Pediatr Clin N Am 51 (2004) 305 – 325

Hydrocephalus
Hugh J. L. Garton, MD, MHSca,*
Joseph H. Piatt, Jr, MD, FAAPb,c
a
Department of Neurosurgery, University of Michigan, Taubman 2128/0338,
1500 E. Medical Center Drive, Ann Arbor, MI 48105, USA
b
Section of Neurosurgery, St. Christopher’s Hospital for Children, Erie Avenue at Front Street,
Philadelphia, PA 19134, USA
c
Departments of Neurological Surgery and Pediatrics, Drexel University College of Medicine,
Philadelphia, PA, USA

Hydrocephalus is not an exotic condition in general pediatric practice, but data

on which to base calculations of the incidence and prevalence of hydrocephalus
are scarce. Working with the 1988 National Health Survey of United States non-
institutionalized population, Bondurant [1] estimated that there were 56,566
children 18 years or younger with a cerebrospinal fluid (CSF) shunt in place.
Taking a denominator from 1988 census data, one can calculate a prevalence of
hydrocephalus of 0.9 per 1000 children or 22.5 affected children per 100,000 total
population. Data from population-based studies in Sweden suggest the incidence
of infantile hydrocephalus rose from 0.48 per 1000 live births between 1967 and
1970 to 0.63 per 1000 live births between 1979 and 1982; the increase was
attributed to survival of premature infants with intraventricular hemorrhage (IVH)
[2]. The pediatric Shunt Design Trial, a multicenter trial conducted in North
America and Europe, enrolled only newly diagnosed patients undergoing initial
shunt insertion [3]. The condition categorized by the Swedish population-based
study as infantile hydrocephalus accounted for about half of all patients in the
Shunt Design Trial, suggesting that the overall prevalence of childhood hydro-
cephalus is about twice the Swedish incidence, in the neighborhood of 1.2 per
1000 children assuming minimal mortality after shunt placement. If an ordinary
pediatric practice carries between 2000 and 4000 patients per physician, and if

* Corresponding author. Department of Neurosurgery, University of Michigan, Taubman 2128/

0338, 1500 E. Medical Center Drive, Ann Arbor, MI 48105.
E-mail address: [email protected] (H.J.L Garton).

0031-3955/04/$ – see front matter D 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.pcl.2003.12.002
306 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

affected children are distributed uniformly among pediatric practices (probably

not a valid assumption), a general pediatrician might expect to serve 2 to 5
children with CSF shunts.

Basic physiology
Hydrocephalus is a disturbance of CSF physiology. The secretion of CSF by
the choroid plexus is a metabolically active process involving ion pumps and
enzyme systems similar to those found in the distal tubule of the kidney. CSF is
indistinguishable from brain extracellular fluid, and because water and electro-
lytes pass freely in and out of the brain across the ependymal surfaces of the
ventricular system, the brain itself is believed to be responsible for a small frac-
tion of total CSF production. CSF secretion continues at an essentially constant
rate, about 20 mL/hour in adult humans, regardless of intracranial pressure (ICP)
as long as the choroid plexus and the brain itself are perfused. Age, body mass,
and various disease states undoubtedly affect the rate of CSF secretion, but
methodologies for studying these affects in humans are problematic. Unlike CSF
secretion, CSF reabsorption is a purely passive process driven in a linear fashion
by the pressure differential between the subarachnoid space and the venous
circulation, specifically, the major dural venous sinuses within the cranial cavity.
Thus the intradural compartment does not stray far from a steady state charac-
terized by equal CSF secretion and reabsorption and ICP within the normal range.
With the rare exception of choroid plexus papilloma, a tumor of the choroid
plexus that causes excessive CSF secretion, the diseases that cause hydrocepha-
lus do so by interfering with CSF reabsorption. A higher pressure gradient is
required to drive CSF back into the venous circulation, so, although all but the
most acutely unstable patients with hydrocephalus eventually achieve a steady
state between CSF secretion and reabsorption, they do so only at an abnormally
high ICP.
Out of respect for historical tradition, hydrocephalus sometimes been charac-
terized further as either communicating or obstructive. These terms date to the era
of pneumoencephalography. If air introduced into the lumbar theca appeared
eventually within the ventricles of the brain, the ventricles and the lumbar sub-
arachnoid space were said to communicate. If air filled the dilated ventricles,
such communicating hydrocephalus was presumed to be caused by obliteration of
the subarachnoid spaces or the arachnoid granulations (the sites of CSF reab-
sorption in the dural venous sinuses) by diffuse inflammatory disease processes,
such as meningitis or subarachnoid hemorrhage. Cases of hydrocephalus in which
the ventricles could not be filled with air introduced from below were attributed
to lesions obstructing bulk flow of CSF. This dichotomy has limited useful-
ness today. Determining the category to which a patient belongs on the basis of
history, physical examination, and even noninvasive imaging studies can be dif-
ficult, as contemporary experience with endoscopic third ventriculostomy (ETV)
has shown.
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 307

Box 1. Causes of hydrocephalus

Prematurity (posthemorrhagic hydrocephalus)
Myelomeningocele
Other congenital or developmental conditions affecting the brain
Dandy-Walker malformation
Arachnoid cysts
Interhemispheric cysts
Aqueductal stenosis
Encephalocele
Brain tumor
Subarachnoid hemorrhage
Traumatic brain injury
Aneurismal subarachnoid hemorrhage
Congenital or developmental conditions affecting the skull
Crouzon, Pfeiffer syndromes
Achondroplasia
Meningitis

Causes of hydrocephalus
Hydrocephalus can present in all age groups; however, the management and
prognosis differ significantly depending on the cause and age at presentation.
Common causes for hydrocephalus are listed in Box 1.
The patterns of hydrocephalus encountered in an individual practice or at a
particular institution may vary widely depending on programmatic or referral
factors. The older literature, mostly reports of institutional experiences, cites
myelomeningocele as the leading distinct cause of childhood hydrocephalus
[4,5], but in more recent, multicenter treatment trials, posthemorrhagic hydro-
cephalus of prematurity is at least as common [3,6,7]. Posthemorrhagic hydro-
cephalus is reviewed in detail in this article. The management of the newborn
with myelomeningocele is discussed elsewhere in this volume, as is the man-
agement of antenatally diagnosed congenital hydrocephalus.

Intraventricular hemorrhage and posthemorrhagic hydrocephalus

Incidence and pathophysiology

Hydrocephalus appears most commonly in the newborn period as a result of
an IVH originating from periventricular germinal matrix. In the premature infant
in particular, the walls of blood vessels in the germinal matrix region lack certain
structural elements present in more mature vessels, and they lack substantial
308 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

external tissue support. They supply blood to the rapidly dividing cells of the
germinal matrix, which is the site of origin for a both neuronal and glial cells
ultimately destined for cortex. These fragile vessels are exposed to arterial and
venous hemodynamic surges in the premature infant and can rupture, usually
within the first 72 hours of life [8].
The severity of the hemorrhage is traditionally grade by Papile’s criteria [9].
There are several modifications of this scale, but as reported by Whitelaw
[8], grade 1 is germinal matrix hemorrhage without extension into the ventricle,
grade 2 is IVH involving up to 50% of ventricular area and not dilating the
ventricle, and grade 3 is IVH involving greater than 50% ventricular area and
dilating the ventricle. Grade 4 is traditionally considered to have an intrapa-
renchymal component outside the germinal matrix region, although Volpe [10]
argues that intraparenchymal hemorrhage should be reported separately from the
grading of the IVH. In addition, Volpe’s interpretation of IVH grades does not
specifically refer to ventricular dilatation as part of the criteria, limiting consid-
eration to percentage of ventricular volume occupied by the clot.
The mechanism by which IVH leads to hydrocephalus has been conjectured
to be occlusion of the arachnoid granulations by the breakdown products of
hemorrhage. Autopsy studies, however, suggest that the arachnoid granulations
are poorly developed in the prenatal period [11], and alternative lymphatic, peri-
vascular, and dural pathways may be involved. Although a sudden increase in
ventricular size can occur after IVH, presumably caused by focal obstruction of
the intraventricular outflow tract, more commonly there is a lag of days to weeks
between the identification of the hemorrhage and the subsequent development of
ventricular dilatation. This delay suggests that the breakdown products of the
hemorrhage may be responsible. In addition to mechanical obstruction of the
outflow pathways, it is possible that growth factors such as transforming growth
factor-beta 1, which is elevated in CSF after IVH and is highest in patients
subsequently requiring shunt placement [12], together with factors present in CSF
as a consequence of the hemorrhage such as thrombin, promote the growth of
connective tissue within the leptomeninges, leading to an arachnoiditis and
contributing to reduced CSF absorption [13].

