Clin Perinatol 30 (2003) 531 – 539

Fetal hydrocephalus
George H. Davis, DO
Division of Maternal-Fetal Medicine, Section of Fetal Diagnosis and Therapy,
Vanderbilt University Medical Center, VUMC B-1100, MCN Nashville, TN 37232-2519, USA

Hydrocephalus is a dynamic process resulting in a progressive increase in

ventricular volume caused by a relative or complete obstruction between the sites
of production and absorption or overproduction of cerebrospinal fluid (CSF). In
common usage, the term hydrocephalus encompasses a variety of malformations
with widely different pathologic and sonographic findings, pathogeneses, and
prognoses. In general, the mortality of fetuses that have hydrocephalus varies
directly with the presence and severity of extra- central nervous system (CNS)
anomalies, whereas the neurological outcome is determined by the underlying
CNS malformation.
CSF is formed within the lateral ventricles at the choroid plexus. Under
normal circumstances it circulates slowly to the third and then the fourth ven-
tricles. It continues its flow through the foramen magnum and into the spinal
canal. At the level of the fourth ventricle it also communicates, by way of the
foramina of Luschka and Magendie, into the subarachnoid space, bathing the
cerebral hemispheres.

Etiology and associated anomalies

The precise prevalence of congenital hydrocephalus is difficult to determine
because of inaccurate diagnosis and underdetection, but its rate has been re-
ported as 0.3 to 1.5 per 1000 live births [1]. Congenital hydrocephalus has been
found in association with many genetic syndromes, and isolated hydrocephaly is
rare. Associated intracranial anomalies are found in 37% of cases [2] and
extracranial anomalies are found in 63% of cases, with spinal dysraphism being
the most common.
Oi and colleagues classified 61 cases of fetal hydrocephalus as follows:

Primary: communicating hydrocephalus, aqueductal stenosis, foramen atresia,

and other forms of obstructive hydrocephalus

E-mail address: george.davis@vanderbilt.edu

0095-5108/03/$ – see front matter D 2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0095-5108(03)00053-8
532 G.H. Davis / Clin Perinatol 30 (2003) 531–539

Dysgenetic: associated with spina bifida, bifid cranium, Dandy-Walker mal-

formation, holoprosencephaly, hydranencephaly, lissencephaly, congenital
cyst, and others
Secondary: caused by brain tumor, hemorrhagic or other vascular diseases, in-
fection, trauma, subdural fluid collection, and others [3]

Nineteen cases (31%) were primary, 34 (56%) were dysgenetic, and eight
(13%) were secondary. The mean IQ was 74.2 (range 20 –132) in patients who
had primary hydrocephalus, 52.4 (range 20 –120) in patients who had dysgenetic
hydrocephalus, and 26 (range 5– 70) in patients who had secondary hydrocepha-
lus [4].
Aqueductal stenosis resulting in hydrocephalus can be X-linked, infectious,
teratogenic, or neoplastic in origin [5]. It is characterized by dilation of the lateral
ventricle and the third ventricle without any other abnormalities. Because the
mean IQ of such patients is about 70 with this diagnosis [6], an accurate diagnosis
is desirable for counseling. Communicating hydrocephalus, characterized by
dilation of the subarachnoid cistern, should also be differentiated from other
forms of hydrocephalus. In this case, the prognosis is far better [7] and prenatal
intervention is not indicated.
In a report on 50 fetuses that had hydrocephalus, 72% died in the immediate
neonatal period and 84% of these fetuses had associated anomalies [7]. In a
similar report of 61 fetuses that had hydrocephalus, 84% had one or more major
CNS anomalies (63%) or extra-CNS anomalies (49%). The overall uncorrected
mortality rate was 67% with deaths occurring in all fetuses that had multiple
extra-CNS anomalies, 57% of fetuses that had an isolated extra-CNS anomaly,
and 37% of fetuses that had no extra-CNS anomaly [3].

