Analysis of a Cohort of Children With Sensory Hearing Loss Using the SCALE Systematic Nomenclature

Transcription

1 The Laryngoscope Lippincott Williams & Wilkins, Inc., Philadelphia 2000 The American Laryngological, Rhinological and Otological Society, Inc. Analysis of a Cohort of Children With Sensory Hearing Loss Using the SCALE Systematic Nomenclature Nancy Sculerati, MD Objectives: What characteristics identify clinical types of childhood deafness? Which aspects of the otological evaluation best delineate them? To approach these related questions, a classification for deafness consistent with current medical concepts was constructed using a systematic nomenclature and then applied to a pediatric cohort of 168 children with sensorineural hearing loss (SNHL) who were referred for private consultation. A major aim of the analysis was to test the utility of SCALE, the new systematic nomenclature. Methods: Patients with SNHL were identified through the office records of a single faculty member of the Department of Otolaryngology in a medical school situated in a major US city. Inclusion criteria required bone conduction thresholds above 30 db or equivalent in at least one of the frequencies from 250 Hz to 4 khz on either behavioral audiogram or on electrophysiological testing. All identified patients had initial visits during an 8-year period from late 1990 to early Patients were excluded if age at first consultation was 19 years or more, if records were insufficient to confirm SNHL, or if further evaluation revealed that SNHL had been misdiagnosed. A formal nomenclature was designed to systematically encode clinical features with simple descriptive terms according to an acronym (SCALE [sidedness, component function, age of onset, lesion, and etiology]) for all included patients. Results: One hundred sixty-eight study patients were analyzed; sensory hearing loss was bilateral in 82% (137/168) and unilateral in 18% (31/168). The etiology of this impairment was determined to be intrinsic in 40% of children (67/168), either secondary to genotype (57/ 67), or to named congenital syndromes without known extrinsic cause (10/67). Recessive single gene Presented as a Candidate s Thesis to the American Laryngological, Rhinological and Otological Society, Inc. From New York University School of Medicine, Department of Otolaryngology, New York, New York. Editor s Note: This Manuscript was accepted for publication January 5, Send Correspondence to Nancy Sculerati, MD, 13 Tracy Road, Pawling, NY 12564, U.S.A. mutations diagnosed by family history, recognition of syndrome, or determination of homozygous 35delG mutations in the gap junction protein gene, Connexin 26, accounted for bilateral sensory hearing loss in 33 children (24% of all bilateral cases). One girl had an X-linked dominant syndrome (Coffin-Lowry syndrome) with auditory brainstem response documented childhood onset of SNHL. Nine patients (5%) had chromosomal aneuploidy, and 12 patients (7%) had either a family history of dominant deafness (7/ 12) or a recognizable autosomal dominant syndrome (5/12), most commonly, Waardenburg syndrome type 1 (4/5). Extrinsic causes of deafness were identified in only 13% of children (21/168) and included a relatively large number of referrals from neurosurgery (9/21). Three of these children had chronic middle ear disease and sensory hearing loss associated with inflammatory and bony changes on temporal bone imaging suggestive of chronic osteitis; all had a history of active otitis media during cranial irradiation. Congenital cytomegalovirus infections were documented in only 4 cases, but 41 patients could have had this as a cause or did not have this cause ruled out. An idiopathic cause or origin was assigned to 36% of patients (61/168) including patients with unnamed syndromic patterns of multiple anomalies. Conclusions: The SCALE nomenclature facilitated complete descriptions of hearing-impaired children and provided a classification scheme applicable to broad categories of human disease. The single most useful diagnostic test was screening for Cx26 mutations. Computed tomography scan of the temporal bones was helpful in establishing etiology for selected patients and invaluable in patients with chronic ear disease. Magnetic resonance imaging scan was a superior diagnostic modality in one child with a posterior fossa arachnoid cyst. Key Words: Hearing loss, deafness, Connexin 26 etiology, nomenclature. Laryngoscope, 110: , 2000 INTRODUCTION How are hearing-impaired children medically evaluated in the United States? There is such variation between 787

2 the practice styles of physicians that no single answer is possible. As a minimal standard of care, these evaluations do include a medical history, family history, focused otolaryngological examination, and review of audiometric test results. The extent of any additional laboratory testing, including referral for imaging studies or molecular genetic testing and referrals for subspecialty consultations, is not only variable but also is controversial. 1 3 Increasing attention to the cost of health care is likely only to emphasize such controversies in coming years. Universally, however, some diagnostic evaluation is performed for each child to classify the cause and severity of hearing impairment. This classification ordinarily stands as the diagnosis in the patient s medical, school, and disability records, even if little specific information is conveyed by the terms given. 4 Conflicts Within Traditional Classification of Deafness The current classification of deafness was constructed in an era when history, physical examination, conventional blood tests and urine analysis, along with behavioral audiometry, made up the tools available to distinguish types of adult and childhood sensorineural hearing loss (SNHL). 5 Using these tools alone, unless the patient had a history of meningitis or stigmata of congenital infection, most SNHL has been classed as idiopathic or attributed to a genetic origin without specific medical diagnosis and without an anatomical or pathophysiological description of the lesion. Even the severity of hearing loss has been described by qualitative terms that have no uniformly accepted definitions among specialists (including audiologists and otologists). The nomenclature presently used to classify childhood SNHL stems from work begun in the 19th century and refined in post World War II epidemiological analysis of students at schools for the deaf. This work led to a generally accepted, but informal, classification of deafness based on distinguishing characteristics. In most groups of deaf children such attributes were few and composed distinct causes (where known by history) or features evident by either physical examination or by available laboratory tests of that era. Accordingly, the hearing-impaired child who had been born deaf (as far as could be determined) and who had never had any known trauma or infection that might account for hearing loss in infancy was designated as having congenital deafness. The infant who survived bacterial meningitis was placed in a category of acquired deafness. 