Evaluation
In most cases, IVH occurring in premature infants is diagnosed on routine
ultrasound performed within 3 days of birth. In other cases the IVH may be
suspected on the bases of neurologic activity such as seizures, depression of re-
sponsiveness, or an unexplained drop in hematocrit. Hydrocephalus occurs as a
consequence of IVH variably and in a fashion predicted in large part by the grade
of IVH. Among very low birth weight infants (< 1500 g), the incidence of IVH is
approximately 22%. For example, in a natural history study of 248 infants (mean
gestation, 26.8 weeks) Murphy and colleagues [14] found a 25% incidence of
progressive posthemorrhagic ventricular dilatation (PVD). Of these patients, 38%
had spontaneous arrest of PVD; 48% were treated pharmacologically with no
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 309

Fig. 1. Coronal and Sagittal measurements of ventricular size. DW, diagonal width (anterior horn
width); VH, ventricular height; VI, ventricular index.

need for shunt; 34% needed shunt or a reservoir; and 18% died [14]. Among the
most severely affected children – those with a birth weight below 1500 g– the
incidence of PVD can be estimated according to IVH grade: 4% had grade I
PVD, 11% had grade II PVD, 76% had grade III PVD, and 71% had grade IV
PVD[14]. The natural history of PVD is variable, and it does not inevitably lead
to the need for surgical intervention. Some patients demonstrate an arrest or even
reduction in ventriculomegaly with no treatment, with medical management
alone, or with serial lumbar punctures. In the study by Murphy and colleagues
[14], shunt rates were reported by grade for infants with birth weights below
1500 g: shunts were required in none of 94 infants (0%) with grade I PVD, in 1 of
52 infants (2%) with grade II PVD, in 10 of 31 infants (32%) with grade III PVD,
and in 8 of 17 infants (47%) with grade IV PVD [14].
In monitoring children with PVD, the authors’ practice is to obtain head ultra-
sound studies twice weekly for several weeks after the IVH is detected. Besides
observing the general size and configuration of the ventricles, several standard
measurements can be obtained on the ultrasound (Fig. 1). These measurements
include the ventricular index, measured in the coronal plain from the falx cerebri to
the lateral aspect of the ventricle at the level of the foramen of Munro just in front
of the choroids plexus. More recently, ventricular height and diagonal ventricular
width (anterior horn width) have been proposed as more sensitive measures of
ventricular dilation and apparently correlate better with expert clinical impression
of ventricular size [15,16]. Readers of the medical literature on this subject are
cautioned that the term ventricular width is often used without specific reference
to exactly what is being measured, and care is required in interpreting and using
such studies.

Prognosis
Epidemiologic studies in southwestern Sweden suggest that for children born
between 1967 and 1982 the prognosis after IVH was very poor, with 18 of 23
310 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

(78%) surviving preterm infants with IVH showing significant neurologic dys-
function by age 2 years [17]. In a subsequent study of former premature infants
with infantile hydrocephalus born during the same epoch, 16 of 61 patients (26%)
died before 2 years of age. Among the 45 survivors, 47% had cerebral palsy, 51%
had mental retardation, and 33% had epilepsy [18].
Distinguishing the outcome of posthemorrhagic hydrocephalus from the out-
come of IVH is difficult. For all neonates with IVH, the grade of IVH correlates
with outcome. Volpe [10], summarizing date from 20 publications, estimated the
incidence neurologic sequelae for patients with IVH of varying severity: grade I,
5%; grade II, 15%; grade III, 35%; and grade IV, 90%. Guzzetta and colleagues
[19] reported prognosis after IVH was greatly influenced by the presence of
intraparenchymal hemorrhage. Among surviving patients, 79% had a major
motor deficit, and just under half had significant cognitive deficits. At the other
end of the spectrum, in one series 9 of 12 patients with grade II IVH evaluated
with standard developmental instruments had normal intelligence, all were am-
bulatory, and one had spastic diplegia [20].
Neurodevelopmental outcomes of children with transient PVD have been
compared with the outcomes of children whose PVD was persistent but not
symptomatic enough to prompt treatment. Of 52 patients with arrested progres-
sion of hydrocephalus (transient PVD) during the neonatal period, 52% had nor-
mal developmental outcomes compared with 37% of patients with asymptomatic
severe hydrocephalus (persistent but asymptomatic PVD). The mean IQ for the
two groups was similar (73 versus 75, respectively) [21].
With more specific regard to posthemorrhagic hydrocephalus, Pikus [52]
evaluated 52 patients with grade IV IVH and progressive hydrocephalus treated
initially between 1977 and 1987. Mortality was 60%, and 78% of survivors had
intellectual function more than 2 SD below the mean for age. All survivors had
some form of spasticity. These children underwent an average of 6.9 shunt
revisions over an 8- to 18-year follow-up. In an accompanying literature review,
the overall mortality for grade IV IVH seems to have fallen in the 1990s as
compared with series published in the 1980s, but a trend toward improving
cognitive outcomes is less clearly evident [22]. Boynton [23] reported outcomes
for 50 preterm infants with posthemorrhagic hydrocephalus, 92% of whom had
either grade III or IV hemorrhages, treated between 1978 and 1984. All but one of
this cohort weighed less than 2000 g at shunt insertion. Mean follow-up was not
reported but probably ranged from 2 to 8 years. There was 7% mortality, with a
median shunt revision rate of four per patient and a shunt infection rate of 19%
per procedure, with a shunt failure rate of 92% in children treated in the first
3 weeks of life. Neurodevelopmental outcomes included blindness (28%), strabis-
mus (40%), and hearing loss (24%). Seizures occurred in 38% of patients, with
most requiring chronic anticonvulsants. Significant limitations in motor function
occurred in 49%, whereas 47% and 24% had severe and moderate developmental
delays, respectively.
At least one recent series paints a somewhat less bleak picture [24]. In a series
of 76 children treated between 1984 and 1999 for IVH, 42 developed hydro-
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 311

cephalus (2 with grade II, 23 with grade III, and 7 with grade III or higher pe-
riventricular leukomalacia [PVL], and 10 with grade IV IVH). With a mean
follow-up of 8 years, mortality was 7% (3 patients), the rate of shunt infection
was 7%, and the mean number of shunt revisions per patient was 1.57. In the
surviving patients, motor function was normal in 15 patients (35%), mildly
impaired without functional significance in 8 patients (19%), and moderately or
severely impaired in the remaining 19 patients (45%). No patient with normal
motor function at follow-up had had PVL. On developmental assessments using a
standardized instrument, 13 (30%) were normal, 12 had minor delays, and the
remaining 14 living patients (33%) had major delays. In 7 patients with PVL, the
outcome was worse than in patients with similar IVH grades for both motor and
developmental assessments [24]. The association of PVL with outcome indepen-
dent of IVH grade has been borne out by other studies, most recently in the
8-year follow-up report from prophylactic indomethacin trial, in which grade III
and grade IV IVH and PVL were independent predictors of poor outcome on a
battery of cognitive tests [25].

