Congenital hydrocephalus in clinical practice: A genetic diagnostic approach

abstract
Congenital hydrocephalus is a common and often disabling disorder. The etiology is very
heterogeneous. Little is known about the genetic causes of congenital hydrocephalus. A
retrospective survey was performed including patients with primary congenital
hydrocephalus referred to the Department of Clinical Genetics between 1985 and 2010 by
perinatologists, (child) neurologists or pediatricians. Patients with hydrocephalus secondary
to other pathology were excluded from this survey. We classified patients with primary
congenital hydrocephalus into two main groups: non-syndromic hydrocephalus (NSH) and
syndromic hydrocephalus (SH). Seventy-five individuals met the inclusion criteria,
comprising 36% (27/75) NSH and 64% (48/75) SH. In 11% (8/75) hydrocephalus was
familial. The cause of hydrocephalus was unknown in 81% (61/75), including all patients
with NSH. The maleefemale ratio in this subgroup was 2.6:1, indicating an X-linked factor
other than the L1CAM gene. In the group of SH patients, 29% (14/48) had a known cause of
hydrocephalus including chromosomal abnormalities, L1 syndrome, MardeneWalker
syndrome, WalkereWarburg syndrome and hemifacial microsomia. We performed this
survey in order to evaluate current knowledge on the genetic etiology of primary congenital
hydrocephalus and to identify new candidate genes or regulatory pathways for congenital
hydrocephalus. Recommendations were made concerning the evaluation and genetic workup
of patients with primary congenital hydrocephalus. We conclude that further molecular and
functional analysis is needed to identify new genetic forms of congenital hydrocephalus.

Introduction
Hydrocephalus is defined as an increase in the cerebral ventricular size and/or subarachnoid
space [1]. It is caused by an imbalance between the production, circulation and resorption of
cerebrospinal fluid (CSF). This excludes ventriculomegaly caused by primary cerebral
atrophy. Most forms of hydrocephalus are caused by obstruction of the flow of CSF. In non-
communicating hydrocephalus, the obstruction site lies within the ventricles (e.g. aqueduct
stenosis) or at the junction between the ventricular and subarachnoid space. Communicating
hydrocephalus results from an obstruction within the subarachnoid space (e.g. venous sinus
occlusion). Non-obstructive hydrocephalus is caused by overproduction of CSF (e.g. choroid
plexus papillomata) [2].

Fetal cerebral ventriculomegaly is evident in most of the patients with congenital

hydrocephalus [3]. Clinical features in newborns and children up to two years are
characterized by macrocephaly, frontal bossing, bulging anterior fontanel, prominent scalp
veins, sunset phenomenon and increased muscle tone. Parents may report poor feeding,
irritability and vomiting. In older children, head size may be normal if fontanels are closed.
Symptoms include vomiting, headaches, vision loss secondary to papilledema and optic nerve
atrophy, disturbances of consciousness, abnormal hypothalamic functions and difficulty in
walking secondary to spastic lower limbs.
Congenital hydrocephalus accounts for approximately 50% of all forms of hydrocephalus [4].
The etiology of congenital hydrocephalus is very heterogeneous. The majority consists of
secondary forms, caused by intra-uterine infections, intracranial hemorrhages, trauma,
teratogens and tumors (Table 1). Congenital hydrocephalus may also result from neural tube
defects and is found in association with other central nervous system malformations. Primary
congenital hydrocephalus occurs in 0.2e0.8 per 1000 live births [5]. A possible genetic
etiology is present in about 40% of patients with congenital hydrocephalus. This includes
cytogenetic abnormalities, monogenic or complex inherited conditions and multifactorial
disorders. However, in the majority of cases the cause of primary congenital hydrocephalus is
still unknown.
2. Materials and methods
The local Medical Ethical Committee of the Maastricht University Medical Centerþ approved
this retrospective survey. The selected individuals included patients referred for genetic
counseling and diagnostics on congenital hydrocephalus at the Department of Clinical
Genetics of the South-East part of the Netherlands between 1985 and 2010. These patients
were clinically re-evaluated for primary versus secondary and dysgenetic forms of
hydrocephalus using inclusion and exclusion criteria (Table 2). In six patients with congenital
hydrocephalus radiological examination and/or autopsy of the brain was not performed or too
much data were lost. They could therefore not be included in this survey. The majority of the
patients was excluded because hydrocephalus had a secondary cause, mainly neural tube
defects. Patients with dysgenetic hydrocephalus resulting from disturbances in early
embryonic development (e.g. holoprosencephaly, hydrancephaly and lissencephaly) were
also excluded. In addition, patients with known classical cytogenetic abnormalities (e.g.
trisomies 13, 18, 21 and triploidy) at referral were removed from this survey. Finally, 75
patients were included in this survey. In these patients pre- or postnatal conventional
karyotyping and, since its introduction at our department in 2009, genome-wide single
nucleotide polymorphism (SNP) and copy number variation (CNV) analysis was offered to
screen for small chromosomal imbalances. In males L1CAM gene mutation analysis was
performed. We clinically subdivided patients with congenital hydrocephalus into two main
categories: non-syndromic hydrocephalus and syndromic hydrocephalus. Non-syndromic
hydrocephalus (NSH) was defined as congenital hydrocephalus without other major
congenital abnormalities and with a maximum of two minor congenital anomalies.
Syndromic hydrocephalus (SH) was defined as hydrocephalus accompanied by at least one
major congenital abnormality (other than major brain malformations) or three minor
congenital anomalies. Classification of phenotypic findings as “minor” or “major” was
adapted from Merks et al. (2003) [6]

