Waardenburg syndrome (WS) is one of the more common genetic causes of deafness and accounts for about 2% of the congenitally deaf [Milunsky, 2007]. This autosomal disorder is characterized by varying degrees of hearing loss associated with pigmentary disturbances of the eye, skin and skin appendages [Milunsky et al., 2007; Waardenburg, 1951]. One of the more distinctive features is a prominent white forelock. The presentation of WS is quite variable and attempts have been made to categorize subtypes on the basis of clinical features. It was first divided into two major subgroups based on the presence (WS type I) or absence (WS type II) of dystopia canthorum (lateral displacement of the inner canthi) [Arias, 1971]. As additional clinical features that distinguished different families were described, the classification scheme was further broadened to include four subtypes. WS type 2 remains the “pure” subtype. The presence of musculoskeletal abnormalities separates WS type 3 (Klein-Waardenburg syndrome) from WS type 1 [Klein, 1983]. Skeletal malformations in WS type 3 range from minimal contractures of the fingers to hypoplasia of the upper limbs and/or pectoral girdle [Hoth et al., 1993; Klein 1983; Senrui 1984; Tekin et al., 2001], and in some cases there is mental retardation. WS type 4 (Shah-Waardenburg syndrome) combines the features of WS type 2 with the presence of Hirschsprung disease [Shah et al., 1981]. A rare subtype, craniofacial-deafness-hand syndrome (CDHS, OMIM 122880) [Sommer et al., 1983; Asher et al., 1996; Sommer and Bartholomew, 2003] was described in a three-generation family that was evaluated several times over a period of 24 years. Affected individuals have severe sensorineural hearing impairment and show clinical features of flat facial profile, hypertelorism and dystopia canthorum, downslanting palpebral fissures, depressed nasal bridge, small “pursed” mouth, ulnar deviation and contractures of the hand, and unique imaging findings of absence or hypoplasia of the nasal bones and a hypoplastic ulnar styloid. The imaging findings and distinct facial features was felt to distinguish CDHS from WS type 3.

Six genes responsible for WS have been identified, that allow objective identification of affected individuals. Heterozygous mutations in the paired box gene 3 (PAX3) were first reported in both familial and sporadic cases of WS type 1 [Baldwin et al., 1992; Hoth et al., 1993; Wollnik et al., 2003] and are responsible for the majority of cases. It was later shown that WS type 3 is an allelic disorder caused by heterozygous, homozygous or compound heterozygous mutations in PAX3 (e.g., [Hoth et al., 1993; Read and Newton, 1997; Zlotogora et al., 1995; Tassabehji et al., 1995; Tekin et al., 2001; Wollnik et al., 2003]. CDHS is also an allelic variant, and may be caused by PAX3 mutations affecting the paired box domain [Asher et al., 1996]. More than 50 point mutations and 20 unique deletions and insertions spread throughout the PAX3 gene have been reported (http://www.hgmd.cf.ac.uk/docs/login.html accessed February 4, 2008). The majority of these mutations are unique to a specific family or person. Even within a single family there can be marked clinical variability.

WS type 2 has been associated with abnormalities in the microphthalmia-associated transcription factor (MITF) gene [Hughes et al., 1994; Liu et al., 1995] or the encoding snail homolog 2 (SNAI2) gene [Sanchez-Martin et al., 2002]. Missense, splicing and small deletion mutations but no rearrangements have been detected in MITF (http://www.hgmd.cf.ac.uk/docs/login.html accessed February 4, 2008) and homozygous deletions of SNAI2 have been found in two families [Sanchez-Martin et al., 2002], although the molecular etiology of most cases of WS type 2 is unknown. WS type 4 is heterogeneous and is associated with mutations in EDNRB (endothelin receptor type B) [Puffenberger et al., 1994; Syrris et al., 1999], EDN3 (endothelin-3) [Baynash et al., 1994; Edery et al., 1996], and SOX10 (sex determining region y-box-10) [Pingault et al., 1998]. Deletions in SOX10 were recently identified in patients with WS types 2 and 4 [Bondurand et al., 2007].