Prevention
Ideally the best solution to posthemorrhagic hydrocephalus is to prevent the
hemorrhage from occurring in the first place. Unfortunately the remarkable
successes of neonatologists in the management of pulmonary immaturity have
not been repeated with IVH. Antenatal administration of phenobarbital [8] and
vitamin K [26] has not shown benefit in a series of randomized, controlled trials.
Antenatal corticosteroids have been shown to reduce the incidence of IVH in
premature children in a meta-analysis of previously published trials. The combined
odds ratio for development of IVH was 0.38 in a comparison of antenatal
corticosteroids with placebo [27]. Postnatal interventions to protect the developing
infant have included the use of phenobarbital, indomethacin, vitamin E, ethamsy-
late, and pharmacologic paralysis. Indomethacin administered antenatally has
produced modest reductions both for all degrees of IVH and for severe IVH.
Cognitive improvements have yet to be convincingly demonstrated. Similarly
vitamin E and ethamsylate have demonstrated ability to reduce the incidence of
IVH, but their value in reducing severe IVH or long-term disability is unproven [8].

Medical therapies
Once a hemorrhage has occurred, medical therapies and serial lumbar punc-
tures may contain developing PVD. These strategies are almost always appro-
priate as initial therapy for PVD, except in cases of very rapid expansion of the
ventricular system, when direct access to the ventricular system is probably more
effective. At what point is treatment indicated? No firm evidence-based answer
can be provided to this question. Numerous laboratory studies speak to the
adverse effect of even mild ventricular dilatation [28]; however, clinical studies
demonstrating that aggressive control of ventricular size in the neonatal period
312 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

produces a better outcome are lacking. In this instance the absence of evidence
should perhaps not be construed as evidence of absence of benefit, because of
the difficulty of detecting the potential effect of control of hydrocephalus against
the background noise of the other insults to which the premature brain is subject,
particularly PVL and hypoxic/ischemic encephalopathy. Given these limitations,
one reasonable option is to use the common entry criterion for treatment trials
for IVH, which has typically been a head circumference 4 mm beyond the ninety-
seventh percentile for age [29].
Diuretics, specifically acetazolamide and furosemide, either singly or in com-
bination, have been used in the past to reduce CSF output in cases of PVD. These
agents act on molecular substrates that the choroid plexus shares with the distal
renal tubule, and because they work by different mechanisms, they act syner-
gistically to reduce CSF production significantly in neonates. One small study
(N = 16) showed a nonsignificant trend toward reduced shunt rates with diuretic
therapy [30]. A much larger clinical trial (N = 177), however, reported an in-
crease in mortality (19% versus 13%), greater need for CSF shunts (51% versus
38%), and increased neurologic disability at 1 year in the diuretic plus standard
management arm of the trial, as opposed to standard management with serial
lumbar punctures [31,32].

Fibrinolytics
The literature on the use of fibrinolytics to prevent the development of
hydrocephalus after IVH is conflicting. A trial comparing intraventricular uro-
kinase given for IVH at the point of detection of ventriculomegaly demonstrated
no reduction in shunt rate over placebo [33]. Hudgins [34] administered low-
and high-dose urokinase to 18 premature infants with PVD. Compared with
historical controls, a smaller fraction of the low-dose group required permanent
CSF shunts. The urokinase patients who ultimately needed permanent shunts
also required fewer subsequent shunt revisions than the historical controls. Un-
controlled experiences with urokinase have yielded similar results [35]. A review
of the controlled and uncontrolled evidence published to 1999 suggested that
no conclusions could be drawn with respect to efficacy [36]. Whitelaw and
colleagues [29] recently completed a phase I trial of a treatment regime of intra-
ventricular tissue plasminogen activator (TPA) followed by 72 hours of con-
tinuous ventricular irrigation with inlet/outlet frontal and occipital catheters in
a population of premature infants with grade IIII and IV IVH. Of the 24 patients,
1 died, 2 had repeat hemorrhages, but only 6 (26% of survivors) required shunt
placement. Among 19 survivors at 1-year follow-up, 8 were normal, 7 had a
single disability, and 4 had multiple disabilities. Two central nervous system
infections occurred [29]. The conceptual problem with all fibrinolytic therapies is
that many patients must be subjected to the risks of invasive neurosurgical
procedures to prevent hydrocephalus that is otherwise destined to develop in only
a fraction of them. Whether the relatively low complication rates reported from
protocol-driven care at the hands of interested surgeons and experienced neonatal
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 313

ICU personnel can be duplicated in ordinary clinical settings is doubtful as well.

Thus in the absence of supportive data from properly controlled, multicenter
clinical trials, fibrinolytic therapy cannot be recommended.

Sequential lumbar punctures

Sequential lumbar punctures are popular as a tactic to prevent posthemor-
rhagic hydrocephalus after high-grade IVH and as a therapy to manage symp-
tomatic hydrocephalus after it has become established. In the prophylactic
approach, CSF is drained on a schedule in the hope that its removal will maintain
ventricular size close to normal and that early evacuation of the byproducts of the
IVH will reduce the likelihood of subsequent hydrocephalus. In symptomatic
hydrocephalus, CSF drainage is instituted to control the symptoms and signs of
active hydrocephalus, such as accelerated head growth, tension of the anterior
fontanel, separation of sutures, apnea, bradycardia, seizures, or rapidly progres-
sive ventriculomegaly on ultrasound. Additional indications could also include
deteriorating somatosensory-evoked potential responses or decreased diastolic
flow velocity on the intracranial vasculature on Doppler analysis [37]. The two
strategies have been compared in several randomized, controlled trials. The larg-
est and most recent of these is the Ventriculomegaly Trial Group, which com-
pared 157 infants with IVH enrolled from 1984 to 1987 who were randomly
assigned to early use of lumbar punctures, performed as often as needed to
prevent ventricular expansion, or to a standard management regimen that called
for lumbar puncture based on head expansion or clinical symptoms or signs of
increased pressure [38]. Early lumbar punctures were successful in restricting
ventricular expansion compared with conservative management. At the 1-year
outcome assessment, however, 62% of each group had required shunt placement,
death rates were similar (15% versus 17%), and ventriculitis was more common
(9% versus 5%) in the early-treatment group. There were no significant differ-
ences in functional outcome at 1 year [38]. Functional outcome could in theory
have been affected adversely by the increased use of ventricular puncture in the
early-drainage group (41% versus 20% had at least one ventricular puncture).
More recently, a meta-analysis of this and three other smaller trials supported this
study’s conclusions that early tapping offers no clear benefit [37]. In addition, the
Ventriculomegaly Trial Group results offer no support to the proposition that
modest ventriculomegaly in the neonate is, independent of other risk factors, a
cause of neurologic injury.
In routine clinical practice the decision to employ sequential lumbar punctures
should be guided by all the available information including fontanel examina-
tion, head size change, neurologic and cardiovascular symptoms, and ultrasound
studies. Typically, the aim is to remove enough fluid to render the fontanel soft
and sunken, usually in the range of 10 cm3/kg/tap. The frequency of the taps is
determined on the basis of the assessment described previously. The use of ven-
tricular, transfontanel taps has been advocated when lumbar punctures are
ineffective in containing ventricular size or controlling symptoms. Typically
314 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

these procedures are reserved for situations in which surgical ventricular catheter
placement is contraindicated by urgency, operating room inaccessibility, or CSF
infection. Ventricular cannulation can be achieved with a small spinal needle with
entrance through the lateral corner of the fontanel. The experienced practitioner
can perform ventricular puncture either by dead reckoning based on anatomic
landmarks or by ultrasound guidance. Typically 10 cm3/kg is withdrawn, al-
though this volume can be modified based on fontanel tension, and the proce-
dure should be discontinued if bradycardia ensues. Transcortical ventricular
puncture in the neonate has been associated with the development of porence-
phalic cysts, although the published literature on this topic is based largely on
ventriculographic and autopsy studies from the 1960s. In the most frequently
cited of these studies, nearly 60% of children with multiple ventricular punctures
developed porencephalic cysts. Most of these children, however, had pyogenic
ventriculitis and had been treated with intraventricular antibiotics and cortico-
steroids; the role that the illness or the associated treatments may have played in
the porencephaly cannot be determined [39].

