Clinical manifestations

Presentation and course

Although discrepancies are present among authors concerning the definition of Dandy-Walker syndrome, the following features are not generally contested:

At birth, most infants are asymptomatic or mildly symptomatic. Macrocephaly is the most common presenting sign, and evidence of increased intracranial pressure dominates the clinical presentation. In about 80% of cases the diagnosis is made during the first year (32; 50; 51); however, some patients first come to clinical attention as adults (48; 62). Unusual presentations in the neonatal period are occipital encephalocele and apnea. Although all patients have cerebellar vermis hypoplasia, cerebellar dysfunction evolves in only about one half of children (27). Nystagmus may be present, but cranial nerve signs are exceptional. Intellectual impairment is found in at least one half of patients. In the series reported by Gerszten and Albright (n=20), 45% of patients had moderate to severe cognitive impairment. There was no relationship between volumetric measurements of the posterior fossa structures and the degree of cerebellar dysfunction and mental retardation. In the series (n=20) reported by Boddaert and colleagues in 2003, all children with normal vermian lobulation and without supratentorial anomalies had normal intellectual outcome. Cognitive impairment is not due to increased intracranial pressure but to impaired cerebellar function. Increasing evidence provides that the cerebellum is important for motor and cognitive learning (63). Agenesis or hypogenesis of the corpus callosum is found in about one sixth of Dandy-Walker syndrome patients. Various cardiovascular anomalies (eg, septal defects, tetralogy of Fallot, dextro cardia, and coarctation of aorta) have been noted (47; 32; 48). Other associated neuropathologic and neuroimaging findings and systemic manifestations are listed in Table 2.

Table 2. Reported Associated Findings in Dandy-Walker Syndrome

Central nervous system

Systemic

Based on: (45; 26; 51; 57; 61).

It is relevant to stick to strict diagnostic criteria to avoid confusion. A review on “common comorbidities in Dandy-Walker syndrome” is misleading, as it includes patients with Dandy-Walker syndrome, Dandy-Walker variant, mega cisterna magna, defined syndromes (as Joubert), and chromosomal aberrations (66).

Prognosis and complications

The overall prognosis depends mainly on the following:

• Possible associated findings (eg, cardiac malformations, additional CNS malformations, and chromosomal aberrations). Type and number of associated anomalies have impacts on survival and outcome (57; 61). Intellectual prognosis is considered favorable if vermian lobulation is preserved and supratentorial anomalies (in particular callosal dysgenesis) are absent (08). Remarkably, this experience, based on 20 patients, has not been addressed in further studies.

Along the line of the importance of clear diagnostic criteria, one has to be careful to imply prognostic implications from article titles: this point is exemplified quoting 2 articles on “schizophrenia” and “psychosis,” respectively, in patients with inferior vermis hypoplasia, not Dandy-Walker malformation (20; 65).

Bolduc and Limperopoulos drew attention to limited methodology in many neurodevelopmental outcome studies in children with cerebellar malformations, including Dandy-Walker malformation (10). Outcome is less favorable in series with prenatal diagnosis due to higher frequency of associated findings including chromosomal aberrations (22; 34). A study on “the natural history of the Dandy-Walker syndrome in the United States” is not informative as it also includes patients with mega cisterna magna, Blake’s pouch, and Dandy-Walker variant (41). A study from Montreal, Canada, on outcomes in children with congenital cerebellar malformations includes data on 10 patients with Dandy malformation (53). At a median age of 13:6 years, 7 children had global developmental delay; for 8 of 10 patients, no exact information about cognitive abilities is available (too young for reliable testing, information not known). Six of 10 children had associated cerebral malformations. A selection bias towards more severely affected patients is possible.

• Complications related to shunt dysfunction

Even in centers with considerable expertise, the mortality of nontumoral infantile hydrocephalus during childhood is in the range of 10%. Sudden unexpected death in patients with Dandy-Walker syndrome not related to shunt problems has been observed (23). In an earlier, large Dandy-Walker syndrome series (operated before 1970), the overall mortality was as high as 24% (48).

