients with HFM. We also sought to assess a possible relationship between the clinical and dysmorphic features of these patients with the abnormalities in the temporal bone computed tomographic (CT) scans and the severity and type of hearing loss.

The medical records of 162 patients with the diagnosis of HFM seen in the Craniofacial Center at The Children's Hospital, Boston, Mass, were reviewed. Only those patients with complete medical records who had undergone temporal bone CT scan and audiologic workup were included in the study. Forty patients were eligible for this retrospective study.

Medical records were reviewed and data were recorded for the following: the patient's sex, age, family history, facial nerve function, history of otologic disease, and audiologic evaluation. Temporal bone CT scans were reviewed for abnormalities of the external auditory canal (EAC), middle ear, ossicles, inner ear, mastoid development, facial nerve canal, and mandibular condyle.

The OMENS classification system¹¹ was used to grade the abnormal components in patients with HFM. Each letter of the acronym indicates 1 of the 5 major dysmorphic manifestations: O, orbital asymmetry; M, mandibular hypoplasia; E, auricular deformity; N, nerve involvement; and S, soft tissue deficiency. Each of these 5 anatomical features was graded from 0 to 3 (ie, 0 indicates none; 3, worse) according to the severity using physical examination findings, photographs, and radiographs, including posteroanterior, lateral, and cephalometry. A total OMENS score was obtained by summing the 5 anatomical features (Table 1).

A single neuroradiologist (C.D.R.) reviewed the temporal bone CT scans. For analysis, radiographic findings of each anatomical category were assigned an ordinal integer score. Data were recorded on the following: EAC (normal [0], stenosis [1], or atresia [2]), middle ear (normal [0], hypoplastic [1], or atretic [2]). ossicles (normal [0], malformed and fused [1], or unidentified [2]), mastoid (normal [0], poor pneumatization [1], or absent pneumatization [2]), facial nerve (normal [0], displaced [1], or could not be identified [2]), condyle (normal [0], displaced [1], or hypoplastic [2]), and inner ear abnormality (normal [0] or abnormal [1]). A total radiographic score was obtained by summing the scores for each category.

Individual and total OMENS scores were compared with the total radiographic score, type, and degree of hearing loss. Patients manifesting unilateral and bilateral anomalies were analyzed separately. For patients with bilateral involvement comparisons were performed on the basis of side rather than patient since each side was evaluated separately. For analysis, the degree of hearing loss was assigned an ordinal integer score, ie, 0 (normal), 1 (mild), 2 (moderate), 3 (moderate-severe), 4 (severe), or 5 (profound).

Since numerical results were ordinal scores rather than continuous variables, nonparametric procedures were used for all analyses. Correlation between the OMENS scores and both the hearing loss and the radiographic score was assessed using the Spearman rank correlation coefficient (ρ). Comparison of OMENS scores on the basis of type of hearing loss was performed using the Kruskal-Wallis procedure.⁵ Patient summaries were expressed as medians, ranges, and frequency distributions. Owing to the nature of this study, the conservative Bonferroni correction for multiple testing was not applied. However, to provide a higher degree of protection to the experimentwise type I error rate, we considered the results to be statistically significant if P<.01. Statistical analysis was performed using SAS version 6.12 (SAS Institute, Cary, NC).

There were 17 male (43%) and 23 female patients (57%), ranging in age from 2 to 37 years (mean age, 12 years). None of these patients had chromosomal anomalies, a history of teratogenesis, or a family history of craniofacial abnormalities. There were 17 (43%) left-sided, 16 (40%) right-sided, and 7 (17%) bilateral presentations (Table 2).

The 5 major OMENS features (ie, orbital, mandible, ear, nerve, and soft tissue) are listed in Table 1. Abnormal orbital size and position were noted in 4 patients (12%) with unilateral HFM and 1 patient (14%) with bilateral HFM. Mandibular hypoplasia, ranging from a small short ramus to complete absence of a ramus, was noted in 32 patients (97%) with unilateral HFM and 7 patients (100%) with bilateral HFM. Unilateral HFM as defined in Table 1 presented as follows: E1 (n = 6 [18%]), E2 (n = 6 [18%]), and E3 (n = 19 [58%]). Bilateral HFM presented as E1 (n = 2 [29%]), E2 (n = 2 [29%]), and E3 (n = 3 [42%]). Six patients (15%) had preauricular tags.