3. Results
Using the definitions for non-syndromic (NSH) and syndromic hydrocephalus (SH) in these
75 patients, 36% (27/75) had NSH and 64% (48/75) SH. Hydrocephalus was familial in 11%
(8/75) of the included patients. In 81% (61/75) the cause of hydrocephalus was unknown. The
maleefemale ratio in this subgroup was 2.6:1. In 75% (56/75) patients conventional or
molecular karyotyping was performed. L1CAM gene mutations analysis was carried out in
18 males. Pregnancy outcome was lethal in 27% (20/75) cases, consisting of 2 miscarriages,
10 induced abortions and 8 stillbirths.

3.1. Phenotype in non-syndromic hydrocephalus (NSH)

In all 27 patients with NSH the cause was unknown. One patient had affected male family
members, suggestive of an X-linked form of NSH. L1CAM gene mutation analysis did not
reveal any mutation. In general, outcome was not favorable: death occurred in 30% (8/27)
before the age of one year. Ages in the survivors ranged from 6 to 37 years. Remarkably, in
14/27 patients with NSH radiological examination of the brain showed additional
abnormalities including aqueduct stenosis (8/27), corpus callosum agenesis (2/27), septum
pellucidum aplasia (1/27), secondary cerebral atrophy (2/27) and cerebellar hypoplasia
(1/27). A small number of patients displayed additional neurological features such as general
or axial hypotonia (3/27), spastic diplegia (1/27) or seizures (2/27). Developmental delay
and/or mental retardation was present in 2 of 4 individuals with complete clinical follow-up
until the age of 18 years. In some patients minor features were present involving the eye,
skin, hair and kidney (Table 3).

3.2. Phenotype in syndromic hydrocephalus (SH)

The SH group consisted of 48 patients, of which 29% (14/48) had a known cause of
hydrocephalus (Table 4). Five SH patients had a chromosomal abnormality: two de novo
trisomy 9 (Fig. 1A), one familial partial trisomy 9p21.3, one de novo unbalanced
translocation 46,XY,der(6)t(6;13)(p25;q34) (Fig. 1B) and one familial unbalanced
translocation 46,XY,der(6)t(6;18)(q26;p11.2). Six patients had mutations in the L1CAM gene
(c.76þ5G>A; c.78T>A; c.778_785del8; c.938T>C; c.1124-1G>C; c.1267C>T) (Fig. 1C and
see: http://www.l1cammutationdatabase.info). One family had MardeneWalker syndrome
(Fig. 1D), one patient had WalkereWarburg syndrome (Fig. 1E) with proven defective O-
glycosylation of alpha-dystroglycan and one patient had hemifacial microsomia (Fig. 1F).
Two other patients with SH had affected family members. In one family, a male and his
maternal uncle presented with hydrocephalus indicating a possible X-linked form of SH.
L1CAM gene mutation analysis was normal. In another family an X-linked dominant skeletal
phenotype was present in carrier women with a lethal form of hydrocephalus in affected
males (Fig. 1G). Carrier women showed 100% non-random X-inactivation in their blood
lymphocytes. X-exome sequencing, segregation analysis and functional analysis is still
pending to identify the underlying gene defect for this new syndromic form of congenital
hydrocephalus. General prognosis in the 34 patients with SH of unknown cause was poor.
Death occurred in 32% (11/34) before the age of one year. Ages in the survivors ranged from
2 to 53 years. In 50% (17/34) patients radiological examination of the brain showed
additional abnormalities including aqueduct stenosis (6/34), corpus callosum agenesis (8/34),
septum pellucidum aplasia (3/34), corticospinal tract abnormalities (1/34), secondary cerebral
atrophy (3/34) and cerebellar hypoplasia (1/34). About one third of the patients displayed
additional neurological features such as general or axial hypotonia (3/34), spastic hemiplegia
(2/34) or seizures (5/34). Developmental delay and/or mental retardation was present in seven
of nine individuals with complete clinical follow-up until the age of 18 years. A large number
of patients (18/34) had ocular abnormalities, including eye movements disorders and visual
disturbances. Skeletal and digital abnormalities were present in 41% (14/34) patients. Nine
patients had genitourinary abnormalities, such as renal agenesis (3/34), hydronephrosis (2/34)
and micropenis (2/34). Congenital heart defects were present in five patients, including septal
defects and supravalvular pulmonary stenosis. Congenital diaphragmatic hernia was present
in three patients. Three other patients had inguinal hernias (Table 3). Furthermore, two
patients had anemia. In one patient a T-cell immunodeficiency was confirmed and presented
with severe complications.