An occasional late complication is the development of syringomyelia, resulting from a herniation of the Dandy-Walker cyst through the foramen magnum (31).

Clinical vignette

Case 1. The proband was the first child of healthy nonconsanguineous young parents. The pregnancy was uneventful. Prenatal ultrasound investigation at 32 weeks' gestation revealed a posterior fossa malformation. Dandy-Walker syndrome was diagnosed. Spontaneous delivery at 38 weeks was uneventful. Clinical examination was normal, as was renal ultrasound. Head circumference was 33 cm (50th to 90th percentile). Dandy-Walker syndrome was confirmed by MRI.

Follow-up examination at 10 months disclosed marked developmental delay with milestones corresponding to 4 to 5 months. Head circumference was 43 cm (10th percentile). There was a convergent squint. Clinical examination was otherwise normal.

Case 2. The diagnosis of Dandy-Walker syndrome was made 4 weeks before delivery. Head circumference at term birth was 40 cm, rapidly increasing postnatally. A ventriculoperitoneal shunt was inserted at 5 weeks. At 2 years of age, an additional cyst drainage was installed due to further increase of head size and a history of occasional vomiting. Subsequently, the patient did well, having only minor truncal cerebellar ataxia. When last seen at 7 years of age, cognitive abilities were normal for her age. She has a capillary nevus of her right frontal region.

Biological basis

Etiology and pathogenesis

The etiology of Dandy-Walker syndrome is poorly understood and probably multifactorial. Dandy-Walker syndrome can be part of defined Mendelian disorders and other syndromes (listed in Table 1). Dandy-Walker syndrome can also be found in association with various chromosomal aberrations (45; 11; 72; 22). However, in the majority of patients diagnosed postnatally, Dandy-Walker syndrome is not associated with other syndromes and chromosomal aberrations; it is an isolated finding.

The pathogenetic mechanism of Dandy-Walker syndrome is still speculative (59). In a broad sense, the disease appears to result from a developmental hindbrain disturbance occurring in early embryonic life, perhaps in the sixth to the seventh week. The initial theory relating Dandy-Walker syndrome causally to impairment of the fourth ventricle outlets has been questioned in light of demonstrated foraminal patencies in a considerable proportion of patients. The dilatation of the fourth ventricle is most likely due to persistence of the anterior membranous area that forms the roof of the early fourth ventricle without the normal regression and disappearance as the choroid plexus and vermis develop (12).

Genetic investigations and animal studies on mouse embryos have drifted the “mechanical” hypothesis towards the interplay of the developing cerebellum and the posterior fossa mesenchyme, at least in a subset of patients. Grinberg and colleagues identified 7 children with Dandy-Walker malformation and de novo interstitial deletions of chromosome 3q24-3q25.33 (28). This region includes 2 Zinc finger genes, ZIC1 and ZIC4. Mice with a heterozygous deletion of these 2 linked genes have Dandy-Walker malformation-like cerebellar hypoplasia, suggesting that heterozygous loss of ZIC1 and ZIC4 is causing Dandy-Walker malformation in individuals with 3q deletion. By reporting on 3 additional patients from a larger group of individuals with Dandy-Walker malformation, Ferraris and coworkers expanded the knowledge of the heterogeneous genotype-phenotype correlation of 3q deletions encompassing ZIC1 and ZIC4 mutations (25). They concluded that homozygosity for ZIC1-ZIC4 genes is neither necessary nor sufficient per se to cause this condition. One patient with developmental delay and dysmorphic features had normal cerebellum and posterior fossa structures. It is debatable whether this constellation with normal neuroimaging should be called Dandy-Walker malformation.