6 Genetic causes were recognized as responsible for a major portion of childhood SNHL and, since families were typically large and geographically static, family history was very likely to reveal known affected relatives if deafness was due to hereditary. Therefore genetic deafness (like other genetic disease) was called familial rather than sporadic. Patients who had a recognizable pattern of unusual or abnormal features that included hearing impairment were placed in a syndromic category, in contradistinction to hearing-impaired children who could not be otherwise distinguished from normal-hearing children. The majority of children lacking a confirmed acquired cause or origin were diagnosed with nonsyndromic idiopathic deafness. 788 The concept that some newborns were at high risk for hearing loss because of factors that could be determined by perinatal history and physical examination was accepted among physicians and audiologists by the third quarter of this century adding another etiological category to the scheme. Identification of a risk factor does not provide insight into the specific lesion causing hearing loss, however, and even with this addition, in the single largest group of children with SNHL the origin remains idiopathic. 2,3 Advances in otology since 1990 are altering many previous assumptions concerning human deafness. The management of irreversible deafness has changed with cochlear implantation. The diagnosis of medical causes of SNHL is becoming precise with the clinical application of molecular biology techniques. Elucidation of the biology of deafness in humans and animals, particularly in deaf mutant mice, has advanced understanding of normal cochlear function and allowed a more meaningful description of disease in some cases. The relative inaccessibility of the functioning cochlea, along with the microscopic scale of functional inner ear components, has always complicated basic investigation of cochlear physiology. As DNA mutations in deafness are identified, in situ hybridization and other localization techniques are being used to mark the cells and sites expressing deafness gene products. At the same time, different forms of inherited SNHL are being clarified clinically as specific patient populations are for the first time being delineated using molecular diagnostic tests for mutations in deafness genes. The etiology of SNHL is understood today in terms of a biochemical pathophysiological lesion, such as abnormal ionic gradients produced by defective potassium channels located in outer hair cells, 7 rather than in terms of a broad etiological category of genetic deafness in some of these cases. The presently used classification of deafness does not readily accommodate these refinements in diagnosis or pathophysiology. As more is being understood, the traditional categories of congenital versus acquired and nonsyndromic versus syndromic and separate etiological categories of infectious or ototoxic or genetic causes of deafness have been rendered much less useful even for nongenetic cases. It is clear, for example, that SNHL may be acquired from infectious causes in utero and is both acquired and congenital in such cases. The onset of infant deafness may be postnatal, despite the proximate cause of hearing impairment (e.g., cytomegalovirus [CMV] infection or inheritance of the Pendred syndrome [PDS] gene mutation) having definitely occurred before birth. 8,9 Previously, nonsyndromic forms of deafness were assumed necessarily to have different causes (probably different abnormal genes) than deafness that was associated with other physical features in a recognizable syndrome. 9 It is now apparent that the very same gene product can be responsible for both syndromic and apparently nonsyndromic forms of deafness countering a basic rationale for the categorical distinction. This finding has occurred as more patients are tested for mutations in mapped deafness genes, and results have indicated that some of the

3 recognizable syndromic forms of deafness include only a fraction of a larger affected population. For example, Pendred syndrome patients have goiter, SNHL, partial Mondini deformities with wide vestibular aqueducts (WVA) on temporal bone imaging, and abnormal perchlorate discharge test results because of an underlying abnormality in the organification of iodine. Some deaf patients without any thyroid abnormalities on either physical examination or thyroid function tests, but who also have mild Mondini deformities associated with WVA and abnormal perchlorate discharge tests, test positive for PDS gene mutations similar or identical to those present in patients with the full syndromic stigmata.10,11 In many patients, deafness following administration of ototoxic drugs is due to an increased susceptibility from specific gene mutations (e.g., mitochondrial 1555 mutation), rather than to an inappropriately high serum level of medication. In these cases of aminoglycoside toxicity, acquired deafness is actually a form of hereditary hearing loss.12 Sporadic versus familial categories are not useful terms for most of the population in the developed nations, where typical families are so small (and relatives so dispersed) that a positive family history is statistically unlikely even if deafness is inherited by recessive genes or by a dominant gene with incomplete penetrance. Classifications of deafness are valuable in both clinical medicine and auditory science when meaningful information is encoded. An optimal classification allows a complete description of patient characteristics at initial presentation but will also accommodate results of further evaluation and any change in hearing status over time, as well as alterations of specific etiological diagnosis that are likely to result from future advances in this field. To these ends, a modified descriptive classification for deafness, the SCALE system of nomenclature, was developed. The acronym SCALE represents sidedness (unilateral or bilateral), component function (sensory thresholds, conductive thresholds, vestibular function), age of onset, lesion, and etiology (Fig. 1). Fig. 1. Patient (case 127) was referred at age 4 years by her pediatrician. A diagnosis of Coffin-Lowry syndrome was not established until age 6 years, despite consultation with genetics. Habilitation of global developmental delay included physical and speech therapies. A mild hearing loss in the sound field associated with bilateral otitis media resolved after myringotomy tube placements, and normal hearing was confirmed with auditory brainstem response (ABR) testing. A left-side primary acquired cholesteatoma was found on computed tomography (CT) scan; otic capsules appeared normal. Subsequently a sensory loss was noted, first on the right, then bilaterally. Amplification was immediate. Speech and language were age-appropriate. Complete SCALE description is given as follows. Asymmetric sensory hearing loss (SHL); R severe, L mild (low-frequency) with recurrent conductive overlay; otitis media, cholesteatoma (L); late-childhood onset SHL (8 y); unknown lesion, presumed cochlear; X-linked dominant inheritance, Xp22.2-p22.1; positive for Coffin-Lowry syndrome. 789