Neurosurgical management
In patients who have rapidly progressive ventriculomegaly or progressive or
symptomatic ventriculomegaly despite lumbar punctures, placement of a direct
ventricular drainage system is indicated. Although definitive treatment with a
ventriculoperitoneal shunt can be considered, most patients in this situation are
still quite small and have persistently proteinaceous and cellular CSF. They are at
high risk for shunt obstruction and shunt infection [40,41]. Furthermore, not all
infants who need ventricular drainage at some point in their preterm care will
evolve to permanent, shunt-dependent hydrocephalus by term, so a technique for
temporary ventricular decompression is expeditious. There are three principle
options: an external ventricular drain, a subgaleal ventricular reservoir or so-
called ‘‘ventricular access device,’’ and a subgaleal shunt. Although favorable
experiences have been reported [42 –45], external ventricular drainage has many
disadvantages: The drain interferes with nursing care. Because of the small size
of the head, the drain is easily dislodged. Because of low CSF flow volumes,
drain obstruction is frequent. And with prolonged drainage there is a formidable
risk of infection. The writers do not favor external drainage. A ventricular access
device is a small, flat-bottomed reservoir attached to a ventricular catheter. The
reservoir sits on the surface of the skull under the galea of the scalp. Percutaneous
puncture of the reservoir with aspiration of CSF on a daily or every-other-day
schedule serves to keep the ventricular system decompressed. Clinical endpoints
are arrest of ventricular dilatation, control of head growth, and elimination of
symptoms and signs of elevated ICP [46 – 49]. A subgaleal shunt is a ventricular
access device with an outlet (Fig. 2). The surgeon places the reservoir with its
outlet in a large subgaleal pocket created by sweeping a finger or blunt dissecting
instrument over the surface of the skull. Typically the pocket covers the entire
hemicranium on the side of the shunt. Drainage of CSF through the reservoir, out
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 315

Fig. 2. Subgaleal shunt. The reservoir and outlet of the shunt sit in the subgaleal space, and CSF drains
into a subgaleal pocket. The plug is removed if the surgeon intends to use the device as a shunt. The
plug can be left in place, and the device can be used as a simple ventricular reservoir, a so-called
‘‘ventricular access device.’’

the outlet, into the subgaleal pocket decompresses the ventricular system. The
subgaleal space probably has some absorptive capacity, but the pocket also serves
a simple mechanical function as a high-compliance receptacle for ventricular
CSF. While the pocket persists, no needle aspiration is necessary. Over the course
of weeks, scarring of the scalp to the periosteum of the skull slowly obliterates
the pocket, and if the patient continues to require ventricular decompression,
periodic needle aspiration of the shunt reservoir can be initiated, as with the
ventricular access device [50]. The subgaleal shunt obviates the need for frequent
needle procedures, minimizing patient stress and, presumably, infection risk [51].
Infection rates ranging between 0% and 10% have been reported [52]. Some
authors believe that subgaleal shunts control ventricular volume more consist-
ently as well [53]. As many as 25% of patients who require ventricular access
devices or subgaleal shunts for posthemorrhagic hydrocephalus ultimately re-
cover and avoid permanent ventriculoperitoneal CSF shunts [47 – 51,53].
A common practice is to employ one of these temporizing measures until the
infant reaches term and a weight of 2 kg. Delay allows PVD to resolve in some
instances, and in the remaining cases of persistent, active hydrocephalus, delay
allows some clearance of protein and cellular debris from the CSF. High CSF
protein is probably not an adverse factor for shunt function, but high cell counts
seem to be. Surgeons believe that wound dehiscence and skin breakdown over
shunt hardware are less frequent in larger infants as well. Insertion of a
ventriculoperitoneal shunt for management of posthemorrhagic hydrocephalus
in the former premature who has reached term does not differ materially from the
management of other forms of hydrocephalus and is discussed in general terms.
316 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

Cerebrospinal fluid shunts

Components
CSF shunts consist typically of three parts: a ventricular catheter, a valve, and
a distal catheter. The ventricular catheter passes from the ventricle through the
cortical mantle out of the skull through a burr hole to reach the external surface
of the skull, where it is joined to the inlet of the valve. There are a variety of
valve designs that incorporate various clever mechanical mechanisms, but the
purpose of each is to prevent excessive CSF drainage. Excessive ventricular
drainage can be responsible for troublesome complications such as postural
headaches, subdural hemorrhage, and chronic changes in brain compliance that
promote recurrent proximal shunt failure. Most valves include a reservoir that can
be punctured percutaneously with a needle for diagnostic purposes. The valve is
connected in turn to a distal catheter that carries the CSF to a body site where it
can be reabsorbed into the venous circulation at low pressure. Popular sites for
CSF diversion are the peritoneal cavity, the pleural cavity, the gall bladder, the
right atrium of the heart, and the internal jugular vein. Lumbar CSF shunts are
schematically similar except that a lumbar intrathecal catheter takes the place of
the ventricular catheter.

Complications
The attractive feature of the CSF shunt is its nearly universal applicability in
the management of hydrocephalus regardless of cause. The burden of the CSF
shunt is a high rate of complications. The complications fall into two broad
categories: infection and mechanical failure. Infection is believed to be caused
most often by contamination of the shunt at surgery [54 – 56]. CSF fistulization of
the surgical wound, wound dehiscence, and skin breakdown over shunt hardware
account for most of the balance of infectious complications. Consequently, most
shunt infections present within 3 months of the previous surgical procedure, and
almost all present within 6 months. Late CSF shunt infections are curiosities and
probably have different mechanisms. Infection rates observed in recent, multi-
center surgical trials cluster in the neighborhood of 8% to 10% per surgical
procedure [3,6,7]. Mechanical failure occurs most commonly from tissue debris
plugging the lumen of the shunt, but separation of components, fracture, and
migration are other possibilities. Excessive CSF drainage with postural headaches
or subdural hemorrhage is also a mode of mechanical failure. Both retrospective
and prospective studies have consistently documented that 30% to 40% of shunts
fail on either a mechanical or an infectious basis in the first year after initial
placement, and an additional 15% fail in the second year. After 2 years the failure
rate seems to drop to between 1% and 7% per year [3,7,41,57,58]. The mortality
from initial shunt insertion seems to be about 0.1% [3,7,59]. The mortality from
shunt failure is estimated to be approximately 1% to 4% [4,60,61].
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 317

Third ventriculostomy, an alternative to cerebrospinal fluid shunts

The life-long requirement for surgical maintenance entailed in the use of CSF
shunts has driven surgeons to seek alternative treatment strategies. Third ven-
triculostomy is a surgical opening in the floor of the third ventricle that allows
CSF to escape the ventricular system, bypassing obstructive lesions downstream.
Third ventriculostomy through a major craniotomy exposure was originally
described more than 80 years ago, but recent technologic developments have
made endoscopic performance of this procedure through a burr hole feasible and
safe. ETV may be considered for patients with masses or obstructive congenital
lesions in the aqueduct, in the fourth ventricle, or at its outlets. About 65% to 70%
of the best candidates can avoid initial CSF shunt placement with this technique
[62 – 64]. ETV is also being attempted with varying success in a number of other
specific instances, including CSF shunt failure or infection [65,66], hydrocephalus
associated with the Chiari malformation type 1 [67], and occasional cases of
communicating hydrocephalus, where it is theorized that the obstruction is within
the subarachnoid spaces of the posterior fossa [68]. (The old diagnostic categories,
obstructive and communicating hydrocephalus, have been recast in modern
neurosurgical practice as the operational categories of ETV-responsive and
ETV-nonresponsive hydrocephalus.) Success rates for these indications are lower
(eg, 23% for neonates with IVH) [69]. ETV has a higher operative risk than shunt
placement. Complications including subarachnoid hemorrhage, basilar artery
injury, and hypothalamic/pituitary dysfunction have been reported. Mortality after
the procedure has been estimated at 1% [70]. Most failures of this therapy occur
shortly after the procedure is performed, although at least one failure occurring 6
years after the procedure has been reported [64]. In one large institutional series,
the durability of ETV did not seem to be any better than that of CSF shunts [71].
ETV must not be considered by the physician or presented to the family as a
permanent cure, and patients with hydrocephalus who have undergone ETV
require the same life-long neurosurgical follow-up as patients with CSF shunts.
Formal explorations of the risk/benefit trade-offs for ETV as compared with CSF
shunt treatment are limited at this point, but no substantial advantages for ETV
within the initial few years after treatment begins have been demonstrated. Longer
follow-up studies are necessary [72,73].