Prenatal diagnosis
Obstetrical ultrasonography now permits prenatal diagnosis of most cases of
hydrocephalus. Ultrasonographic criteria for the diagnosis of congenital obstruc-
tive hydrocephalus include isolated, marked ventricular enlargement with macro-
cephaly or progressive enlargement on serial examinations. Misclassification of
fetuses as having isolated versus nonisolated mild ventriculomegaly occurred in
less than 2% of cases [8]. In the presence of marked ventriculomegaly, however, the
posterior fossa and cerebellum are usually examined only with great difficulty.
MRI has a different biophysical basis than ultrasound, with superior soft tissue
contrast. MRI also allows sagittal and coronal scans of fetal organs, views not
commonly obtainable with advancing gestation because of fetal position.
In a study of 36 complicated pregnancies at 29 to 41 weeks’ gestation, six
fetuses had ventriculomegaly diagnosed with high-resolution real-time ultraso-
nography. The authors found that the axial view of the fetal head was satisfactory
for identifying cerebral ventricular enlargement, thickness of the cerebral mantle,
and most of the midline brain structures such as the third and fourth ventricles,
 G.H. Davis / Clin Perinatol 30 (2003) 531–539 533

the falx, and the thalamus. Ossification of the calvarium, however, impaired
visualization of the posterior fossa, and sagittal imaging was impossible even
when a vaginal approach was used. In those cases, MRI provided better
visualization of the cerebellum and corpus callosum. MRI alone could image
the pons, the base of the skull, and the cervical spine. As a result, in every case
MRI provided diagnostic information that was not obtained by ultrasonography.
Additional findings included hemorrhage, cerebellar size, corpus callosum, and
outlines of skull bases, the brain stem, and the cervical spine [9].

Management
Depending on the gestational age of the fetus, management options include
pregnancy termination, expectant management with delivery at or near term, or
preterm delivery. In discussing these options it is important to consider associated
anomalies and findings while keeping in mind the high rate of mortality when
multiple anomalies are present [3]. In spinal dysraphism, among others, amnio-
centesis might be considered to exclude such aneuploid conditions as trisomy
18 and 21. Chromosomal abnormalities occur in 11% of fetuses that have
hydrocephalus [2].
When hydrocephaly is secondary to spina bifida and genetic evaluation is
normal, careful attention should be paid to the lesion level, the degree of
ventriculomegaly, and the presence of kyphosis or scoliosis. See Ch. 7.
The route of delivery depends on the presumptive prognosis for the fetus and
the obstetrical problems posed by the enlarged fetal head. With a large fetal head
and poor prognosis, one available option is fetal cerebral decompression by
cephalocentesis and vaginal delivery. Such a procedure usually results in intra-
partum or neonatal demise. Historically, cephalocentesis has been employed
extensively in the management of the severely enlarged head. With the advent of
ultrasound and safer Cesarean delivery, this option has fallen into some disfavor,
but it might still be considered when the prognosis is lethal. The author
recommends working through the institutional ethics committee when this option
is desired.
Cesarean delivery can also be used to deliver the fetus that has an enlarged
cranium. In families for whom pregnancy termination or cephalocentesis are not
options and who desire Cesarean delivery with aggressive neonatal care including
surgery, progressive hydrocephalus presents a dilemma. In this case, delivery
must often be delayed until the fetus is mature enough to have a reasonable
chance of survival in the nursery. If hydrocephalus progresses during that time,
ongoing brain damage might occur while the fetus matures. Fetuses that have
isolated hydrocephalus experience a worsening mental outcome with increasing
length of gestation after diagnosis. In this situation, another alternative is
shunting fetal hydrocephalus before birth.
If drainage of CSF in cases of hydrocephalus is not provided, progressive
ventricular dilation and brain damage might result. The adverse effect of fetal
534 G.H. Davis / Clin Perinatol 30 (2003) 531–539

hydrocephalus on the developing brain is not well understood. Histologic studies

have implicated acute and severe hydrocephalus in causing brain damage. One
hypothesis proposed that increased CSF pressure leads to flattening of the
ependyma and interruption of the CSF –brain barrier followed by penetration
of the periventricular white matter by CSF. Demyelination secondary to axonal
degeneration results in irreversible tissue damage [10,11].
Although the exact point of irreversible tissue damage is unknown, the time of
onset and severity of hydrocephalus are important prognostic factors. Oi et al
evaluated the long-term follow-up results of 20 fetuses that had hydrocephalus
diagnosed in the third trimester. The only significant factor affecting outcome was
the length of gestation after diagnosis of hydrocephalus. The authors suggest that
the adverse effects of hydrocephalus might become irreversible before delivery
[11]. Further evidence of this concept is provided by an analysis of 19 fetuses that
had uncomplicated hydrocephalus, nine diagnosed between 22 and 32 weeks’
gestation and 10 identified between 32 and 40 weeks’ gestation. The mean IQ of the
former group was 67.5 (range 35– 100), whereas the IQ of the latter group was 76.6
(range 20 –132). Fetuses that had a longer duration of hydrocephalus before
delivery experienced a significantly worse outcome (P < 0.05) [12].
The concept of shunting hydrocephalus is based on the assumption that the
ventriculomegaly is caused by intraventricular hypertension. This elevated
pressure is presumed to be caused by obstruction of the normal circulation of
CSF through the ventricular system, the cerebral aqueduct, and the subarachnoid
space (ie, obstructive hydrocephalus).