4 Rationale for SCALE System of Nomenclature General considerations. The systematic nomenclature of SCALE was designed to provide an automatic guideline for patient evaluation. Categories were chosen to be both consistent with current concepts and likely to accommodate advances in the field. How can future advances be predicted? When asked to project research trends over the next 25 years, Peter Dallos 13 (president of the Association for Research in Otolaryngology, 1998) prefaced his answer with a quote from the New York Yankee baseball manager, Yogi Berra: Prediction is very hard, especially when it is about the future. In SNHL, the assumption that application of molecular genetic technology will transform otological diagnosis and treatment is based on biological facts: single gene mutations and cochlear malfunction on an ultrastructural and molecular level are the most common proximal causes of human deafness. Descriptions of features that are important for taking care of patients are unlikely to become obsolete with the addition of new levels of knowledge. In sensory deafness, such features include the risk that the underlying cause can be transmitted to others, the prognosis of aural habilitation, the likelihood that further deterioration in hearing level will occur, and the probability that specific additional abnormalities, impairments, or disease states are present. The major clinical focus for the newly identified hearing-impaired child is, appropriately, on habilitation and management rather than on diagnosis and classification. Certain etiological aspects of human deafness are liable to remain important because the prognosis for aural habilitation partly depends on them. For example, both the thresholds of remaining hearing acuity and preservation of tonotopic organization of the auditory system must contribute to the auditory abilities of the hearing impaired and are likely to influence habilitation independent of future improvements in technologic aid by amplification, electrical stimulation, or other strategy Each component in the SCALE system was chosen using the general considerations discussed above, according to the following specific rationales. Side (laterality). There is an obvious qualitative difference between bilateral and unilateral hearing impairment. No doubt, even unilateral sensory loss can impede aural comprehension and speech development (particularly in association with learning disabilities), but this distinction remains functionally important. The distinction is so basic that classifications of the hearing-impaired, aimed strictly at delineation of disability, often completely exclude individuals with unilateral impairment. Since the most basic biology of deafness is unchanged by laterality, both classes are included in this medical classification of deafness. For example, a number of specific diagnoses produce either unilateral or bilateral cochlear impairment (e.g., Waardenburg syndrome and congenital CMV infection). Conversely, a deafness-causing lesion, particularly a tumor, is sometimes likely to have a certain etiology only if present bilaterally. Many bilateral neoplasms are hereditary in origin, such as acoustic neuroma in neurofibromatosis type 2. In most cases of bilateral SNHL, the peripheral hearing acuity is equal, or nearly equal, on each side. In cases 790 where the hearing level is so different that a different class of hearing sensitivity exists on each side, the term asymmetric bilateral SNHL (AsBSNHL) is used. This definition of asymmetric hearing loss does not prevent audiological description of asymmetry at one or more particular frequencies in cases of bilateral SNHL. In asymmetric SNHL, the acuity class on each side is described sequentially. Component function: bone and air conduction thresholds and vestibular status. Hearing impairment is typed as sensory or conductive followed by a qualitative description of acuity based on thresholds for each component. The terms sensory and sensorineural are used interchangeably. The former is preferred as simpler. The term mixed is not excluded from use but is not included as a formal part of this nomenclature. Instead, each component is described separately, if present. Although vestibular function is not a part of auditory acuity, the status of the peripheral vestibular system is an important feature in delineating known types of deafness. Therefore vestibular function is described as a component function. These minor adjustments in terminology simplify terminology and add information. Central auditory, eighth cranial nerve, and inner hair cell receptor dysfunctions are all indisputably part of the sensory nervous system; in fact, literal nerve deafness is unusual in humans. This semantic distinction was perhaps once trivial, but now has clinical importance in communicating with patients and families concerning cochlear and brainstem implants. The concept that an underlying sensory hearing acuity class exists in all people with the possibility for decrease in sound perception if a conductive overlay is added in all those except the most profoundly deaf is a sufficiently important concept for physicians (including primary care physicians) to intuit when caring for children that this explicit nomenclature is warranted. There are classes of residual hearing acuity. Residual acuity correlates with the functional deficit in deafness. As John Ballantyne 17 pointed out 30 years ago, there is no really satisfactory classification of the various degrees of deafness, and no two authorities will agree on where to draw the line between one degree and the next. Compounding the question of just which threshold level separates classes of impairment, differences in the quality of sound perception also exist, may be unrelated to peripheral thresholds, and are difficult to measure with speech discrimination tests in young children. Therefore the classification for degree of hearing loss in this nomenclature (and in any nomenclature) is arbitrary. However, a standard meaning for the terms used makes them more powerful. Acuity classes for this system have been defined as follows: mild loss, 25 to 45 db; moderate loss, 46 to 65 db; severe loss, 66 to 85 db; and profound loss, greater than 86 db. The auditory brainstem response (ABR) thresholds are similarly defined as the lowest intensity level at which wave V could be detected and replicated at equivalent dbnhl. Ability to understand the spoken word is the underlying basis for the acuity classes. Clearly, although many individuals have fairly constant residual acuity throughout the frequency range required for most languages,