Recognition of complications
The challenge for the primary pediatrician is to distinguish the symptoms
and signs of CSF shunt failure from the symptoms and signs of more common
childhood conditions such as teething, viral respiratory and gastrointestinal syn-
dromes, and otitis. Two nonspecific factors should heighten the pediatrician’s
level of suspicion. The first is young age. Infants with shunts making sick visits
to the office are significantly more likely to have shunt failure than older patients
with shunts [41,60]. The second is the elapsed time since the most recent shunt
operation. Patients making sick visits within 6 months of surgery are much more
318 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

likely to have shunt failure than patients who have been neurosurgically stable for
longer than 6 months [41].
The presentations of CSF shunt failure generally fall into one of three modes:
acute ICP elevation; chronic, insidiously progressive ICP elevation, and infection.
The symptoms and signs of acute ICP elevation are familiar: headache, nausea,
vomiting, acute papilledema, and occasionally VI nerve palsy. Parental descrip-
tions of distention of scalp veins or puffiness around the eyes should not be
dismissed, even though the physiologic mechanisms and the findings themselves
may not be obvious to the physician. They should be considered symptoms rather
than signs. Headache awakening the patient from sleep in the early morning hours
is particularly suggestive, as is headache exacerbated by recumbency, straining, or
exertion. Symptoms and signs associated with a chronic course of shunt failure
include accelerated head growth, loss of developmental milestones or difficulty in
school, and chronic papilledema or optic atrophy. In addition to fever, signs that
call attention to the possibility of infection are redness and swelling at the surgical
sites, drainage of pus or, more likely, CSF from a wound, nuchal rigidity, ab-
dominal pain, and signs of peritoneal irritation. Patients with related neurologic
conditions may manifest CSF shunt failure in ways particular to those underlying
conditions. The patient with epilepsy, for example, may come to attention with an
otherwise unexplained increase in seizure frequency [74]. A patient with
myelomeningocele may present with dysphonia, dysphagia, or disturbances of
ventilatory regulation related to the underlying Chiari malformation type 2 or with
progressive myelopathy related to underlying syringomyelia.
Factors specific to the shunt itself deserve consideration. Patients with pleural
CSF shunts may complain of dyspnea and pleuritic pain if failure of reabsorption
of CSF leads to a large pleural collection. Patients who have had atrial CSF
shunts in place for some years are at risk for chronic pulmonary thromboembo-
lism and cor pulmonale. Unfortunately, this condition seldom comes to atten-
tion before it is irreversible [75]. Chronic, indolent infection of a ventriculoatrial
shunt can lead to glomerulonephritis, so-called ‘‘shunt nephritis,’’ on the basis of
antigen-antibody complex deposition. The usual offending organism is coagu-
lase-negative staphylococcus. Physical examination of the shunt may sometime
be informative. Persistence of fluid along the course of the shunt more than 2 or
3 weeks after surgery is suspicious. Old shunts can become brittle and fracture,
in which case a palpable gap may be identified between the fracture fragments.
Most shunt valves include a pump mechanism for moving fluid orthograde
through the shunt at the time of surgery. Attempts are sometimes made to extract
diagnostic information from the behavior of the pump mechanism. If the pump
reservoir can be depressed with little resistance, one might think that the shunt is
patent distally, and if the pump reservoir refills promptly, one might think that the
shunt is patent proximally. Unfortunately, even in the hands of the pediatric
neurosurgeon, pumping the shunt is not a reliable diagnostic test [76,77]. A
decision to forego further investigation of possible CSF shunt failure must never
be made solely on the basis of examination of the valve. Finally, a symptom
generally not associated with shunt failure is pain along the course of the shunt.
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 319

The shuntalgia syndrome presents with a highly focal discomfort at the site of the
valve or, commonly, along the distal catheter in the posterior triangle of the neck.
There may be associated tenderness, and a hard, fibrotic sleeve of subcutaneous
scar tissue may be palpable around the shunt catheter, but there is never any
associated swelling, fluctuance, or redness. Shuntalgia is most common among
adolescents and may be related to tethering of the shunt in the subcutaneous
tissues during growth spurts. The syndrome is self-limiting, but the discomfort is
often annoying and resistant to nonnarcotic analgesics. It may infrequently herald
shunt fracture or separation of shunt components, so neurosurgical reassessment
is appropriate.
In this era of evidence-based medicine, the evaluation of the child who may
have CSF shunt failure deserves to be put on a rational footing. Unfortunately
research in this area is limited, and the available data are not yet readily ap-
plicable to routine pediatric practice [73,78,79]. Estimates of the sensitivity
and specificity of certain symptoms, signs, and diagnostic tests are available.
What most practitioners do not know, however, is the prevalence of real CSF
shunt failure among the patients presenting to their offices for investigation
for possible CSF shunt failure. In the jargon of evidence-based medicine, the
likelihood ratio attached to some symptom, sign, or test result may be known,

Table 1
Predictive values for symptoms and signs of shunt failure in a group of predominantly very young
children (Median age 4.6 months) seen within the first 5 months after initial shunt insertion
PPV % LR LR Sensitivity Specificity
Symptom or Sign (95% CI) 1-NPV (95% CI) Neg (%) (%)
Nausea and vomiting 79 (61 – 91) 19 10.4 (4.7 – 23.0) 0.7 36 97
Headache 57 (29 – 82) 21 4.4 (1.6 – 12.1) 0.9 15 97
Irritability 78(61 – 90) 18 9.8 (4.7 – 20.5) 0.6 39 96
Loss of milestones 20 (3 – 56) 25 0.8 (0.2 – 3.5) 1.0 3 96
Decreased level 100 (59 – 100) 25 NR 0.9 10 100
of consciousness
Bulging fontanel 92 (79 – 98) 15 33.1(10.5 – 104.2) 0.5 51 98
Papilledema 100 (16 – 100) 27 NR 1 3 100
Abnormal increase 67 (51 – 100) 17 5.7 (3.3 – 9.9) 0.6 46 92
in OFC
Fluid tracking around 75 (51 – 91) 23 8.3 (3.1 – 22.0) 0.8 21 98
the shunt
Shunt reservoir does 80 (28 – 99) 21 13.4 (1.5 – 115.8) 0.9 11 99
not depress
Shunt reservoir does 86 (42 – 100) 20 19.9 (2.5 – 159.7) 0.8 17 99
not refill
Fever 89 (67 – 99) 22 23.6 (5.6 – 99.8) 0.8 23 99
CI, confidence interval; LR, Likelihood ratio for presence of symptom; LR Neg, likelihood ratio for
the absence of symptoms; NPV, negative predictive value (1-NPV: % without symptom who will still
have shunt failure); NR, result could not be calculated because of division by zero; PPV, positive
predictive value (% with symptom who will have shunt failure).
Modified from Garton HJ, Kestle JR, Drake JM. Predicting shunt failure on the basis of clinical
symptoms and signs in children. J Neurosurg 2001;94(2):202 – 10.
320 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