Facial nerve weakness as defined in Table 1 was documented in 16 patients (48%) with unilateral HFM. Four patients had weakness of N1 (frontal or zygomatic), 7 patients had weakness of N2 (ie, buccal, marginal mandibular, or cervical), and 5 had weakness of N3 (complete facial nerve). Two of these patients also had weakness of the unilateral hypoglossal nerve (CN12). Four patients (57%) with bilateral HFM had facial nerve paresis. Two of these patients had bilateral and 2 had unilateral manifestations of facial nerve weakness (Table 2). Soft tissue or muscle deficiency was noted in 19 patients (58%) with unilateral HFM and 2 patients (29%) with bilateral HFM.

A history of chronic otitis media was noted in 20 patients (60%) with unilateral HFM and 3 patients (43%) with bilateral HFM. Fifteen patients (45%) with unilateral HFM and 1 patient (14%) with bilateral HFM had undergone placement of tympanostomy tubes. Five patients (15%) with unilateral HFM and 5 patients (71%) with bilateral HFM were successfully using hearing aids.

Patients with unilateral HFM (Table 2) initially had the following: normal hearing (n = 1 [3%]), mild conductive hearing loss (CHL) (n = 2 [6%]), moderate CHL (n = 1 [3%]), moderate-severe CHL (n = 5 [15%]), severe CHL (n = 14 [42%]), profound CHL (n = 3 [9%]), profound sensorineural hearing loss (SNHL) (n = 2 [6%]), and bilateral moderate to severe CHL (n = 2 [6%]). Three patients (9%) initially had bilateral hearing loss (Table 2). Patients with bilateral HFM (Table 2) initially had the following: bilateral moderate to severe CHL (n = 3 [43%]), bilateral severe CHL (n = 2 [29%]), bilateral moderate mixed hearing loss (n = 1 [14%]), or unilateral severe CHL (n = 1 [14%]).

Atresia of the EAC was noted in 20 patients (61%) with unilateral HFM and in 4 patients (57%) with bilateral HFM. Canal stenosis was observed in 12 patients (36%) with unilateral HFM and in 2 patients (28%) with bilateral HFM. Only 1 patient with unilateral HFM had a normal EAC. One patient with bilateral HFM had atresia on one side and stenosis on the other (Figure 1 and Figure 2).

Patients with unilateral HFM had the following findings: normal middle ear (ME) (n = 4 [12%]), hypoplastic ME (n = 23 [70%], and atretic ME (n = 6 [18%]). All patients with bilateral HFM had hypoplastic MEs. Ossicles were malformed and fused in 24 patients (73%) with unilateral HFM and 6 patients (86%) with bilateral HFM. Ossicles could not be identified in 8 patients (24%) with unilateral HFM because of severe atresia of the ME.

Mastoid air cells were absent in 14 (42%) and poorly developed in 8 patients (24%) with unilateral HFM; and absent in 4 (57%) and poorly developed in 2 (29%) patients with bilateral HFM. Thirty patients (90%) with unilateral HFM and 7 patients (100%) with bilateral HFM had displacement and some degree of condylar hypoplasia.

Abnormalities of the facial nerve canal were noted in 35 (88%) of the 40 patients. Anterior displacement of the facial nerve, most commonly in the mastoid segment, was seen in 21 patients (64%) with unilateral HFM and in 4 patients (57%) with bilateral HFM. The facial nerve canal could not be identified in 8 patients (24%) with unilateral HFM and in 2 patients 29% with bilateral HFM.

Hypoplastic oval window was the most common inner ear abnormality and was noted in 12 patients (36%) with unilateral HFM and in 2 patients (29%) with bilateral HFM. An abnormal cochlea was noted in 1 patient with unilateral HFM and in 1 patient with bilateral HFM; these 2 patients had profound SNHL and bilateral moderate mixed hearing loss, respectively. Other inner ear abnormalities noted in patients with unilateral HFM were as follows: hypoplastic vestibule (n = 2), hypoplastic semicircular canals (n = 1), atretic facial nerve recess (n = 1), and atretic round window (n = 2).