5 there are patterns of residual hearing that are frequency dependent. The term low-frequency hearing loss (LfHL) is used for decreases in hearing acuity occurring at or below 500 Hz, with otherwise normal hearing for that component. If there is hearing impairment for the component (e.g., a moderate sensory loss with a severe sensory loss at 500 Hz), the impairment may be described as SHL, moderate (severe low-frequency). If, bilaterally, a moderate sensory loss has a mixed low-frequency component, the component acuity is described as bilateral SHL, moderate with low-frequency conductive overlay. High-frequency hearing loss (HfHL) is defined as a decrease in hearing acuity for pitches at or above 2,000- to 4,000-Hz frequencies, with normal mid-frequency and low-frequency thresholds for that component. Midfrequency hearing loss is similarly defined for the primary pitch of the speech frequencies (1,000 Hz). Very-highfrequency thresholds were not examined in this study. In this report, conductive impairments were described only if there was an underlying SNHL, but the SCALE nomenclature can be used to code for isolated conductive impairments. Classes of vestibular function were not defined; delineation of vestibular function was not required to describe any of the entered patients. None had reliable vestibular testing. Major progression, minor progression, and fluctuating hearing loss were considered. For the individuals affected, there is an enormous difference between stable and progressive deafness. Progressive SNHL in childhood has been reported in the literature with an extremely variable prevalence (from 4% to 30%) depending on the report. 18 Generally, audiologists and otologists consider threshold changes of 10 db significant. It has been well documented that a large percentage of young people with SNHL will have drops and fluctuations in measured acuity over years and decades. This kind of progression has not ordinarily been distinguished from changes in the basic acuity class of residual hearing, perhaps partly accounting for the range of reported progression in pediatric SNHL and for different clinical opinions concerning progressive deafness. There is a clinical schism, however, between minor progression and major functional deterioration in hearing acuity that occurs when, for example, a child s impairment in the speech frequencies changes from a moderate to a severe or profound level, particularly if the change occurs over a few days, weeks, or months. In the SCALE nomenclature, this type of change is termed major progression. It includes late or childhood-onset hearing loss where levels change from normal to hearing impaired if further major deterioration occurs; transitions are marked by the symbol. Progression and fluctuations that do not change the acuity class of the hearing impairment are termed minor progression and minor fluctuation. Minor progression was defined as sensorineural decrease in hearing of 10 db or more at any one frequency for reliable behavioral or ABR threshold. Minor fluctuation was defined as an improvement of at least 10 db, similarly measured, at one or more times. Progressive and fluctuating hearing losses were assigned only if there was no concurrent middle ear disease that might influence threshold variation. Age of onset. The age of onset of hearing loss is likely to influence auditory performance. Although none of the patients in this report were given a prenatal time point, there is no doubt that in some cases a functioning ear or other auditory anatomical area was never formed or was sufficiently functionally impaired during development to change spontaneous activity influencing other areas of the central nervous system. The aural language skills of young children with SHL are well known to vary significantly enough to be qualitatively different in different individuals with apparently similar peripheral hearing thresholds, age at intervention, methods of habilitation, and intellectual function. There are children with so-called corner audiograms who, despite minimal measurable residual hearing and late identification, perform at or above age-appropriate language levels using conventional amplification, as well as cochlear implant recipients who rapidly achieve levels of performance beyond their peers in an implant program. Conversely, children with normal peripheral hearing thresholds are also known to have learning disabilities involving auditory processing. These individuals have poorer skills for speech discrimination in noise and lower than expected levels of speech and language skills in the absence of habilitation, on an auditory basis, despite normal hearing thresholds. Although language centers are not specifically auditory, these differences reflect auditory skills rather than symbolic language ability per se. Known attributes of the developmental biology of the auditory system in particular, and the central nervous system in general, suggest that, at least partly, the refinement of synaptic organization and preservation of higher nuclei in the face of decreased peripheral auditory stimulus is responsible for these differences. The age of onset of SHL and of the presence of fluctuating or persistent conductive loss is likely to influence these attributes. For these reasons, the age of onset of hearing loss or major progression is reported if known (Fig. 1). Lesion. Even in cases where there is a known specific origin or cause of deafness, the lesion often remains obscure. Temporal bone pathology has shed some light on specific causes and pathophysiology of human deafness, but has inherent limitations in evaluating current patients. Temporal bone imaging has allowed diagnosis of anatomical dysplasias of the otic capsule, but such anomalies are present in only a small fraction of deafness. Nuclear magnetic resonance imaging (MRI) reveals some central lesions, including eighth cranial nerve lesions, in the relatively unusual cases of hearing impairment with such disease. The usual lesion in deafness is cochlear and membranous, that is, on a cellular, ultrastructural, or molecular level within the peripheral auditory sense organ. Hair cell failure is a common end result of many primary insults, and at this time, only exceptional cases of 791

