Intracranial pressure monitoring and lumbar puncture after endoscopic third ventriculostomy in children.
ABSTRACT The aim of this study is to analyze changes in intracranial pressure (ICP) after endoscopic third ventriculostomy (ETV) performed in children affected by noncommunicating hydrocephalus.
ICP was continuously recorded for an average of 7 days in 64 children who underwent 68 ETVs for obstructive triventricular hydrocephalus of various etiology. In the first group (44 children), ETV was performed as the primary treatment; in the second group (20 children), the patients presented with shunt malfunction and underwent ETV and shunt removal. Three of the patients in the second group were reoperated for obstruction of the stoma: two were reoperated once and one was reoperated twice.
ICP changes after ETV were not homogeneous and varied according to etiology: the highest values were observed in patients affected by posterior fossa tumors and the lowest values were seen in patients operated on during shunt malfunction and who had their shunt removed. After 31 procedures (45.6%), ICP remained normal (< 20 mmHg) for the entire duration of the monitoring. After 37 procedures (54.5%), ICP was persistently high on Day 1 (mean, 29.7) and decreased very slowly in the subsequent days, remaining high for 2-9 days (mean, 4.5). After 20 of the 37 procedures with high postoperative ICP, patients presented symptoms of intracranial hypertension that resolved, in most of the cases, with one or two lumbar punctures. Lumbar puncture was noted to be effective in bringing about fast normalization of the ICP and resolution of the symptoms. In 13 patients (19.1%), ETV failed and a ventriculoperitoneal shunt was implanted. After four procedures, the stoma obstructed and the patients were treated, reopening the stoma. Postoperative ICP was not statistically significant higher in the patients in whom ETV failed.
The high ICP observed in a group of patients in the early postoperative days is probably related to the slow permeation of the subarachnoid spaces by the cerebrospinal fluid flowing out of the third ventriculostomy. Management of intracranial hypertension after ETV remains a matter of controversy. The role of the lumbar puncture in the faster normalization of the ICP is examined in this article. By increasing the compliance and the buffering capacities of the spinal subarachnoid spaces, it probably decreases the cerebrospinal fluid outflow resistance from the ventricular system, facilitating the decrease of the ventricular volume and allowing faster permeation of the intracranial subarachnoid spaces. High postoperative ICP can account for persistent symptoms of intracranial hypertension and ventricular dilatation on computed tomographic scans after third ventriculostomy. A cycle of one to three lumbar punctures should always be performed in patients who remain symptomatic and who show increasing ventricular dilatation after ETV, before ETV is assumed to have failed and an extracranial cerebrospinal fluid shunt is implanted.
- SourceAvailable from: library.tasmc.org.il[Show abstract] [Hide abstract]
ABSTRACT: This study evaluates the safety, efficacy, and indications for continuous lumbar drainage (CLD) in patients following endoscopic third ventriculostomy (ETV). We retrospectively reviewed the clinical data of 22 consecutive patients treated between 1996 and 2010 with CLD after ETV. The decision to insert a CLD was made in selected patients only. CLD was inserted in cases of high measured intracranial pressure (12 patients), clinical symptoms indicative of continuing hydrocephalus (2 patients), and "prophylactically" in 8 patients, based either on the clinical condition of patients before ETV or on technical difficulties during the ETV procedure, which seemed to increase the risk of ETV failure. CLD insertion took place either in the operating room immediately following the ETV procedure or under very specific conditions and with close patient monitoring in an ICU setting. Only four patients eventually required shunting, all within 1 month after ETV. Therefore, the overall ETV success rate was 81.8% (18/22 patients). Of the 14 patients suffering from measured or clinically observed continuing hydrocephalus, 12 (85%) ultimately recovered without the need for a permanent shunt. Without the CLD, some of these patients would probably have been declared "failures" and referred for a standard shunt. CLD provided a time window following ETV for the absorption system to recover and return to full functionality. Selective usage of CLD is a reasonable and safe method to