but the pretest odds ratio is usually not known [80]. For example, Barnes and
colleagues [81] evaluated the clinical presentations of 53 children presenting to a
hospital unit dedicated to CSF shunt management. The predictive value of the
combination of nausea, vomiting, and drowsiness was high: 82% of patients with
this constellation of symptoms were found to have a shunt failure. Because of
the specialized nature of the unit, however, fully 69% of all patients were found
to have shunt failure, so the high rate of prediction is unsurprising and difficult to
translate into a more typical pediatric office or emergency practice, where most
children presenting with similar symptoms will have alternative diagnoses. Simi-
larly, Garton and colleagues [79] evaluated 276 young children being seen by
neurosurgeons in the first 5 months after shunt insertion as part of a controlled
clinical trial; 26% of these children were determined to have shunt failure at the
time of the encounter [3,79]. Table 1 lists the predictive values for various
symptoms and signs of shunt failure for this cohort of patients. For example,
patients presenting with irritability had a 78% chance of having a shunt failure.
The direct applicability of these data to pediatric office or emergency room prac-
tice can be questioned because of the way in which the study cohort was as-
sembled, and the reader will note the relatively wide confidence intervals around
the predictive values, but it is clear that patients presenting with typical symp-
toms and signs early after shunt placement have a high likelihood of malfunc-
tion. If an alternative explanation is not readily at hand, neurosurgical referral is
highly advisable.

General principles and aphorisms

As in so many areas of pediatric office practice, when the child with
hydrocephalus makes a sick visit, the most important question is whether the
presenting symptoms reflect a dangerous underlying disease process. If a firm
alternative diagnosis can be established on the basis of positive evidence, no
further study of the shunt is indicated, but careful observation until symptoms
resolve is prudent. The question of shunt failure is never really closed until the
patient is well again. If no alternative diagnosis can be made, the next step must
be referral back to the treating neurosurgeon.
Box 2 lists diagnostic tests that the neurosurgeon may use to assess
shunt function.
As is true of the symptoms and physical signs of shunt failure, these tests each
have sensitivities and specificities that are lower than one might wish. There is no
definitive standard. Brain imaging deserves particular mention. When a CSF
shunt is functioning properly, ventricular volume decreases in comparison with
the pretreatment state. When a CSF shunt fails, ventricular volume usually but
not always increases in comparison with the symptom-free, posttreatment
baseline. Brain imaging at the time of symptoms cannot be interpreted without
comparison to posttreatment baseline imaging, and even if previous images are
available, the changes in the appearance of the brain caused by shunt failure may
be subtle or nil. Iskandar [82] reviewed the written interpretations of brain
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 321

Box 2. Diagnostic measures useful in the evaluation of

cerebrospinal fluid shunt function
Radiography of the shunt system
Brain imaging
Ultrasound (early infancy)
CT
MRI
Shunt tap
Radionuclide shuntography
Intracranial pressure monitoring

imaging studies in an institutional series of 68 patients with CSF shunt failure

confirmed by surgery and subsequent resolution of symptoms. In 24% of the
interpretations there was no mention of the possibility of shunt failure. In another
study 16% of surgically proven shunt failures had CT scans that were unchanged
from previous baseline studies [81]. Pediatricians should not substitute brain
imaging for neurosurgical consultation.
Pediatricians (and neurosurgeons) are wise to pay attention to the observations
of experienced parents. In a review of admissions for possible CSF shunt failure
at The Hospital for Sick Children at Great Ormond Street, London, parents’
opinions were found to have greater diagnostic accuracy than the opinions of
referring pediatricians [78]. Parents were about as accurate as brain imaging, al-
though both produced false negatives.
Finally, one aspect of creating a medical home for the child with complex,
chronic illness is referral for appropriate subspecialty care and establishment of a
communicative relationship among all the involved parties: the family, the
pediatrician, and the subspecialist [83]. The family of the child with hydroceph-
alus usually becomes attached to the treating neurosurgeon, but in this mobile
society, such relationships seldom last forever. Relocations for employment and,
regrettably, changes of insurance often necessitate transfers of medical care. The
receiving pediatrician provides an important service to the new patient with
hydrocephalus by consulting the neurosurgeon for a baseline evaluation before
problems develop. Evaluation of the ill child with symptoms consistent with acute
CSF shunt failure is greatly facilitated by the presence of a previous neurosurgical
encounter documenting neurologic and imaging findings when the child was well.

References
[1] Bondurant C, Jimenez D. Epidemiology of cerebrospinal fluid shunting. Pediatr Neurosurg
1995;23:254 – 8.
[2] Fernell E, Hagberg B, Hagberg G, von Wendt L. Epidemiology of infantile hydrocephalus in
Sweden. I. Birth prevalence and general data. Acta Paediatr Scand 1986;75(6):975 – 81.
322 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

[3] Drake JM, Kestle JR, Milner R, Cinalli G, Boop F, Piatt J, et al. Randomized trial of cerebro-
spinal fluid shunt valve design in pediatric hydrocephalus. Neurosurgery 1998;43(2):294 – 303.
[4] Sainte-Rose C, Piatt JH, Renier D, Pierre-Kahn A, Hirsch JF, Hoffman HJ, et al. Mechanical
complications in shunts. Pediatr Neurosurg 1991;17(1):2 – 9.
[5] Piatt JH, Carlson CV. A search for determinants of cerebrospinal fluid shunt survival: retrospec-
tive analysis of a 14-year institutional experience. Pediatr Neurosurg 1993;19(5):233 – 41.
[6] Pollack IF, Albright AL, Adelson PD. A randomized, controlled study of a programmable shunt
valve versus a conventional valve for patients with hydrocephalus. Hakim-Medos Investigator
Group. Neurosurgery 1999;45(6):1399 – 408.
[7] Kestle JR, Drake JM, Cochrane DD, Milner R, Walker ML, Abbott III R, et al. Lack of benefit
of endoscopic ventriculoperitoneal shunt insertion: a multicenter randomized trial. J Neurosurg
2003;98(2):284 – 90.
[8] Whitelaw A. Intraventricular haemorrhage and posthaemorrhagic hydrocephalus: pathogenesis,
prevention and future interventions. Semin Neonatol 2001;6(2):135 – 46.
[9] Papile LA, Burstein J, Burstein R, Koffler H. Incidence and evolution of subependymal and
intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm. J Pediatr
1978;92(4):529 – 34.
[10] Volpe JJ. Neurology of the newborn. Philadelphia: WB Saunders; 2001.
[11] Fox RJ, Walji AH, Mielke B, Petruk KC, Aronyk KE. Anatomic details of intradural channels
in the parasagittal dura: a possible pathway for flow of cerebrospinal fluid. Neurosurgery 1996;
39(1):84 – 90.
[12] Whitelaw A, Christie S, Pople I. Transforming growth factor-beta1: a possible signal molecule
for posthemorrhagic hydrocephalus? Pediatr Res 1999;46(5):576 – 80.
[13] Motohashi O, Suzuki M, Yanai N, Umezawa K, Shida N, Yoshimoto T. Thrombin and TGF-
beta promote human leptomeningeal cell proliferation in vitro. Neurosci Lett 1995;190(2):
105 – 8.
[14] Murphy BP, Inder TE, Rooks V, Taylor GA, Anderson NJ, Mogridge N, et al. Posthaemorrhagic
ventricular dilatation in the premature infant: natural history and predictors of outcome. Arch
Dis Child Fetal Neonatal Ed 2002;87(1):F37 – 41.
[15] London DA, Carroll BA, Enzmann DR. Sonography of ventricular size and germinal matrix
hemorrhage in premature infants. AJR Am J Roentgenol 1980;135(3):559 – 64.
[16] Grasby DC, Esterman A, Marshall P. Ultrasound grading of cerebral ventricular dilatation in
preterm neonates. J Paediatr Child Health 2003;39(3):186 – 90.
[17] Fernell E, Hagberg B, Hagberg G, von Wendt L. Epidemiology of infantile hydrocephalus in
Sweden. III. Origin in preterm infants. Acta Paediatr Scand 1987;76(3):418 – 23.
[18] Fernell E, Hagberg B, Hagberg G, Hult G, von Wendt L. Epidemiology of infantile hydro-
cephalus in Sweden. Current aspects of the outcome in preterm infants. Neuropediatrics 1988;
19(3):143 – 5.
[19] Guzzetta F, Shackelford GD, Volpe S, Perlman JM, Volpe JJ. Periventricular intraparenchymal
echodensities in the premature newborn: critical determinant of neurologic outcome. Pediatrics
1986;78(6):995 – 1006.
[20] Krishnamoorthy KS, Kuehnle KJ, Todres ID, DeLong GR. Neurodevelopmental outcome
of survivors with posthemorrhagic hydrocephalus following grade II neonatal intraventricular
hemorrhage. Ann Neurol 1984;15(2):201 – 4.
[21] Dykes FD, Dunbar B, Lazarra A, Ahmann PA. Posthemorrhagic hydrocephalus in high-
risk preterm infants: natural history, management, and long-term outcome. J Pediatr 1989;
114(4 Pt 1):611 – 8.
[22] Pikus HJ, Levy ML, Gans W, Mendel E, McComb JG. Outcome, cost analysis, and long-term
follow-up in preterm infants with massive grade IV germinal matrix hemorrhage and progressive
hydrocephalus. Neurosurgery 1997;40(5):983 – 9.
[23] Boynton BR, Boynton CA, Merritt TA, Vaucher YE, James HE, Bejar RF. Ventriculoperitoneal
shunts in low birth weight infants with intracranial hemorrhage: neurodevelopmental outcome.
Neurosurgery 1986;18(2):141 – 5.
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 323