6 792 cochlear deafness bear distinguishing physical or laboratory findings. Specific information about the lesion, when known, does add valuable information when recorded as part of a diagnostic classification and therefore is listed explicitly. Etiology: intrinsic, extrinsic, or induced. Three basic etiological categories are designated. The first two categories, intrinsic and extrinsic, distinguish whether a specific cause for hearing loss has originated either from within or from outside the body without any implied age of onset. Thus a congenital CMV infection is an extrinsic cause of deafness, whereas congenital Trisomy 13 is an intrinsic cause of deafness. In this delineation, the term body means the inherent individual organism. Neonatal hyperbilirubinemia is considered an extrinsic cause of deafness for the same reasons that secretions inspissated to solid state are considered foreign bodies and that myoglobinemia from muscle trauma would be considered an extrinsic cause of renal failure. A third category, induced, is reserved for impairment produced by a specific set of intrinsic and extrinsic factors. In hearing, the best-known entity in this category is ototoxic SHL produced by even low doses of aminoglycosides in individuals with particular mitochondrial DNA mutations (e.g., 1555). It is likely that there are other induced causes of SHL which have not yet been described, particularly involving presbycusis, since it is well known that underlying cochlear damage in animal models predisposes to development of hearing loss with noise exposure, ischemia, and other extrinsic causes of hearing loss. It is also possible that some of the variation seen in the hearing acuity of patients with a common specific genetic origin is induced by extrinsic factors (e.g., Connexin 26). Etiology-specific medical diagnosis was considered. As has been discussed, the known causes and origins of deafness range from specific proximate causes (e.g., homozygous 35delG Connexin 26 mutations) to factors that statistically place an individual at high risk for SHL (e.g., prematurity). A diagnosis of genetic deafness has different degrees of certainty depending on the supporting evidence. Therefore etiology is preferably specified with supporting information. Genetic causes and origins are divided as to pattern of inheritance, autosomal dominant, autosomal recessive, X-linked, and maternal (mitochondrial). Where evidence is not sufficient for a full delineation, the most specific term retaining accuracy is used (e.g., recessive or autosomal). The etiology of SNHL has and will change according to the era and particular population in question. In the case of genetic mutations and maternal infections, this systematic nomenclature would facilitate routine documentation of these causes and origins whenever a diagnosis of hearing impairment is a part of patient records. In this manner, pertinent information can be extracted retrospectively. Concerning transmissibility and family risks, the pattern of inheritance for known genetic causes of hearing loss are given, along with the basis for attributing a diagnosis of genetic hearing loss, in this nomenclature because of the importance of the diagnosis for other family members. For example, a diagnosis of autosomal recessive SHL ranges from likely, in the case of an otherwise unremarkable child whose exhaustive workup is negative, to unless proven otherwise, in the case of siblings with idiopathic deafness and hearing parents, to proven, in the case of positive DNA tests for gene mutations with an expected clinical presentation. The diagnosis of genetic hearing loss carries consequences for future siblings and future children and therefore is best described explicitly with supporting information so that the certainty of the assessment can be judged. When a specific gene mutation is diagnosed, the term mutant is not appropriately applied to a patient (unlike a mouse). The term used in this system is deafness gene alleles. The mapped locus, gene, and sequence mutation are named, if known. Although the term allele is not common in lay language at present, it is an easy word, and it is scientifically more accurate than mutation, since many inherited changes in nucleotide array are polymorphisms that do not code for differing traits. More important, this term bears no pejorative connotation. At best, terms of medical nomenclature are informative and aid communication. Words that are used with different meanings in other contexts (particularly pejoratively) can impede rapport with both patients and family members, as well as interfere with professional communications. Associated physical and laboratory findings, including present or absent physical markers and distinct positive findings, are listed at the end of the description. Such findings include syndromes, unique syndromic anomalies, and isolated findings. PATIENTS AND METHODS Aims and Objectives The value of the newly constructed SCALE nomenclature was tested by classifying a cohort of hearing-impaired children. The resulting data were used to answer two clinically pertinent questions: 1. What specific etiologies account for deafness? 2. Which aspects of the otological evaluation were diagnostic of a specific etiology? Patient Population and Study Design All patients in the study were examined by the author in private consultation. Parents and children were interviewed, and past medical records were requested and reviewed. Follow-up data were obtained where possible. Entry criteria included sensory threshold at or above 30 db in at least one frequency from 250 to 4,000 Hz. The lower limit of 30 db was chosen to avoid borderline cases. Exclusion criteria included age at first visit of 19 years or more, insufficient clinical follow-up or audiological information to establish veracity of sensory hearing loss, or further follow-up establishing that the original finding of SHL was an audiological misdiagnosis. Serial inspection of paper office files, performed twice sequentially, yielded 189 patients meeting entry criteria. First visits occurred over an 8-year period ending in early Sixteen patients were excluded because available records were insufficient to confirm sensory loss, five children were excluded because further evaluation showed that SHL was an audiological misdiagnosis, and the remaining 168 children were entered for clinical classification of hearing loss and analysis of otological evaluation.

7 RESULTS There were 83 boys and 85 girls in the subject population, yielding a male-to-female sex ratio of 1.00:1.02. Age at entry ranged from 2 weeks to 18 years. Sidedness There were 137 cases of bilateral (82%) and 31 cases of unilateral SHL (18%) in the patient cohort. Of the bilateral cases, 18 of 137 (13%) were classed as asymmetric. Component Acuity The component acuity ranged from profound to mild hearing impairment. 