In utero therapy
In the fetus that has obstructive hydrocephalus, shunting the excess CSF from
the lateral ventricle stops progressive ventricular enlargement and prevents
further brain damage. In a series of newborns that had overt hydrocephalus, less
than 15% of untreated infants survived, all of whom had neurological impair-
ment. By contrast, more than 80% of infants treated with early shunting survived,
and two thirds of these infants were normal at follow-up [13,14]. Based on the
information outlined previously, intrauterine ventricular shunting before brain
damage might potentially improve the outcome of affected children.
Initial attempts to place a shunt in utero consisted of ultrasonographically
guided, percutaneous placement of a needle into the uterus through the parietal
bone of the fetus and into the lateral ventricle. The site of entry into the fetal brain
was intended to be located posterior to the major motor cortex and anterior to the
visual area, comparable to the site used for shunt placement in the newborn. The
shunt (Denver Shunt) was pushed through this needle until it was in the proper
position in the fetal head, and the needle was withdrawn such that the distal end
of the shunt was located in the amniotic fluid space.
Two major technical problems with this procedure remain unresolved. The
first is that the Denver Shunt was easily removed from the fetal head. It was
 G.H. Davis / Clin Perinatol 30 (2003) 531–539 535

anchored in the head by small silicone rubber wings that were thin and soft
enough to fold flat along the device as it passed through the needle. The wings
provided only a weak anchoring strength. Several centimeters of shunt tube
protruded from the fetal scalp into the amniotic fluid. Shunt dislodgement from
the fetal head, whether caused by active pulling by the fetus or incidental to
normal fetal movement, accounted for most cases of intrauterine shunt failure.
The second cause of shunt failure was obstruction. Cases were reported in which
the shunt became obstructed with tissue from within the brain and by vernix from
the amniotic fluid.
Perhaps of greater importance than the technical problems were problems of
case selection. Twenty-two percent of 41 treated fetuses were found to have
associated anomalies after birth. Thus, poor patient selection contributed subs-
tantially to the poor outcome data derived from fetal hydrocephalus shunting
[15]. The best results were obtained in 32 fetuses that had aqueductal stenosis.
These fetuses had a survival rate of 87%; however, only 42% of survivors had
normal intellectual outcome, 7.2% had mild to moderate handicaps, and 50% had
severe handicaps. The procedure-related mortality was 9.75% [16].
In one small series conducted contemporaneously with these percutaneous
shunts, the control group had a better outcome than the shunted group [17]. As a
result, a moratorium on percutaneous shunting of fetal hydrocephalus was called
for and agreed to at the third annual meeting of the International Fetal Medicine
and Surgery Society in 1985.
It seems likely that with careful patient selection a small group of fetuses that
have severe, progressive hydrocephalus at an early gestational age would benefit
from intrauterine shunting [16]. Needle shunting, as described previously, is not
now possible because of the unsolved technical problems associated with the
device. With the moratorium in effect, such shunts are not commercially
available. There have, however, been recent calls from varied quarters to revisit
this issue now that there is a better understanding of the etiology and natural
history of the disease and better prenatal diagnosis and patient selection.
Treatment of hydrocephalus in the newborn is performed by placing a tube into
the lateral ventricle and connecting it to a valve that then sends the fluid through a
tube, which migrates subcutaneously downward into the peritoneal cavity. If a
similar apparatus and shunt could be placed at hysterotomy, the efficacy might be
vastly improved. As with in utero treatment of spina bifida (Ch. 7), careful
consideration would need to be given to the risks and benefits of the procedure.
With standard neurosurgical placement of the proximal shunt apparatus in the
affected fetus through a hysterotomy, the distal end of the drain exits between the
fetal scapulae and functions as a ventriculoamniotic shunt. The major risk to
the fetus is preterm delivery with resultant possible severe or permanent damage.
At Vanderbilt University Medical Center, patients carrying a fetus that has
isolated aqueductal stenosis (obstructive hydrocephalus) with progressive ven-
triculomegaly (ventricular enlargement > 1.5 mm on serial examinations) [18]
can be offered open fetal surgery to place a ventriculoamniotic shunt. Potential
candidates are counseled by the examining obstetrical ultrasonologist. In addi-
536 G.H. Davis / Clin Perinatol 30 (2003) 531–539