[24] Reinprecht A, Dietrich W, Berger A, Bavinzski G, Weninger M, Czech T. Posthemorrhagic

hydrocephalus in preterm infants: long-term follow-up and shunt-related complications. Childs
Nerv Syst 2001;17(11):663 – 9.
[25] Vohr BR, Allan WC, Westerveld M, Schneider KC, Katz KH, Makuch RW, et al. School-age
outcomes of very low birth weight infants in the indomethacin intraventricular hemorrhage
prevention trial. Pediatrics 2003;111(4 Pt 1):e340 – 6.
[26] Crowther CA, Henderson-Smart DJ. Vitamin K prior to preterm birth for preventing neonatal
periventricular haemorrhage. Cochrane Database Syst Rev 2001;(1):CD000229.
[27] Crowley PA. Antenatal corticosteroid therapy: a meta-analysis of the randomized trials, 1972 to
1994. Am J Obstet Gynecol 1995;173(1):322 – 35.
[28] Del Bigio MR. Pathophysiologic consequences of hydrocephalus. Neurosurg Clin N Am 2001;
12(4):639 – 49.
[29] Whitelaw A, Pople I, Cherian S, Evans D, Thoresen M. Phase 1 trial of prevention of hydro-
cephalus after intraventricular hemorrhage in newborn infants by drainage, irrigation, and fibri-
nolytic therapy. Pediatrics 2003;111(4 Pt 1):759 – 65.
[30] Libenson MH, Kaye EM, Rosman NP, Gilmore HE. Acetazolamide and furosemide for post-
hemorrhagic hydrocephalus of the newborn. Pediatr Neurol 1999;20(3):185 – 91.
[31] International PHVD Drug Trial Group. International randomised controlled trial of acetazol-
amide and furosemide in posthaemorrhagic ventricular dilatation in infancy. Lancet 1998;
352(9126):433 – 40.
[32] Kennedy CR, Ayers S, Campbell MJ, Elbourne D, Hope P, Johnson A. Randomized, controlled
trial of acetazolamide and furosemide in posthemorrhagic ventricular dilation in infancy: follow-
up at 1 year. Pediatrics 2001;108(3):597 – 607.
[33] Luciano R, Velardi F, Romagnoli C, Papacci P, De Stefano V, Tortorolo G. Failure of fibrino-
lytic endoventricular treatment to prevent neonatal post-haemorrhagic hydrocephalus. A case-
control trial. Childs Nerv Syst 1997;13(2):73 – 6.
[34] Hudgins RJ, Boydston WR, Hudgins PA, Morris R, Adler SM, Gilreath CL. Intrathecal uroki-
nase as a treatment for intraventricular hemorrhage in the preterm infant. Pediatr Neurosurg
1997;26(6):281 – 7.
[35] Hansen AR, Volpe JJ, Goumnerova LC, Madsen JR. Intraventricular urokinase for the treatment
of posthemorrhagic hydrocephalus. Pediatr Neurol 1997;17(3):213 – 7.
[36] Haines SJ, Lapointe M. Fibrinolytic agents in the management of posthemorrhagic hydrocepha-
lus in preterm infants: the evidence. Childs Nerv Syst 1999;15(5):226 – 34.
[37] Whitelaw A. Repeated lumbar or ventricular punctures in newborns with intraventricular hemor-
rhage. Cochrane Database Syst Rev 2001;(1):CD000216.
[38] Ventriculomegaly Trial Group. Randomised trial of early tapping in neonatal posthaemorrhagic
ventricular dilatation. Arch Dis Child 1990;65(1 Spec No):3 – 10.
[39] Lorber J, Emery JL. Intracerebral cysts complicating ventricular needling in hydrocephalic
infants: a clinico-pathological study. Dev Med Child Neurol 1964;86:125 – 39.
[40] Hanlo PW, Cinalli G, Vandertop WP, Faber JA, Bogeskov L, Borgesen SE, et al. Treatment of
hydrocephalus determined by the European Orbis Sigma Valve II survey: a multicenter prospec-
tive 5-year shunt survival study in children and adults in whom a flow-regulating shunt was used.
J Neurosurg 2003;99(1):52 – 7.
[41] Tuli S, Drake J, Lawless J, Wigg M, Lamberti-Pasculli M. Risk factors for repeated cerebrospinal
shunt failures in pediatric patients with hydrocephalus. J Neurosurg 2000;92(1):31 – 8.
[42] Weninger M, Salzer HR, Pollak A, Rosenkranz M, Vorkapic P, Korn A, et al. External ventricular
drainage for treatment of rapidly progressive posthemorrhagic hydrocephalus. Neurosurgery
1992;31(1):52 – 7.
[43] Harbaugh RE, Saunders RL, Edwards WH. External ventricular drainage for control of post-
hemorrhagic hydrocephalus in premature infants. J Neurosurg 1981;55(5):766 – 70.
[44] Rhodes TT, Edwards WH, Saunders RL, Harbaugh RE, Little CL, Morgan LJ, et al. External
ventricular drainage for initial treatment of neonatal posthemorrhagic hydrocephalus: surgical
and neurodevelopmental outcome. Pediatr Neurosci 1987;13(5):255 – 62.
324 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325