A second locus on chromosome 6p25.3 linked to Dandy-Walker malformation was identified by Aldinger and colleagues. They showed that deletions or duplications encompassing FOXC1 are associated with cerebellar and posterior fossa malformations, including cerebellar vermis hypoplasia, mega cisterna magna, and Dandy-Walker malformation (01). In 6 individuals (3 fetuses) with 6p25 deletions, including FOXC1, various combinations of cerebellar and ocular (anterior chamber defects consistently, and others) malformations were confirmed (21). Studies on mouse embryos have further elucidated the interaction of the developing cerebellum and the posterior fossa mesenchyme, “communicating” with each other as they develop. Foxc1 is widely expressed in the extracellular matrix surrounding the cerebellar anlage. Foxc1-deficient mice have a lower number of cerebellar radial glial cells, disrupting cerebellar development (30).

Haldipur and colleagues have expanded their studies on mice models and comparisons with observations in humans with Dandy-Walker malformation (29). They have demonstrated that hypomorphic Foxc1 mutant mice have granule and Purkinje cell abnormalities causing subsequent disruptions in postnatal cerebellar foliation and lamination. A partially formed posterior lobule echoes the posterior vermis “tail sign” observed in human neuroimaging. In a large exome sequencing study of 100 families with either cerebellar hypoplasia or Dandy-Walker malformation, remarkable differences were found between the 2 cohorts and between the subgroups of Dandy-Walker malformations with or without an unpaired caudal lobule or “tail sign” (02). Among 36 individuals with sufficiently detailed neuroimaging, 30 patients had a “tail sign.” The rate of a genetic diagnosis was lowest in this group. This observation underscores the potential diagnostic significance of the “tail sign.”

A third locus on chromosome 8p21 has been identified in a patient with a de novo 2.3-Mb deletion encompassing 19 genes (76). FGF17 (Fibroblast Growth Factor 17) was downregulated in peripheral lymphocytes and skin fibroblasts. There is insufficient data to yet determine whether FGF17 alone causes Dandy-Walker malformation. In mice loss of fgf17 causes vermis hypoplasia and foliation abnormalities.

Based on studies in 3 fetuses with Dandy-Walker malformation (plus intrauterine growth retardation and other congenital anomalies) with 2 de novo loss and 1 de novo gain of the chromosome 7p21.3 region, a novel locus at 7p21.3 was postulated (37). The critical region includes NDUFA4 and PHF14, but mutation analysis was not reported. This finding needs to be confirmed, particularly as the overall constellation is beyond "pure" Dandy-Walker malformation.

The OMIM (Online Mendelian Inheritance in Man catalogue) lists "Dandy-Walker malformation with encephalocele, autosomal dominant" (%609222) based on previous publications. Darbro and coworkers reexamined 2 previously reported pedigrees (19). They found a nonsense mutation in the NID1 gene in the Vietnamese family, but not in the Indian family. However, further analyses of genes interacting with NID1 led to identification of a heterozygous missense mutation in LAMC1 in the second family. Remarkably both genes are "players" involved in the extracellular matrix, which is considered essential in the developing posterior fossa mesenchyma and its derivatives. It is worth mentioning that the Vietnamese kindred were reported as "autosomal dominant occipital cephalocele" (05). Imaging revealed occipital skull defects and variable retrocerebellar fluid accumulation, but no cerebellar vermis anomalies as seen in Dandy-Walker malformation proper. The same heterozygous ND1 variant (c.1162 C> T) reported by Dabro and coworkers was later confirmed in 4 members of a 3-generation family by McNiven and coworkers (42).

AP1S2 mutations were reported in "X-linked Dandy-Walker malformation with intellectual disability" in a family previously reported as Pettigrew syndrome (14). This is another X-linked mental retardation syndrome, rather than a Dandy-Walker malformation proper, as imaging revealed inferior vermis hypoplasia and megacisterna magna, respectively.