gain time and possibly facilitates the recovery of absorption capacity following ETV. CLD should be considered before conceding a post-ETV patient as a failure.Child s Nervous System 08/2011; 27(11):1973-8. · 1.24 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: The aim of this study was to assess the mid-term results, success rates, and time-to-failure of secondary endoscopic third ventriculostomy (secondary ETV), as well as the complex management of preoperative and postoperative cares. To this purpose, a retrospective analysis of a pediatric population of 22 children who underwent endoscopic third ventriculostomy (ETV) after shunt malfunction (secondary ETV) was performed. The failure rate, given by the percentage of new shunt replacement in the first 3 months after ETV, was 36%, with a mean time to failure of 14.3 days. All the failures were evident within 1 month after the ETV. Despite the small number of patients in our series, we found no significant correlation between ETV failure and both patient age and hydrocephalus etiology (p = 0.47 and p = 0.78, respectively). In our experience, ETV secondary to shunt malfunction in pediatric patients has a success rate of 64%. As it is a safe and rapid treatment option even in emergency conditions, it is worth performing this procedure in previously shunted children.Child s Nervous System 03/2010; 26(7):937-43. · 1.24 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Brain stem glioma accounts for 6-9% of brain tumors in children. Tumor progression may lead to CSF pathway obstruction and development of hydrocephalus. We retrospectively reviewed charts of patients consecutively treated in our institution with diffuse intrinsic pontine glioma in order to assess incidence of hydrocephalus, its management, and its impact on overall survival. All patients had brain stem glioma not amenable to surgery. Cases with exophytic brain stem glioma were excluded. Fifty-one children were treated from January 2005 to December 2010 for brain stem glioma in the Pediatric Neurosurgery Department of Necker Enfants Malades, Paris, France. Hydrocephalus occurred in 11 of them (22%). They were six boys and five girls; the average and median time from tumor diagnosis to onset of hydrocephalus were 5.3 and 3.2 months, respectively, while average and median time from onset of hydrocephalus to death were 5.3 and 2.8 months, respectively. Hydrocephalus was treated in nine patients by a ventriculoperitoneal (VP) shunt and in two patients by an endoscopic third ventriculostomy. Because of early failure, a VP shunt was implanted in one child. The overall 1-year survival rate was 33%. Survival rate of patients with such obstructive hydrocephalus was not significantly different from patients harboring brain stem glioma who did not develop hydrocephalus. Furthermore, hydrocephalus was not related to terminal tumor progression. Considering both risks and benefit of treatment, VP shunt could be proposed, on the base of our experience, as the first option in spite of the apparently obstructive nature of the hydrocephalus associated to a brain stem tumor.Child s Nervous System 10/2011; 27(10):1735-9. · 1.24 Impact Factor
INTRACRANIAL PRESSURE MONITORING AND LUMBAR
PUNCTURE AFTER ENDOSCOPIC THIRD
VENTRICULOSTOMY IN CHILDREN
Giuseppe Cinalli, M.D.
Department of Pediatric
Santobono Children’s Hospital,
Pietro Spennato, M.D.
Department of Pediatric
Santobono Children’s Hospital,
Claudio Ruggiero, M.D.
Department of Pediatric
Santobono Children’s Hospital,
Ferdinando Aliberti, M.D.
Department of Pediatric
Santobono Children’s Hospital,
Michel Zerah, M.D.
Department of Pediatric
Ho ˆpital Necker-Enfants Malades,
Vincenzo Trischitta, M.D.
Department of Pediatric
Emilio Cianciulli, M.D.
Department of Pediatric
Giuseppe Maggi, M.D.
Department of Pediatric
Giuseppe Cinalli, M.D.,
Via Gennaro Serra N. 75,
80132 Naples, Italy.
Received, January 19, 2005.
Accepted, August 5, 2005.
OBJECTIVE: The aim of this study is to analyze changes in intracranial pressure (ICP)
after endoscopic third ventriculostomy (ETV) performed in children affected by non-
METHODS: ICP was continuously recorded for an average of 7 days in 64 children who
underwent 68 ETVs for obstructive triventricular hydrocephalus of various etiology. In the
first group (44 children), ETV was performed as the primary treatment; in the second group
(20 children), the patients presented with shunt malfunction and underwent ETV and shunt
removal. Three of the patients in the second group were reoperated for obstruction of the
stoma: two were reoperated once and one was reoperated twice.