tion, the patient is counseled by a consulting genetic counselor. Diagnostic

transabdominal amniocentesis is performed and amniotic fluid is sampled for
fetal karyotype, cytomegalovirus (CMV) by polymerase chain reaction (PCR)
and culture, and toxoplasmosis by PCR and culture. Maternal blood is drawn for
serologic testing for syphilis and IgG- and IgM-specific antibodies to CMV and
toxoplasmosis. If the patient is rubella nonimmune, rubella-specific IgG and IgM
antibodies are also measured. A fetal MRI scan is obtained. The patient is
counseled by a consulting pediatric neurosurgeon and a social worker. Anesthesia
and neonatal consultation is obtained. A medical ethicist and a chaplain counsel
the patient. Consultation with the patient’s family doctor and a psychologist/
psychiatrist are available as indicated. In addition to pregnancy termination and
expectant management, the experimental procedure described below is offered as
a third option. The Fetal Surgery Subcommittee of the Hospital Ethics Committee
reviews each case, and informed consent is obtained.
At hysterotomy, intrauterine pressure is measured. A ventricular shunt is
placed in the standard neurosurgical fashion. A 1-cm incision is made in the scalp
in the right parietal area above and behind the right ear. A small (3 –4 mm) hole is
made in the skull using the scalpel blade. This incision is easily done because the
fetal skull is not completely ossified. The underlying dural membrane is
coagulated and incised to expose the brain. A standard blunt ventricular needle
is passed into the lateral ventricle using conventional external landmarks for
guidance. Intraventricular pressure is measured in a fashion identical to that
described for intrauterine manometry. The pressure scale is zeroed at the fetal
skin surface. When a free flow of CSF is obtained, the needle is removed and the

Fig. 1. Ventriculostomy with insertion of shunt.

 G.H. Davis / Clin Perinatol 30 (2003) 531–539 537

4-cm ventricular catheter of a Ultrasmall (PS Medical) shunting device with a

low –low pressure valve is passed through the needle track into the ventricle. The
ultrasound probe is used to confirm proper placement of this catheter (Fig. 1).
Next, a subcutaneous tunnel is made from the scalp incision to a point midway
between the scapulae (shoulder blades). A 2-mm incision is made at this point
and the proximal end of the distal drainage catheter that comes with the shunting
device is pulled up from the interscapular incision to the scalp incision. The
device is attached to the shunt valve, allowing fluid to flow from the ventricle
through the tubing and out into the amniotic fluid (Fig. 2).
The interscapular position has been chosen for three reasons: (1) this
placement is possible without repositioning of the fetus, which reduces procedure
time considerably; (2) it prevents the fetus from grasping the catheter and
dislodging it; and (3) the anterior portion of the chest is left untouched so the
shunt can be converted readily to a standard ventriculoperitoneal position at birth.
The valve prevents reflux of amniotic fluid into the ventricle and prevents
overdrainage of CSF. The scalp incision is then closed using a running PDS
(absorbable) suture. Because of its small size, the interscapular incision does not
require closure. Finally, the drainage catheter is trimmed so that only about 2 to
3 cm protrude through the skin. This portion of the procedure takes less than
10 minutes. The hysterotomy and laparotomy incisions are closed as described
previously. After discharge, weekly outpatient visits are performed for evaluation
of the mother and the fetus. Cesarean delivery is performed with the timing to be
determined by the usual obstetrical factors.

Fig. 2. Shunt completion: subcutaneous placement with subscapular drainage to amniotic cavity.
538 G.H. Davis / Clin Perinatol 30 (2003) 531–539

At present, this procedure is being offered only as part of an Institutional

Review Board-approved feasibility study. If it is successful, a more in-depth
evaluation or controlled trial might be considered.