Epidemiology

The prevalence of isolated (ie, not associated with another defined syndrome) Dandy-Walker syndrome was estimated to be about 1 per every 30,000 live births (48). It accounts for about 2% to 4% of infantile hydrocephalus (32). In some, but not all, larger series, a mild preponderance toward females is seen (32; 48); this is particularly evident in patients with facial hemangioma (59; 55). A population-based survey of fetal posterior fossa anomalies from the northern region of England found an incidence of Dandy-Walker malformation of approximately 1 per 11,000 (39).

A comprehensive epidemiological study was reported using data from the European population-based registries of congenital anomalies in the European Surveillance of Congenital Anomalies network (EUROCAT) (58). Data were collected from 28 registries in 17 countries in the observation period from 2002 to 2015. Results are presented for Dandy-Walker malformation and Dandy-Walker variant, diagnosed using the International Classification of Diseases. Although the individual diagnostic allocation and the large regional variability may be questioned, the overall study results are of interest. A total of 734 cases were registered: 562 Dandy-Walker malformation and 172 Dandy-Walker variant cases. About 90% were diagnosed prenatally. The overall prevalence of Dandy-Walker malformation was 6.79 per 100,000 births, with 39.2% livebirths, 56.5% termination of pregnancy, and 4.3% late fetal deaths. The livebirth prevalence was 2.74 per 100,000 births. About 40% had an isolated Dandy-Walker malformation. A substantial proportion had multiple congenital anomalies, chromosomal aberrations, and genetic syndromes.

Prevention

Primary prevention of Dandy-Walker syndrome is not possible. According to Murray and colleagues, in isolated Dandy-Walker syndrome, the recurrence risk for parents of an affected proband is in the order of 1% to 5% (45). Bordarier and Aicardi reported that the risk is "practically negligible" (11). Prenatal diagnosis is possible (see Diagnostic workup).

Differential diagnosis

Differentiation of isolated Dandy-Walker syndrome from other conditions with Dandy-Walker syndrome as a possible association is mainly based on clinical evidence of additional features (eg, dysmorphic signs, cleft palate, growth retardation, joint contractures, skeletal abnormalities, ocular findings). Differentiation from other posterior fossa cystic lesions (ie, mega cisterna magna and retrocerebellar arachnoid cyst) should not be problematic in the postnatal situation if standard neuroimaging criteria are followed (35; 71; 03). The position of the choroid plexus in the fourth ventricle gives important clues to the nature of a posterior fossa cyst: normal in arachnoid cyst, absent in Dandy-Walker, displaced into the superior cyst wall in Blake pouch (46). In difficult situations CSF flow MRI can improve the diagnostic certainty (75). Some authors consider Blake pouch cyst consisting of posterior ballooning of the superior medullary velum into the cisterna magna as an entity within the Dandy-Walker continuum (15). The delineation of Ritscher-Schinzel syndrome from Dandy-Walker syndrome may not be clear-cut; although the older sister of the index patient of our initial report had marked shunt dependent hydrocephalus, the proband did not have ventriculomegaly (56; 36). Joubert syndrome is characterized by near total absence of cerebellar vermis, normal cerebellar hemispheres, a "bat-wing" shaped fourth ventricle, and a "molar-tooth" shaped midbrain; a large posterior fossa cyst is not a feature (68; 03). Exceptionally, a Dandy-Walker malformation may mask recognition of the molar tooth sign (60). A similar exceptional case was included in an article reporting a further gene (ARMC9) associated with Joubert syndrome (73). Unfortunately, patients with clear-cut evidence of a molar tooth sign are still published as having Dandy-Walker malformation rather than Joubert syndrome and related disorders (64). For review of vermian agenesis see Bordarier and Aicardi (11). PHACE(S) association (MIM 606519) is often associated with a facial hemangioma, but the posterior fossa anomalies are distinct from Dandy-Walker syndrome. In a cohort of 55 patients with PHACE syndrome, a classical-type Dandy-Walker malformation was described in a single individual, whereas unilateral cerebellar hypoplasia was the predominant finding (67). Attention has been drawn to acquired cerebellar lesions as a potential complication of extreme prematurity (43) and posthermorrhagic cerebellar disruption respectively (38). These changes may mimic Dandy-Walker syndrome and may erroneously be called Dandy-Walker variant (16). These dynamic changes have been impressively illustrated by Pichiecchio and colleagues in a patient with prenatal cerebellar hemorrhage and postnatal progressive upward displacement of a hypoplastic vermis and enlargement of the posterior fossa mimicking perfectly a “classical” Dandy Walker syndrome (52).