RESULTS: ICP changes after ETV were not homogeneous and varied according to
etiology: the highest values were observed in patients affected by posterior fossa
tumors and the lowest values were seen in patients operated on during shunt mal-
function and who had their shunt removed. After 31 procedures (45.6%), ICP re-
mained normal (?20 mmHg) for the entire duration of the monitoring. After 37
procedures (54.5%), ICP was persistently high on Day 1 (mean, 29.7) and decreased
very slowly in the subsequent days, remaining high for 2–9 days (mean, 4.5). After 20
of the 37 procedures with high postoperative ICP, patients presented symptoms of
intracranial hypertension that resolved, in most of the cases, with one or two lumbar
punctures. Lumbar puncture was noted to be effective in bringing about fast normal-
ization of the ICP and resolution of the symptoms. In 13 patients (19.1%), ETV failed
and a ventriculoperitoneal shunt was implanted. After four procedures, the stoma
obstructed and the patients were treated, reopening the stoma. Postoperative ICP was
not statistically significant higher in the patients in whom ETV failed.
CONCLUSION: The high ICP observed in a group of patients in the early postoperative
days is probably related to the slow permeation of the subarachnoid spaces by the
cerebrospinal fluid flowing out of the third ventriculostomy. Management of intracra-
nial hypertension after ETV remains a matter of controversy. The role of the lumbar
puncture in the faster normalization of the ICP is examined in this article. By increasing
the compliance and the buffering capacities of the spinal subarachnoid spaces, it
probably decreases the cerebrospinal fluid outflow resistance from the ventricular
system, facilitating the decrease of the ventricular volume and allowing faster perme-
ation of the intracranial subarachnoid spaces. High postoperative ICP can account for
persistent symptoms of intracranial hypertension and ventricular dilatation on com-
puted tomographic scans after third ventriculostomy. A cycle of one to three lumbar
punctures should always be performed in patients who remain symptomatic and who
show increasing ventricular dilatation after ETV, before ETV is assumed to have failed
and an extracranial cerebrospinal fluid shunt is implanted.
KEY WORDS: Endoscopic third ventriculostomy, Hydrocephalus, Intracranial pressure, Lumbar puncture,
Neuroendoscopy, Subarachnoid space
Neurosurgery 58:126-136, 2006
126 | VOLUME 58 | NUMBER 1 | JANUARY 2006
in nature. Unfortunately, the success rate is variable and
ranges from 22.2 to 100% (5, 6, 7, 12, 18, 19). If ETV fails, a
further procedure of cerebrospinal fluid (CSF) diversion is
required to treat the hydrocephalus, usually the insertion of a
ventriculoperitoneal (VP) shunt. The best method to manage
the postoperative period and to detect the failures are still
under debate. After a successful ETV, the symptoms of in-
creased intracranial pressure (ICP) usually resolve. However,
in some cases, symptoms can persist or recur in the early
postoperative period. This phenomenon has been well known
since the time of Matson (24), who described the postoperative
period of a ventriculocervical subarachnoid shunt implanted
to bypass aqueductal stenosis. He reported that “occasionally,
the shunt does not seem to be working for several days and
then suddenly, or usually more gradually over the next few
days, free communication between the ventricle and the lum-
bar subarachnoid space becomes established” (24). Thus, dur-
ing this “adaptation period,” a misleading diagnosis of failure
of the procedure can be made and a VP shunt implanted (1,
16). Also, a radiographic follow-up examination is a challenge;
the ventricles often remain enlarged despite clinical improve-
ment (9, 14, 15, 17, 27, 35) and a computed tomographic (CT)
scan performed in the early postoperative days may only
reveal minor modifications. Cine phase-contrast magnetic res-
onance imaging (MRI), useful to detect the flow void signal
through the floor of the third ventricle, is not available in all
institutions for daily control and does not have high sensibility
and specificity in predicting outcome (5, 6, 13).
To better understand the modification of CSF hydrodynam-
ics after ETV and to adequately manage the early postopera-
tive period, we have adopted a policy of continuous monitor-
ing of ICP in children who underwent ETV at our institution.