4. Discussion
4.1. Definition and classification of hydrocephalus
Providing an accurate definition and classification on hydrocephalus is challenging. It has
been part of an ongoing debate since many years. The discussion mainly focuses on the in- or
exclusion of cerebral atrophy and the distinction between primary and dysgenetic forms of
congenital hydrocephalus in the presence of other major central nervous system
malformations [7]. Completely excluding cerebral atrophy from the definition is controversial
as hydrocephalus and cerebral atrophy sometimes coexist independently within the same
clinical entity. The same problem is encountered if other central nervous system
malformations are present next to congenital hydrocephalus. A clear distinction between
primary endysgenetic forms of congenital hydrocephalus is not always possible in these
cases. For example, cilia disorders that affect cerebellar development can lead to
hydrocephalus at the same time. Another example is the nosologic discussion on
MardeneWalker syndrome, which will probably continue until the underlying gene defect is
known. The DandyeWalker malformation with hydrocephalus in this syndrome can lead to
cerebral and cerebellar atrophy, but the reverse can also be true. Thus, these brain
malformations can be either primary, secondary or associated with each other. Further
discussion is needed to reach consensus at these critical points. The lack of a final definition
and classification on hydrocephalus is an important limitation to this survey. It makes our in-
and exclusion criteria disputable and carries the potential risk of selection bias.

4.2. General diagnostic approach to congenital hydrocephalus

Genetic evaluation of congenital hydrocephalus starts with a detailed medical history,
physical examination and brain imaging with close involvement of child neurologists,
pediatricians and clinical geneticists (Fig. 2). The first step is to make a clinical distinction
between primary, secondary and dysgenetic forms of congenital hydrocephalus (Table 2) and
a proper differential diagnosis (Supplemental Tables 1e3). This survey focused exclusively
on the primary forms of (N)SH. Depending on the presence of additional clinical features,
one should decide whether the hydrocephalus is non-syndromic (no major and 2 minor
anomalies) or syndromic (1 major and >2 minor anomalies). The genetic diagnostic approach
differs substantially between both groups.