1. Profound impairments. Fifty-seven patients had a profound sensory loss classed on at least one side at entry. Bilateral profound SHL was present in 33 patients (28% of all cases with bilateral SHL). In unilateral SHL, the impaired side was classed as profound in 13 patients (13/31), comprising 42% of all patients with unilateral SHL. 2. Severe impairments. A total of 36 patients had a severe sensory loss in at least one ear at entry. Bilateral severe SHL was present in 24 patients (20% of all cases with symmetric bilateral SHL). A severe SHL was present in three patients with unilateral SHL (10% of all unilateral cases) 3. Moderate impairments. Moderate impairments were found in at least one ear in 59 patients at entry. Of the patients with bilaterally symmetric class impairment, 30% of cases were moderate. Unilateral moderate SHL was present in 10 cases (32% of all unilateral cases) 4. Mild impairments. Mild impairments were found in at least one ear in 46 patients. Bilateral mild SHL was found in 29 children (23% of all bilateral cases). However, many of these children had significant periods with overlying conductive losses, so the auditory impairment was greater than such a sensory loss might imply. Of patients with unilateral SHL, five had mild losses (16% of unilateral cases). Age of Onset Only one child in the group (identified in the newborn nursery) had a proven congenital onset of sensory hearing loss. In all intrinsic and idiopathic cases, a congenital age of onset was assumed unless there was strong objective evidence to the contrary. Evidence for late onset occurred in eight patients, as documented by prospective audiological testing and attainment of speech and language milestones. In these children, there was antecedent electrophysiological testing documenting normal thresholds. Children with extrinsic causes related to neurological disease ordinarily had a known age of the onset of SHL; these included cases of head trauma, brain tumors, and cranial irradiation. Congenital CMV cases are well known to have both progression and late onset. It was not clear when the hearing loss began in the particular cases of congenital CMV in the series. However, a congenital onset was assumed. There were two cases of sudden unilateral losses in older children that were idiopathic. Of 18 cases classified as showing major progression, six had either an idiopathic or autosomal origin, three were proven to be due to congenital CMV infection, (3/4 CMV patients) and six had cochlear dysplasia or WVA syndrome. One patient was diagnosed as having apparent X-linked stapes gusher, and one child received large amounts of chemotherapy and cranial irradiation to the head with onset of profound bilateral loss more than 1 year after finishing treatment. Lesions Specific lesions causing or intimately associated with childhood SHL were found (or presumed) on the basis of a specific diagnosis in 16% patients as follows: 1. Gross anomalies of temporal bone involving peripheral auditory organ. Otic capsule dysplasias were present in 11 cases (6% of all children). As temporal bone imaging was not performed uniformly and was of varying quality, this percentage does not represent the incidence of gross anatomical anomalies in the group. Since these anomalies are known to be associated with unstable residual hearing thresholds, the high incidence in this series is probably due to selection bias. Many children were referred by audiologists for medical evaluation because hearing appeared to diminish. Review of available scans indicated widened vestibular aqueduct in seven children (cases 9, 16, 25, 58, 77, 85, and 86), partial Mondini deformity with widened vestibular aqueduct in two (cases 5 and 27), and common cavity cochlear dysplasias in two boys, one with Trisomy 13 (case 21) and the other unilaterally in a healthy child with normal hearing in the contralateral ear (case 22). None of these children had a known extrinsic cause. 2. Eighth cranial nerve disease or cochlear nucleus damage. Eighth cranial nerve palsy (or damage to the ipsilateral cochlear nucleus) was present in three patients (cases 91, 131, and 158 [2%]) on the basis of an expanding posterior fossa mass or as a consequence of neurosurgical manipulation of a posterior fossa mass, in all cases. No child had an acoustic neuroma in this series. These patients were all referred by pediatric neurosurgeons or neurologists. 3. Labyrinthitis. Labyrinthitis was presumed on the basis of specific clinical diagnosis and history in all cases; eight children (5%) were assessed as having this lesion. Categories included the following. Radiation-induced disease included three cases of chronic ear disease, sensory loss, and conductive overlay that occurred after cranial irradiation with active otitis media (cases 87, 102, and 154). In one case (case 154), MRI and temporal bone CT supported the diagnosis. Suppurative bacterial disease was present in one patient with a history of bacterial meningitis and profound sensory losses (case 45). Inflammatory immune response included all four cases of congenital CMV. 4. Ultrastructural cochlear lesions included abnormal Connexin-26 gap junction function (3%). Etiology Etiologies were divided into intrinsic and extrinsic classes. 793

8 I. All intrinsic causes accounted for 40% of all patients (67/168). 1. Genotypic cause was present in 57 of 168 patients (one-third [34%]) (5% chromosomal aneuploidy, 29% single-gene related alleles). 2. Epigenetic syndromes (no known extrinsic cause) were present in 10 of 168 patients (6%). 3. Aneuploidy was present in 5% of all cases (6% of all bilateral cases). There were eight children with abnormal numbers of chromosomes. The most common human autosomal aneuploidy, Down syndrome, accounted for more than half of patients in this etiological subgroup (5/8 [62%]). The most common sex chromosomal aneuploidy, Turner syndrome, was present in one patient. The remaining two patients were unusual: a rare survivor of nonmosaic Trisomy 13 and an apparently healthy child with Trisomy 8. Together, these patients accounted for 5% (8/168) of the cohort. All children with abnormal chromosome number had bilateral sensory losses, accounting for nearly 6% of total bilateral cases. 4. Single-gene related deafness was present in 48 of 168 patients (29%). a) Autosomal recessive type was present in 31 of 168 patients (18%). Only four children in this group had a specific gene-related deafness diagnosed, showing a Connexin 26 35delG homozygous status in all cases. In the remaining 27 children, there was a family history of unaffected parents and an affected first-degree relative (including a sibling of one child with positive Connexin 26 results.) Some of the cases classed in this group, in which the affected siblings or other first-degree relatives were all male, could be due to an X-linked recessive pattern of inheritance. b) Autosomal dominant type (proven by family history and/or recognizable syndrome) was present in 12 of 168 patients (7%). c) Autosomal (recessive or dominant with incomplete penetrance) type was present in 4 of 168 patients (2%). d) Autosomal dominant (proven by family history and/or recognizable syndrome) type was present in 12 of 168 patients (7%). Four children with bilateral impairment had apparent autosomal dominant deafness on the basis of family history of childhood deafness with similar audiometric features. One child was the second affected generation, one case the third, and two cases were the fourth affected successive generation (cases 100 and 132). Three children with unilateral impairment had similar histories. Five children had recognizable syndromes involving autosomal dominant genes. Four had Waardenburg type 1 (PAX) and one had fibroblast receptor protein generelated Crouzon syndrome. e) Autosomal (recessive or dominant with incomplete penetrance) type was present in 4 of 168 patients (2%). f) X-linked single gene inheritance was present in 3 of 168 patients (2%). One patient (case 70) had a diagnosis of Opitz syndrome, also known as the hypospadias dysphagia syndrome or telecanthus with associated abnormalities 794 (midline abnormalities such as cleft lip, laryngeal cleft, heart defects, hypospadias, and agenesis of the corpus callosum). Although there are both X-linked and autosomal forms of Opitz syndrome, anteverted nares and posterior pharyngeal cleft are seen only in the X-linked recessive form of the syndrome (gene locus Xp22) and were present in this child. 