She underwent several surgeries including tonsillectomy and adenoidectomy at the age of 5, plastic surgery on her eyelids to correct ptosis at the age of 6 and repair of a septal nasal deformity and airway obstruction at the age of 26. Presurgical imaging studies and medical records were not available for review. She denied constipation or other gastrointestinal problems suggestive of Hirschsprung disease. Family history revealed no mental retardation, skeletal abnormalities or white forelock and no other cases of hearing loss in three generations.

On physical examination her height was 174 cm and her weight was 65.7kg, both above the 75^th centile. Her physical examination was remarkable for a flat facial profile, a short broad nose, depressed nasal bridge, microgenia and retrognathia, and hypertelorism; the innercanthal distance was 4 cm, i.e. 2 standard deviations above the mean (Figs 1A,B), and she had dystopia canthorum with a W index of 2.27 cm. Her philtrum was normal and not smooth. There was no white forelock or evidence of any skin depigmentation. Her hair showed some graying at the age of 34 (within the normal range for Caucasians). Her eyes were normal, without heterochromia of the irides. Her ears were of normal shape and position without pits or tags. She had a high arched palate with normal uvula. The upper front teeth were replaced by a bridge to correct severe malocclusion (not shown). No goiter or other palpable thyroid abnormality was detected. There was limited movement of the wrist with ulnar deviation of the hands and mild flexion contractures of digits 2 through 5 (Fig 1C). In addition, there was clindactaly of the 5^th digits. At the time of consultation, she had profound hearing loss and was adept at lip-reading. Her speech cadence and tone was monotonic and metallic.

Does our patient have a variant of WS? She shares some but not all the characteristics of CDHS. Her hand and wrist radiographs are normal and CT scan of the face did not show absence or hypoplasia of the nasal bones, but revealed underdevelopment of the sinuses ranging from mild hypoplasia to complete aplasia, features not previously associated with WS. To our knowledge only one family with CDHS has been published previously [Asher et al., 1996] and CT imaging was not performed in any of the affected individuals. Therefore, it is not possible to know whether the bone abnormalities observed in our patient are also present in them.

A search of the English literature did not yield previous reports of a syndrome with the combination of abnormalities seen in our patient. The cochlear malformation observed in our patient is a form of Mondini deformity that is associated with sensorineural hearing loss in Pendred syndrome, but is more often seen as an isolated abnormality. Although similar abnormalities have been described in WS (e.g., [Oysu et al., 2001], most of these patients do not have macroscopic inner ear malformations that are visualized by CT or MR imaging. There is one report of a man with similar sinus abnormalities discovered during evaluation of sinusitis symptoms [Haktanir et al., 2005]. No other description of the patient was provided beyond the statement that “no skeletal, systemic or hematological abnormality was detected on clinical and laboratory examinations.”

As is increasingly found for heterogeneous disorders in general, categorization on the basis of molecular etiology will likely be more useful than eponymous terminology or grouping by clinical features. For example, different PAX3 mutations in the paired box domain have been identified in the previously reported CDHS family [Asher et al., 1996] and in families with WS type 3 [Tekin et al., 2001]. Our patient does not carry these particular mutations or another coding or splice site sequence alteration in PAX3. Point mutations have been identified in more than 90% of people with WS type 1 or WS type 3 [DeStefano et al., 1998; Hoth et al., 1993], but a recent study found whole or partial PAX3 gene deletions in approximately 10% of affected individuals [Milunsky et al., 2007]. Partial or full SOX10 deletions have also been identified in WS types 2 and 4 [Bondurand et al., 2007]. The presence of heterozygosity for a PAX3 polymorphism rules out a full gene deletion, but a partial deletion remains a possible disease mechanism in our patient. PAX3 cDNA could be examined to look for deletion of several exons or abnormal splice forms and a copy number variation might be detected by array technology.