4.3. Genetic evaluation of non-syndromic hydrocephalus

So far, L1 syndrome is the most common genetic cause of congenital hydrocephalus. It
accounts for about 5e10% of males with congenital hydrocephalus [4]. The L1 syndrome is
caused by mutations in the neural cell adhesion molecule L1 (L1CAM) gene at Xq28. The
phenotypic spectrum is variable and comprises both NSH and SH: X-linked hydrocephalus
with stenosis of the aqueduct of Sylvius (HSAS), MASA syndrome (Mental retardation,
Aphasia, Spastic paraplegia and Adducted thumbs), X-linked complicated hereditary spastic
paraplegia type 1 and X-linked complicated corpus callosum agenesis [8]. Mutation analysis
of the L1CAM gene should be performed in all males with NSH. As hydrocephalus is
described in w5% of female carriers, mutation analysis of the L1CAM gene should also be
considered in (mildly) affected females, especially if there is a positive family history of
congenital hydrocephalus or in the presence of aqueduct stenosis of Sylvius, corpus callosum
agenesis or adducted thumbs.
In patients with NSH array-based comparative genomic hybridization (CGH) or single
nucleotide polymorphism (SNP) array CNV analysis can be considered to screen for small
chromosomal imbalances. However, the prevalence of pathogenic copy number variants in
NSH patients is thus far unknown and needs further investigation. 4.4. Genetic evaluation of
syndromic hydrocephalus The most frequent cytogenetic abnormalities in SH include
(mosaic) trisomy 9, 9p, 13 and 18 and (mosaic) triploidy [9]. Since the introduction of
genome-wide CNV analysis using microarray platforms several other small chromosomal
aberrations have been reported in SH patients (Supplementary Table 1). In particular, partial
6p terminal deletions are reported rather frequently [10]. The 6p terminal deletion has a
recognizable phenotype. Clinical features include Axenfeld-Rieger anomaly, hearing loss,
congenital heart disease, dental anomalies, developmental delay and a characteristic facial
appearance (Fig. 1B). Maclean et al. (2005) reported a child with a 6pter microdeletion
arising from a de novo translocation (6;18)(p25.1;p11.2) including the FOXC1, FOXF2 and
FOXQ1 forkhead gene cluster at 6p25. Central nervous system anomalies included
hydrocephalus and hypoplasia of the cerebellum, brainstem and corpus callosum with mild to
moderate developmental delay [11]. The most interesting candidate gene for this form of
congenital hydrocephalus is FOXC1: mice with homozygous nonsense mutations of the
Foxc1 gene die at birth because of progressive hydrocephalus [12]. Mutations in the FOXC1
gene have been identified in patients with Axenfeld-Rieger syndrome and more recently, also
in DandyeWalker syndrome [13]. The latter is defined by partial or complete absence of the
cerebellar vermis and cystic dilatation of the fourth ventricle, often accompanied by
congenital hydrocephalus.
In SH patients without a directly recognized syndrome, array platforms combining SNP and
CNV analysis allow for the detection of (partial) uniparental disomy or regions of
homozygosity (ROH). We prefer to use both techniques, especially in patients with presumed
autosomal recessive congenital hydrocephalus (e.g. consanguineous parents or affected sib
pairs). This sometimes leads to identification of a disease-causing gene. Multiplex Ligation-
dependent Probe Amplification (MLPA) analysis remains a valuable, inexpensive tool to
directly confirm clinically suspected (subtelomeric) microdeletions or -duplications.
Conventional karyotyping can still be useful to investigate mosaic (<30%) chromosome
abnormalities and possible (de novo) inversions or balanced translocations in which the
breakpoints can potentially disrupt genes or have a position effect on neighboring genes.
The L1 syndrome should be excluded in every male with SH. Vos et al. (2010) screened 60
patients suspected of having L1 syndrome without L1CAM gene mutation using a high
density X-chromosome specific array-based CGH. Eleven CNVs were identified containing
candidate genes expressed in brain (including e.g. the GDI1 gene) (thesis Y. Vos: Genetics of
L1 syndrome, Chapter 8, 2010). Further research is needed to determine the potential role of
these genes in X-linked congenital SH. Few other families with X-linked SH have been
described. The high maleefemale ratio observed in our cohort (2.6:1) indicates possible
involvement of X linked disease-causing genes other than L1CAM. Candidate genes for
lethal forms of congenital hydrocephalus in males include OFD1, AP1S2 and HDAC6
involved in respectively orofaciodigital syndrome type I [14], X-linked mental retardation
[15] and X linked dominant chondrodysplasia [16]. SH is otherwise mainly inherited in an
autosomal recessive manner. Congenital hydrocephalus has occasionally been described in
numerous other genetic disorders [17]. A nonexhaustive list is available in Supplementary
Table 3. In 21% (7/34) of our patients with unresolved SH clinical features of the
VACTERL-H association were identified. The phenotypic spectrum of the VACTERL-H
association is wide and poorly understood [18]. Hemifacial microsomia and Fanconi anemia
show many overlapping features including vertebral, gastrointestinal, cardiac and limb
defects. Convincing relationships between these conditions and VACTERL-H have been
described in literature [19e21]. As suggested earlier by Faivre [22], chromosome breakage
studies should be considered in patients with congenital hydrocephalus and additional
features of VACTERL-H association.
Metabolic disorders should not be disregarded as a possible cause of hydrocephalus. Given
the mainly autosomal recessive inheritance, it should especially be considered in affected sib
pairs. Metabolic disorders are more likely to cause postnatal hydrocephalus. Congenital
hydrocephalus is a common feature in disorders with glycosylation defects such as muscular
dystrophydystroglycanopathies, including WalkereWarburg syndrome (Supplementary Table
2). Other metabolic disorders associated with congenital hydrocephalus include for example
mucopolysaccharidoses, alpha-mannosidosis and Smith-Lemli-Opitz syndrome
(Supplementary Table 3). Metabolic investigation in SH should at least include assessment of
blood creatine kinase (CK) and lactate, plasma cholesterol metabolites, urinary
mucopolysaccharides and isoelectric focusing for sialotransferrine