Patients' demographics and summarized data are listed in Table 3 and Table 4. Patients with unilateral HFM had a significant correlation between the "E" score (auricular abnormality) and the radiographic score (ρ+0.57, P<.001). However, no significant correlation was found for other individual scores (orbit, mandible, nerve, or soft tissue). Total OMENS score (severity of craniofacial features) correlated significantly with the total radiographic score for both unilateral (ρ = 0.47, P = .006) and bilateral (ρ = 0.83, P<.001) presentations. All significant ρ values were positive, indicating that the high total OMENS scores are associated with high radiographic scores and vice versa. No statistically significant correlation between the OMENS scores and the degree or type of hearing loss was observed for either unilateral or bilateral presentations. The P values for all statistical tests are summarized in Table 3 and Table 4.

The incidence of HFM has been estimated at 1 in 5600 births.¹¹ It usually manifests with varying degrees of hypoplasia and asymmetry of bony and soft tissue of the face that is usually unilateral but not uncommonly bilateral—7 patients in our series. It has been suggested that there is a 3:2 predilection for males and the right side of the face.¹² In our sample, no evidence of sex preference or dominance of facial asymmetry to either side was noted; confirming our earlier study of 121 patients.¹¹

The cause of HFM is thought to be pathogenically heterogeneous. It is as if the defective genes, teratogens, and vascular anomalies singly or collectively can cause disruption of normal development leading to a wide spectrum of the anomalies seen in these patients. Poswillo¹³ devised a chemically induced murine phenocopy and showed that focal hemorrhage from the stapedial artery supplying the first and second pharyngeal arches could be the primary cause. However, it is unlikely that embryonic hematoma formation could account for the wide range of HFM features, particularly those outside the craniofacial regions. Johnston and coworkers^14- 15 showed that exposure to thalidomide and retinoic acid could lead to anomalies similar to HFM. Accutane (13-cis retinoic acid) interferes with proliferation and migration of neural crest cells, causing facial and cardiac abnormalities similar to the HFM. Otani et al¹⁶ proposed a transgenic mouse model with an insertional mutation on chromosome 10. Hemifacial microsomia can also occur in patients with chromosomal disorders such as trisomy 18, trisomy 7, 9 mosaicism, and terminal deletion of 22q. There are also pedigrees with HFM, and in some affected families the history is consistent with autosomal dominant or recessive inheritance.^17- 20

Several classification systems have been proposed to document and analyze the clinical manifestations of HFM. Mandibular hypoplasia seems to be the element central to all these schemes. Pruzansky²¹ described the 3 types of mandibular abnormalities; however, he did not account for other associated manifestations of HFM. Harvold et al²² described 5 types of mandibular anomalies, but did not include auricular, neural, and orbital anomalies. David et al¹² presented the skeletal, auricular, and soft tissue (SAT) system modeled after the TMN classification of tumors. However, this system does not permit independent evaluation of the orbital and cranial nerve abnormalities. The OMENS system, proposed in 1991,¹¹ is an expansion of Pruzansky's classification. It includes 5 of the major craniofacial manifestations of HFM and allows independent grading of the dysmorphic features. In this system, each anatomical abnormality is graded from 0 (normal) to 3 (most severe). Cousley²³ compared the SAT with the OMENS system and concluded that the OMENS system was more sensitive to the wide phenotypic heterogeneity of HFM. The OMENS system was later expanded by Horgan et al²⁴ to OMENS-Plus to include the extracraniofacial anomalies.