Diagnostic workup

The postnatal diagnosis of Dandy-Walker syndrome is based on detailed neuroimaging investigation, preferably MRI. Great care is required not to overlook a molar tooth sign in rare instances of Joubert syndrome mimicking a Dandy-Walker syndrome. Careful, clinical investigation is mandatory to assess if Dandy-Walker syndrome is isolated or an associated finding in the context of other defined syndromes. It is arguable whether chromosomal analysis is indicated in isolated Dandy-Walker syndrome.

A prenatal diagnosis is less straightforward and requires great caution. One has to be aware that an enlarged fourth ventricle may be found at 14 to 16 weeks’ gestation as a transient phenomenon (13); thus, prenatal diagnosis is usually possible toward 20 weeks' gestation (22; 34). For experienced investigators, the diagnosis is usually possible by ultrasound. In some instances, however, differentiation of Dandy-Walker syndrome, mega cisterna magna, and retrocerebellar arachnoid cyst can be problematic (24). This difficulty has been largely mastered with increasing experience and technical advances (17). Pinto and colleagues have raised an alert regarding an important potential pitfall in early second trimester fetal MRI (54). They illustrate 2 fetuses with upward anti-clockwise rotation of the vermis at 21 and 20 weeks’ gestation, respectively, followed by complete normalization on follow-up imaging. They conclude that follow-up imaging is crucial in order not to misinterpret this anatomic situation and to avoid the potential risk of unnecessary pregnancy termination. Authors of another study suggest that the “tail sign” could be helpful in the differential diagnosis between Dandy Walker syndrome, vermian malrotation, and vermian hypoplasia (07). The “tail sign” appears as a “dysmorphic” and elongated end of the posterior vermis on the sagittal plane.

Boemer and colleagues observed (antenatally) a Dandy-Walker malformation in siblings with carnitine palmitoyltransferase II (CPT2) deficiency, a potential diagnostic pitfall (09).

Chromosomal aberrations are more frequent in prenatally diagnosed Dandy-Walker syndrome (72; 22).

Management

The treatment of Dandy-Walker syndrome is surgical. Direct surgery on the posterior cranial fossa and membrane excision, which was attempted previously, has been abandoned (32). Aggressive cyst fenestration has been re-recommended (74). Currently, the treatment of choice if progressive hydrocephalus is evident is some form of shunting. Controversy exists concerning the optimal proximal shunt placement. Some specialists initially prefer a standard ventriculoperitoneal shunt (33), whereas others first attempt a cystoperitoneal derivation (32). Combined shunting of the cyst and lateral ventricles is recommended by Osenbach and Menezes (48). In this subgroup of pediatric hydrocephalus, the possible role of third ventriculostomy versus cerebrospinal fluid shunt is not yet clear. Upward herniation of the posterior compartment through the tentorial incisura is uncommon but should be carefully monitored by follow-up neuroimaging.

In light of the frequently observed delayed milestones and cognitive impairment, evaluation of psychomotor development is essential for provision of supportive measures, including physiotherapy or special education.

Special considerations

Pregnancy

Prenatal diagnosis is usually possible around 20 to 21 weeks' gestation. Potential pitfalls in prenatal diagnosis are discussed in the Diagnostic workup section.