4.5. Future perspectives

Upcoming next generation sequencing technologies are rapidly changing the field of genetic
diagnostics. At present, whole exome and/or genome sequencing are still too expensive for
widespread use. Analysis of the large amount of data generated, e.g. interpreting the
numerous unclassified genetic variants, will become a major challenge. Targeted X-exome
sequencing can be a more readily available option in males with congenital hydrocephalus,
especially in case of a positive family history. Growing evidence from literature indicates that
cilia disorders may play an important role in congenital hydrocephalus. Cilia are
microtubule-based organelles that extend from the surface of almost every cell in the human
body (e.g. in the ependyma of the cerebral ventricles) and are able to propel fluid [23, 24]. As
ciliated ependymal cells play a crucial role in the control of cerebrospinal fluid homeostasis
in the brain, cilia defects can lead to the formation of hydrocephalus [25]. Two cilia-related
disease mechanisms have been identified in hydrocephalus formation:
A. defects in the transport of proteins across the ciliary compartment border leading to the
formation of aberrant cilia with overproduction of CSF, and B. ependymal ciliary dysmotility
leading to an impaired circulation and obstruction of CSF. The phenotypic spectrum of cilia
disorders includes cerebellar hypoplasia, retinopathy, polycystic kidneys, liver cysts,
polydactyly, obesity, leftright asymmetry, situs inversus, infertility and (sometimes)
hydrocephalus. To date, more than 100 human cilia genes are detected. The genetic
heterogeneity of cilia disorders complicates current diagnostics. Implementation of genome-
wide exome
sequencing in the near future will enable us to generate data on the various known cilia genes
in a single experiment. The Ciliary Proteome Database can then be a helpful instrument for
interpreting this data [26].

4.6. Prognosis
The prognosis of a child with congenital hydrocephalus is highly variable and mainly
depends on the underlying cause of hydrocephalus, associated malformations and the timing
and success of surgical treatment [27]. General mortality rate varies from 5% to 15%. About
40% of the children with congenital hydrocephalus is cognitively impaired, ranging from
mild to severe retardation. The degree of cognitive impairment does not necessarily correlate
with head circumference or severity of hydrocephalus [8]. Motor problems are seen in 30%
of the children with congenital hydrocephalus, of which half is wheelchair dependent [28].
Delay of treatment is an important risk factor for poor outcome [2].

4.7. Recurrence risk

The recurrence risk in sporadic patients with NSH is low: approximately 4% in males and 2%
in females [4]. However, when NSH is familial the recurrence risk increases dramatically,
depending on the mode of inheritance. Familial forms of NSH are often X-linked with a 50%
recurrence risk for the male offspring of female carriers. Daughters have a 50% chance of
being a carrier with a low risk (<5%) of manifesting clinical features of congenital
hydrocephalus [4]. In sporadic patients with SH the recurrence risk depends on the
inheritance pattern of the suspected syndrome. In SH patients with hydrocephalus of
unknown origin the estimated recurrence risk, after extensive clinical and molecular
investigations, is 5%. In familial forms of SH the recurrence risk depends on the inheritance
pattern suspected from pedigree data (Fig. 2).

5. Conclusion
We have to conclude that genetic causes of congenital hydrocephalus are still largely
unknown. This retrospective survey provides evidence for new genetically determined forms
of congenital hydrocephalus. Candidate gene and functional analysis is ongoing. We
recommend L1CAM gene mutation analysis and genome-wide SNP/CNV analysis in every
patient with congenital hydrocephalus. In addition, metabolic investigation, conventional
karyotyping and chromosomal breakage studies should be considered in patients with
unknown forms of SH. Further research is needed in order to identify the underlying gene
defects in hitherto unknown, possibly genetic forms of (non-) syndromic congenital
hydrocephalus.