[45] Berger A, Weninger M, Reinprecht A, Haschke N, Kohlhauser C, Pollak A. Long-term experi-

ence with subcutaneously tunneled external ventricular drainage in preterm infants. Childs Nerv
Syst 2000;16(2):103 – 9.
[46] McComb JG, Ramos AD, Platzker AC, Henderson DJ, Segall HD. Management of hydrocepha-
lus secondary to intraventricular hemorrhage in the preterm infant with a subcutaneous ventricu-
lar catheter reservoir. Neurosurgery 1983;13(3):295 – 300.
[47] Hudgins RJ, Boydston WR, Gilreath CL. Treatment of posthemorrhagic hydrocephalus in the
preterm infant with a ventricular access device. Pediatr Neurosurg 1998;29(6):309 – 13.
[48] Brockmeyer DL, Wright LC, Walker ML, Ward RM. Management of posthemorrhagic hydro-
cephalus in the low-birth-weight preterm neonate. Pediatr Neurosci 1989;15(6):302 – 7.
[49] Richard E, Cinalli G, Assis D, Pierre-Kahn A, Lacaze-Masmonteil T. Treatment of post-haemor-
rhage ventricular dilatation with an Ommaya’s reservoir: management and outcome of 64
preterm infants. Childs Nerv Syst 2001;17(6):334 – 40.
[50] Fulmer BB, Grabb PA, Oakes WJ, Mapstone TB. Neonatal ventriculosubgaleal shunts. Neuro-
surgery 2000;47(1):80 – 3.
[51] Rahman S, Teo C, Morris W, Lao D, Boop FA. Ventriculosubgaleal shunt: a treatment option
for progressive posthemorrhagic hydrocephalus. Childs Nerv Syst 1995;11(11):650 – 4.
[52] Tubbs RS, Smyth MD, Wellons III JC, Blount JP, Grabb PA, Oakes WJ. Alternative uses for
the subgaleal shunt in pediatric neurosurgery. Pediatr Neurosurg 2003;39(1):22 – 4.
[53] Sklar F, Adegbite A, Shapiro K, Miller K. Ventriculosubgaleal shunts: management of post-
hemorrhagic hydrocephalus in premature infants. Pediatr Neurosurg 1992;18(5 – 6):263 – 5.
[54] Duhaime AC, Bonner K, McGowan KL, Schut L, Sutton LN, Plotkin S. Distribution of
bacteria in the operating room environment and its relation to ventricular shunt infections:
a prospective study. Childs Nerv Syst 1991;7(4):211 – 4.
[55] Kulkarni AV, Drake JM, Lamberti-Pasculli M. Cerebrospinal fluid shunt infection: a prospec-
tive study of risk factors. J Neurosurg 2001;94(2):195 – 201.
[56] Shapiro S, Boaz J, Kleiman M, Kalsbeck J, Mealey J. Origin of organisms infecting ventricular
shunts. Neurosurgery 1988;22(5):868 – 72.
[57] Rekate HL. Shunt revision: complications and their prevention. Pediatr Neurosurg 1991 – 92;
17(3):155 – 62.
[58] Kestle J, Drake J, Milner R, Sainte-Rose C, Cinalli G, Boop F, et al. Long-term follow-up data
from the shunt design trial. Pediatr Neurosurg 2000;33(5):230 – 6.
[59] Di Rocco C, Marchese E, Velardi F. A survey of the first complication of newly implanted
CSF shunt devices for the treatment of nontumoral hydrocephalus. Cooperative survey of
the 1991 – 1992 Education Committee on the ISPN. Childs Nerv Syst 1994;10(5):321 – 7.
[60] McGirt MJ, Leveque JC, Wellons III JC, Villavicencio AT, Hopkins JS, Fuchs HE, et al.
Cerebrospinal fluid shunt survival and etiology of failures: a seven-year institutional experience.
Pediatr Neurosurg 2002;36(5):248 – 55.
[61] Iskandar BJ, Tubbs S, Mapstone TB, Grabb PA, Bartolucci AA, Oakes WJ. Death in shunted
hydrocephalic children in the 1990s. Pediatr Neurosurg 1998;28(4):173 – 6.
[62] Grunert P, Charalampaki P, Hopf N, Filippi R. The role of third ventriculostomy in the man-
agement of obstructive hydrocephalus. Minim Invasive Neurosurg 2003;46(1):16 – 21.
[63] Hopf NJ, Grunert P, Fries G, Resch KD, Perneczky A. Endoscopic third ventriculostomy:
outcome analysis of 100 consecutive procedures. Neurosurgery 1999;44(4):795 – 804.
[64] Cinalli G, Sainte-Rose C, Chumas P, Zerah M, Brunelle F, Lot G, et al. Failure of third
ventriculostomy in the treatment of aqueductal stenosis in children. J Neurosurg 1999;90(3):
448 – 54.
[65] Buxton N, Macarthur D, Robertson I, Punt J. Neuroendoscopic third ventriculostomy for failed
shunts. Surg Neurol 2003;60(3):201 – 3.
[66] Boschert J, Hellwig D, Krauss JK. Endoscopic third ventriculostomy for shunt dysfunction in
occlusive hydrocephalus: long-term follow up and review. J Neurosurg 2003;98(5):1032 – 9.
[67] Decq P, Le Guerinel C, Sol JC, Brugieres P, Djindjian M, Nguyen JP. Chiari I malformation:
 H.J.L. Garton, J.H. Piatt, Jr / Pediatr Clin N Am 51 (2004) 305–325 325

a rare cause of noncommunicating hydrocephalus treated by third ventriculostomy. J Neurosurg

2001;95(5):783 – 90.
[68] Kehler U, Gliemroth J. Extraventricular intracisternal obstructive hydrocephalus – a hypothesis to
explain successful 3rd ventriculostomy in communicating hydrocephalus. Pediatr Neurosurg
2003;38(2):98 – 101.
[69] Buxton N, Macarthur D, Mallucci C, Punt J, Vloeberghs M. Neuroendoscopic third ventriculo-
stomy in patients less than 1 year old. Pediatr Neurosurg 1998;29(2):73 – 6.
[70] Schroeder HW, Niendorf WR, Gaab MR. Complications of endoscopic third ventriculostomy.
J Neurosurg 2002;96(6):1032 – 40.
[71] Tuli S, Alshail E, Drake J. Third ventriculostomy versus cerebrospinal fluid shunt as a first
procedure in pediatric hydrocephalus. Pediatr Neurosurg 1999;30(1):11 – 5.
[72] Garton HJ, Kestle JR, Cochrane DD, Steinbok P. A cost-effectiveness analysis of endoscopic
third ventriculostomy. Neurosurgery 2002;51(1):69 – 77.
[73] Barlow P, Ching HS. An economic argument in favour of endoscopic third ventriculostomy as
a treatment for obstructive hydrocephalus. Minim Invasive Neurosurg 1997;40(2):37 – 9.
[74] Faillace WJ, Canady AI. Cerebrospinal fluid shunt malfunction signaled by new or recurrent
seizures. Childs Nerv Syst 1990;6(1):37 – 40.
[75] Piatt JH, Hoffman HJ. Cor pulmonale: a lethal complication of ventriculoatrial CSF diversion.
Childs Nerv Syst 1989;5(1):29 – 31.
[76] Piatt JH. Physical examination of patients with cerebrospinal fluid shunts: is there useful infor-
mation in pumping the shunt? Pediatrics 1992;89(3):470 – 3.
[77] Piatt JH. Pumping the shunt revisited. A longitudinal study. Pediatr Neurosurg 1996;25(2):73 – 6.
[78] Watkins L, Hayward R, Andar U, Harkness W. The diagnosis of blocked cerebrospinal fluid
shunts: a prospective study of referral to a paediatric neurosurgical unit. Childs Nerv Syst 1994;
10(2):87 – 90.
[79] Garton HJ, Kestle JR, Drake JM. Predicting shunt failure on the basis of clinical symptoms and
signs in children. J Neurosurg 2001;94(2):202 – 10.
[80] Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical epidemiology: a basic science for
clinical medicine. 2nd edition. Boston: Little, Brown and Company; 1991.
[81] Barnes NP, Jones SJ, Hayward RD, Harkness WJ, Thompson D. Ventriculoperitoneal shunt
block: what are the best predictive clinical indicators? Arch Dis Child 2002;87(3):198 – 201.
[82] Iskandar BJ, McLaughlin C, Mapstone TB, Grabb PA, Oakes WJ. Pitfalls in the diagnosis of
ventricular shunt dysfunction: radiology reports and ventricular size. Pediatrics 1998;101(6):
1031 – 6.
[83] Medical Home Initiatives for Children with Special Needs Project Advisory Committee, Amer-
ican Academy of Pediatrics. The medical home. Pediatrics 2002;110(1 Pt 1):184 – 6.