19 There were two cases of X-linked dominant syndromes; the first (case 74) was a variant of X-linked perilymphatic gusher in a teenage boy who had a history of progressive SHL with increasing conductive overlay and who had lost response to sound in one ear after stapedectomy complicated by perilymphatic gusher. His young adult mother had a long-standing mild SHL, as expected in this syndrome. However, his temporal bone CT did not show any of the typical findings in the main form of this syndrome. Case 127 was a female heterozygote with the X-linked Coffin-Lowry syndrome (Fig. 1). 5. Congenital syndrome (epigenetic or sporadic, no known extrinsic cause). Ten children (6%) had named syndromes not known to be caused by any extrinsic factor (see discussion). In some families, these syndromes are linked to chromosomal sites but are not considered to be uniformly caused by gene alleles. These intrinsic syndromes included syndromes of Goldenhar (three cases), Noonan (three cases), CHARGE (two cases), and VATER (one case), as well as diaphragmatic hernia and omphalocele (1 case). Although some cases of Noonan syndrome have been shown to be an autosomal recessive, single-gene abnormality in several families (12q24), none of the cases reported here were familial or documented to have reported loci abnormalities. Similarly, cases of oculoauriculovertebral dysplasia are usually sporadic, but a few families consistent with autosomal recessive inheritance have been reported, and other families clearly support autosomal dominant inheritance. 6. Autoimmune causes. Western blot testing for antibodies against the 68-kD bovine inner ear antigen were carried out on patients entered with a question of progression. Tests were performed either at the research laboratory of Dr. Jeffry Harris at the University of California at San Diego or at the Immunodiagnostics Company (Buffalo, NY). None of the test results were positive. One boy (case 153) had been treated with multiple courses of steroids for presumed autoimmune hearing loss. The definitive origin in his case was Waardenburg syndrome type 1. II. Extrinsic causes accounted for known etiology in 10% of all patients (16/168 cases). 1. Congenital and neonatal infections accounted for 26 of all cases. a) Cytomegalovirus (4/168 [2%]). There were four cases in which CMV could be attributed as a specific diagnosis; all of these cases were bilateral and three showed major progression. 2. Bacterial meningitis (3/168 [2%]). There was documentation of culture results in only two cases (pneumococcal in both cases).

9 3. Hyperbilirubinemia (1/168 [0.6%]). The child assigned this cause had a total serum bilirubin value of greater than 15 and did not receive aggressive care. He was otherwise unremarkable. 4. Noise trauma. No child had a history of noise trauma that might account for deafness. The possibility that high gain amplification was responsible for minor progression in at least some cases cannot be refuted or proven. Noise may have had a role in one case (case 142 [unilateral SHL, moderate Hf]) where loss occurred in motor vehicle accident; however, head trauma was severe and was the assigned origin. 5. External radiation to temporal bone (1/168 [0.06%]). The clinical presentation and temporal bone scans indicated osteitis, probably a form of radiation osteitis, in one patient (case 165), who underwent cranial irradiation 12 years before entry with otitis media for sole treatment of a brain tumor. There was bilateral chronic ear disease. 6. Combination radiation and ototoxic chemotherapeutic drugs (4/168 [2%]). Four children (cases 40, 87, 102, and 154) had sensory losses secondary to treatment for malignant tumors that included cranial irradiation and potentially ototoxic chemotherapy (cisplatin or carboplatin). In two of these children there was long-term middle ear disease (after treatment) producing conductive overlay. Both had otitis media at the time of incidental temporal bone irradiation that occurred secondary to tumor irradiation. 7. Posterior fossa tumor or cyst, directly, or complication from surgery for excision (3/168 cases [2%]). 8. High-risk extrinsic factors were implicated in 4% of cases, without revealing a clear etiology. a) History of meconium aspiration (3/168 [2%]). Meconium aspiration was documented in three cases through hospital records. b) Prematurity/low birth weight (3/168 cases [2%]). There were three cases of prematurity with very low birth weight; all involved multiple risk factors including anoxia and aminoglycoside administration. DISCUSSION Etiology General etiologies could be designated by diagnostic evaluation in 107 cases (64%). Intrinsic etiologies were most common, accounting for 40% of all cases. Abnormalities of chromosomal number (aneuplodies) were present in 5% of the cohort and included two diagnoses known to be commonly associated with SHL: Trisomy 21 and Turner syndrome. 20 The most common single etiology, intrinsic or extrinsic, was autosomal recessive deafness designated on the basis of either proven homozygous mutations (Connexin 26 35delG) or family history (affected sibling, unaffected parents). Autosomal dominant hearing loss on the basis of family history was not uncommon in this series, accounting for 3% of patients with bilateral SHL and 10% of the children with unilateral SHL. Waardenburg syndrome type 1 was found in four patients and was often misdiagnosed because either intercanthal distance showing dystopia canthorum was not recognized as diagnostic of Waardenburg syndrome type 1 (Fig. 2) or minor pigmentary anomalies in other family members were inaccurately presumed to be diagnostic (e.g., case 126). Of congenital syndromes not due to a single gene anomaly, Noonan and Goldenhar Syndromes were the most common, each accounting for 2% of all patients. 21,22 There were 61 cases that were classed as idiopathic (36%), including 20 that were likely to be due to autosomal inheritance by exclusion. This yield is comparable to other analyzed groups reported during the same era in that the single largest group was idiopathic and the next largest was of genotypic origin. 2,3 Because congenital CMV infections are asymptomatic in about 90% of cases and 41 children lacked tests Fig. 2. Unilateral SHL, progressive lowfrequency loss, positive for Waardenburg syndrome type 1. Patient s diagnosis (case 115) of Waardenburg syndrome type 1 was not given until the time of her entry visit (left). Dystopia canthorum, the single pathognomonic feature of the most common type of Waardenburg syndrome, has been evident since infancy (right) in this attractive child. The pupils are a normal distance apart, but the medial canthi are wide set. There are no obvious nonpigmented areas of skin or hair. Diagnosis was subsequently confirmed on formal genetics consultation. 795