Summary
Fetal hydrocephalus is a dynamic process. Its natural progression is not
completely understood, but it is almost always associated with other intracranial
and extracranial anomalies. Fetal hydrocephalus might begin as mild ventriculo-
megaly and progress to a more serious state. Mild ventricular enlargement without
an increase in head size might be a normal variant. Careful prenatal evaluation is
always indicated when the diagnosis of ventricular enlargement has been made,
and it should include a detailed anatomic survey and genetic amniocentesis.
Therapeutic options need to be discussed in detail with the patient. If multiple
associated anomalies—especially other CNS anomalies—are present, the prog-
nosis for a viable outcome is dismal. If isolated hydrocephalus is seen, gender
determination is appropriate in the consideration of X-linked hydrocephalus
secondary to aqueductal stenosis. Communicating hydrocephalus can also be mis-
taken for obstructive ventriculomegaly, but it can usually be excluded by dem-
onstration of dilation of the subarachnoid cistern.
There is a small subset of fetuses that have isolated hydrocephalus that might
benefit from prenatal surgical intervention. Feasibility studies are in progress to
assess the therapeutic benefit of this type of therapy. Gestational age requirements
and strict criteria are being evaluated. As with the multiple therapies discussed in
this issue, many interventions are becoming more and more feasible, and ante-
natal surgery is likely to be limited only by the ability to think imaginatively and
to act creatively.

References
[1] Habib Z. Genetics and genetic counseling in neonatal hydrocephalus. Obset Gynecol Surv 1981;
36:529.
[2] Charlene FA, Berkowitz RI, Romero R, et al. The diagnosis of fetal hydrocephalus. Am J Obstet
Gynecol 1983;147:703.
[3] Oi S, Honda Y, Hidaka M, Sato O, Matsumoto S. Intrauterine high-resolution magnetic reso-
nance imaging in fetal hydrocephalus and prenatal estimation of postnatal outcomes with ‘‘per-
spective classification. J Neurosurg 1998;88:685 – 94.
[4] Charlene FA, Duncan C, Mint LR, Hobbits JC, McClure M, Scott D, et al. Outcome of fetal
ventriculomegaly. Lancet 1984;2:179 – 81.
[5] McCullough DC, Balzer-Martin LA. Current prognosis in overt neonatal hydrocephalus. J Neuro-
surg 1982;57:378.
[6] Goldstein RB, La Pidus AS, Filly RA, Cardoza J. Mild lateral cerebral ventricular dilatation in
utero: clinical significance and prognosis. Radiology 1990;176:237 – 42.
[7] Salam MZ. Stenosis of the aqueduct of Sylvius. In: Vinken PJ, Bruyn GW, editors. Handbook of
clinical neurology, Volume 30. Amsterdam (Netherlands): Elsevier/North Holland Biomedical
Press; 1977. p. 609 – 22.
 G.H. Davis / Clin Perinatol 30 (2003) 531–539 539

[8] Kirkinen P, Partanen K, Vainio P, Ryynänen M. Ann Med 1996;28:131 – 6.

[9] Milhorat TH, Clark RG, Hammock MK, McGrath PP. Structural, ultrastructural, and permeability
changes in the ependyma and surrounding brain favoring equilibration in progressive hydro-
cephalus. Arch Neurol 1970;22:397 – 407.
[10] Weller R, Shulman K. Infantile hydrocephalus. J Neurosurg 1972;36:255 – 65.
[11] Oi SZ, Yamada H, Kimura M, Ehara K, Matsumoto S, Katayama K, et al. Factors affecting
prognosis of intrauterine hydrocephalus diagnosed in the third trimester. Neurol Med Chir 1990;
30:456 – 61.
[12] Laurence KM, Coates S. The natural history of hydrocephalus: detailed analysis of 182 unop-
erated cases. Arch Dis Child 1962;37:345 – 62.
[13] Young HF, Nulsen FE, Weiss MH, Thomas P. The relationship of intelligence and cerebral
mantle in treated infantile hydrocephalus. Pediatrics 1973;52:38 – 44.
[14] Clewell WH. Hydrocephalus shunt. In: Charlene FA, Isaacson GC, Campbell S, editors. Ultra-
sound in obstetrics and gynecology. Boston: Little, Brown and Company; 1993. p. 1283 – 7.
[15] Manning FA, Harrison MR, Rodeck C. Catheter shunts for fetal hydronephrosis and hydro-
cephalus. Report of the International Fetal Surgery Registry. N Engl J Med 1986;315:336 – 40.
[16] Clewell WH. Presentation to International Fetal Medicine and Surgery Society, Jasper, Al-
berta, 1985.
[17] Goldstein P, Taylor WS, Zisow D, Carson B, Shuster E, Brodner R. Ventriculoamniotic shunt for
treatment of hydrocephalus in one of twins: medical, ethical, and legal considerations. Fetal
Diagn Ther 1990;5:84 – 91.
[18] Cardoza JD, Goldstein RB, Filly RA. Exclusion of fetal ventriculomegaly with a single measure-
ment: the width of the lateral ventricular atrium. Radiology 1988;169:711 – 4.