The association between HFM and anomalies of the other organs (ie, central nervous system, cardiac, pulmonary, renal, gastrointestinal, and skeletal) has been well documented. Sporadic malformations present in a frequency of 1 in 1000 births (ie, congenital lip and palate, tracheoesophageal fistula, atrial septal defect, and tetralogy of Fallot). The term "associated anomaly" is used when a defect occurs in 10% to 15% of the cases and is pathogenically related to the primary abnormality. Associated anomalies in patients with HFM are reported as: central nervous system, 5% to 15%^24- 25; cardiac, 45% to 55%²⁶ and 26.4%³; genitourinary, 5% to 6%²⁵ or 10%²⁴; pulmonary and gastrointestinal, 10%²⁴; and skeletal, 41%.²⁴

Patients with HFM can present with a wide range of anomalies; however, facial nerve weakness and hearing loss are the most common functional deficits. The prevalence of facial nerve palsy with HFM varies in range as seen from the following sources: 10%, Grabb²⁷; 19%, Converse et al²⁸; 25%, Murray et al²⁹; 22%, Bassila and Goldberg³⁰; 45%, Vento et al¹¹; and 22%, Carvalho et al.¹⁰ In our series 20 (50%) of the 40 patients showed some degree of facial nerve weakness, and 2 patients also had unilateral hypoglossal nerve weakness. Facial nerve weakness could be due to the deficiency of the mesoderm of the pharyngeal arches, neural ectoderm, or a combination of both. Evaluations of the temporal bones have shown the following: (1) abnormal course and exit point of the facial nerve over the temporamandibular joint²⁸; (2) development of the nervous intermedius component of the facial nerve, and no development of the mastoid component³¹; (3) hypoplasia of the entire course of the facial nerve.³²

The most common type of hearing deficit in HFM is conductive loss—34 patients (86%) in our series. However, the presentation of SNHL in these patients remains underappreciated. The overall incidence of congenital SNHL in the general population is 0.001% to 0.004%, and 3% to 4% in patients with craniofacial syndromes.³⁰ The incidence of SNHL in our patients was 10% (4 patients) consistent with other reports, 11% (Carvalho et al¹⁰) and 16% (Bassila and Goldberg³⁰).

Hemifacial microsomia can manifest as a structural abnormality in any tissue derived from the first and second pharyngeal arches. The mandibular and auricular deformities are the main components of the dysmorphic features. In our series, 39 patients (97%) had mandibular hypoplasia and 38 patients (95%) showed auricular abnormalities. Our data showed a significant correlation between the severity of the external auricular anomalies ("E" score) and the extent of temporal bone abnormalities in patients with unilateral HFM. We also found that the overall severity of craniofacial features (total OMENS score) correlated significantly with the degree of the temporal bone abnormalities (total radiographic score) for patients with unilateral and bilateral HFM. Thus, in patients with abnormal hearing, temporal bone CT scan must be obtained to further assess the degree of stenosis and atresia of EAC, status of the ossicular chain and ME, and inner ear anomalies. The timing of the temporal CT scan must be individualized based on the the age of the patient, severity or worsening of the hearing status, and no otologic surgery. At our craniofacial center, a 3-dimensional CT study is always done in preparation for mandibular elongation in the age range of 5 to 7 years. Whenever possible, a combined study for temporal bone abnormalities and type of mandibular hypoplasia should be done concurrently.

Also, our data show that there is no correlation between the severity of the dysmorphic features of HFM and the degree or type of hearing loss. Patients with minimal dysmorphic features can present with a moderate-severe degree of hearing loss. Failure to appreciate this finding could delay proper and timely diagnosis, thereby prolonging sensory deprivation and delayed speech and language acquisition. Thus, we strongly recommend a complete audiologic evaluation of every child with a diagnosis of HFM, regardless of the type or the severity of the clinical manifestations.

Our results confirmed that evaluation and interpretation of data using the OMENS classification system provide a logical and comprehensive manner to document, independently analyze, and compare the major dysmorphic features of HFM. Furthermore, we agree with Cousley²³ that an asterisk be added to the "E" component of the acronym (OME*NS) to indicate the type and degree of hearing loss.

Accepted for publication August 11, 2000.

Presented in part at the American Society of Pediatric Otolaryngology Combined Otolaryngological Society Meeting, Orlando, Fla, May 2000.

Corresponding author: Reza Rahbar, DMD, MD, Department of Otolaryngology and Communication Disorders, The Children's Hospital, 300 Longwood Ave, Boston, MA 02115 (e-mail: rahbar@a1.tch.harvard.edu).