10 excluding this cause, it is likely that many of them had sensory hearing impairment on the basis of this extrinsic cause. 23 Only four children were diagnosed with congenital CMV as a cause of deafness, and two of them were known to be infected as neonates. The natural history of congenital CMV infection indicates that sensory hearing loss, usually isolated, occurs in about 15% of children infected at birth by age 6 years. 8 Deafness is known to be caused in many individuals by a combination of a genetic predisposition with an external environmental exposure. Although SHL associated with mitochondrial mutations is well described, no patients were identified with mitochondrial DNA related deafness. Since no testing was performed for mitochondrial mutations, the true incidence of that genetic cause is unknown in the group. Two patients with idiopathic disease had muscle weakness associated with fevers. In both cases the initial episode was temporally associated with diphtheria-pertussis-tetanus (DPT) immunization. Testing for mitochondrial mutations in these patients is planned. Molecular diagnostic tests for the 1555 mutation in children with a general origin of extrinsic deafness secondary to high-risk factors who were exposed to therapeutic doses of aminoglycosides might give a specific diagnosis in some cases. The spectrum and weighted representations of causes in this series of referred patients are strongly influenced by selection bias. The author saw most of these children for evaluation of either newly diagnosed hearing loss (sometimes made at entry) or because of changing audiometric test results. Neither cochlear implantation nor excision of acoustic neuroma was performed during the study period. Although children were referred from many sources, the most common referral source of hearing-impaired children was from audiologists at four separate programs for hearing-impaired children. The children with deafness from posterior fossa tumors were referred by pediatric neurosurgeons whose primary interest is in brainstem tumors. The paucity of children with bacterial meningitis referred is probably due to the lack of uncertainty as to their diagnosis. Specific Workup The two most useful tests were laboratory screening for Connexin 26 mutations and CT of the temporal bones. All molecular diagnostic screening was performed at another university under the direction of a molecular otologist. Three percent of cases (5/167) were diagnosed by screening for Connexin 26 mutations. The epidemiology of single-gene mutations associated with SHL is beginning to be elucidated; one of the most common known deafness genes codes for a gap junction protein (Connexin 26). Recommendations for gene testing were only made in the latter part of the study period; fewer than 20 patients have been screened. The true incidence of Connexin 26 gene related hearing loss is not known in this cohort. However, this is a common form of deafness worldwide, and laboratory screening is now practical for clinical diagnosis. 24, Four children were identified as homozygous for the 35delG mutation. In most cases the establishment of a specific diagnosis was clinically invaluable. The sister (case 20) of one girl (case 21) who tested homozygous for Connexin 26 35delG was not individually tested, but can be reliably given a specific diagnosis of Connexin 26 gene related deafness on the basis of her sibling s positive test result. Although the etiological factor of autosomal recessive deafness was already known in these sisters by family history, her parents expressed a tremendous relief when the specific diagnosis was made. The knowledge that the carrier status of the third normalhearing daughter (and that of any future son-in-law) could be determined was perceived as a loss of helplessness. At entry, the parents of one patient (case 90) reported that he had Waardenburg syndrome. This diagnosis first appeared in the patient s records as a probable diagnosis in a pediatric endocrinology consultation summary letter and was then assumed to be the proven or probable diagnosis in later notes from many sources. The basis of that diagnosis was the presence of a patch of depigmented scalp hair in the child s brother. This raised a natural question in the parents minds of who of the two of them was affected by this dominant gene, and responsible for passing it to both their sons. Additional opinions raised doubt concerning the certainty of the diagnosis, but did not dispel concerns entirely, as Waardenburg type 2 was certainly possible. Case 90 had been classed as idiopathic in the study records because CMV was not ruled out as a possible diagnosis; immunoglobulin G (IgG) CMV titers were positive and urine culture for CMV results were not available. The clear confirmation of Connexin 26 gene related deafness was helpful for every member of this family. There was also prognostic value for the child s hearing. Although progression has been reported in Connexin 26 gene related deafness secondary to homozygous 35delG mutations, it is not usual. On the other hand, major progression occurs commonly in both congenital CMV and Waardenburg syndrome related deafness. Imaging Temporal bone imaging was helpful in making a diagnosis of an anatomical lesion in 10 cases, widened vestibular aqueducts with or without Mondini dysplasias and a common cavity cochlear dysplasia. CT scan was preferred because of the fine detail of the temporal bone image, unless the patient s cause was likely to be primarily neurological. In case 91, for example, CT of the temporal bone was nondiagnostic, but MRI showed expansion of a posterior arachnoid cyst that accounted for the new unilateral SHL. Neurosurgical revision was prompted by that study. In one case, the finding of dilated vestibular aqueducts was clinically important. The cause in case 77 was idiopathic before CT. This exceptionally intelligent 5-year-old girl had a moderate hearing loss, just diagnosed. Her parents had adopted her newly born, knowing the history of maternal drug use. Her increasing difficulty with language and recent change in behavior were postulated as evidence of previously unsuspected neurological damage by several professionals. The diagnosis of WVA