In the present study, we analyze changes in ICP after ETV is
performed as the primary treatment or as an alternative to
shunt revision in children affected by noncommunicating hy-
drocephalus. In addition, the role of lumbar punctures as a
control for increases in ICP is examined.
uring the past decade, endoscopic third ventriculos-
tomy (ETV) has became the treatment of choice for
any form of hydrocephalus that is purely obstructive
PATIENTS AND METHODS
Between September 1999 and December 2004, 126 ETVs
were performed in the Department of Neurosurgery at the
Santobono Children’s Hospital in Naples to treat obstructive
hydrocephalus. ICP was continuously monitored after 68 pro-
cedures in 64 children. Two patients were operated on and
recorded in the statistics twice, and one was operated on and
recorded three times for obstructions of the stoma. The pa-
tients’ ages ranged from 3 months to 18 years (mean, 78.1 mo;
median, 78 mo). Three patients were younger than 1 year old,
but had a closed fontanel at the time of the operation and ICP
monitoring. ETV was performed as a primary treatment in 44
children (Group 1). All the patients were affected by triven-
tricular hydrocephalus. The triventricular hydrocephalus was
secondary to posterior fossa tumors in 19 children, to aque-
ductal stenosis in 11, to periaqueductal tumors in 12, and to
other conditions in two.
In a second group (20 children, 24 procedures) the patients
presented with shunt malfunction and underwent third ven-
triculostomy and shunt removal as an alternative to shunt
revision. This group included seven patients with aqueductal
stenosis, four with hydrocephalus secondary to tectal tumors,
three with posthemorrhagic hydrocephalus, three with myelo-
meningocele, and three patients with other conditions. Pa-
tients presenting with shunt malfunction were considered
candidates for ETV if, at the time of shunt malfunction, they
presented with a triventricular dilatation on CT scan and/or
clear MRI evidence of aqueductal obstruction on 2-mm thick
sagittal T2-weighted MRI drive sequences and/or MRI cine
phase-contrast evidence of lack of CSF flow through the aq-
ueduct and absence of any anatomical factor contraindicating
the procedure (i.e., associated malformations, slit-like ventri-
All endoscopic procedures were performed under general
anaesthesia. The surgical technique has been extensively de-
scribed elsewhere (6). Until September 2000, a steerable fiber-
scope was used (Codman & Shurtleff, Raynham, MA). Begin-
ning inOctober 2000,a
Neurologic Technologies, Goleta, CA) was used for the pro-
cedures. During the procedure, care was taken to limit the loss
of CSF and to use irrigation only when it was truly indispens-
able, in an effort to reduce significant modifications in ICP
immediately after the procedure. In patients harboring a pos-
terior fossa tumor, ETV was performed before tumor removal
in all cases except one. At the end of the operation in Group 1
patients (primary presentation), an ICP transducer (Codman
ICP transducer; Codman & Shurtleff, Raynham, MA) was left
in the lateral ventricle and was entered with the endoscope
through the endoscopic track. In patients of the second group
(shunt malfunction), the transducer was part of an EVD sys-
tem (Codman Microsensor Ventricular Catheter; Codman &
Shurtleff, Raynham, MA). In patients presenting with shunt
malfunction and who had their shunt removed (Group 2), the
safety ventricular catheter that was part of the transducer-
EVD system was connected to an external drainage bag that
was left closed in all patients for the duration of ICP monitor-
ing. In the ward, the transducer was connected to the monitor
(ICP Express-Codman; Codman & Shurtleff, Raynham, MA).
ICP values displayed on the monitor were recorded every 15
minutes; the 96 ICP values obtained each day were averaged
to obtain a daily average value (DAV). This was calculated for
each day after the procedure (the day of the operation was
considered Day 1). Throughout the day, the ICP oscillations
were wide, with extremely high (up to 80 mmHg) or ex-
tremely low values (especially after lumbar punctures), and
depended very much on crying, vomiting, and other external
factors in children. To obtain more reliable values from the
oscillation range, we decided to average the five highest val-
ues and the five lowest values during a given day. Therefore,
for each day, we calculated a DAV (average of the 96 record-
rigid fiberscope (Medtronic
INTRACRANIAL PRESSURE MONITORING AFTER ENDOSCOPIC THIRD VENTRICULOSTOMY
VOLUME 58 | NUMBER 1 | JANUARY 2006 | 127
ings), a High Daily Average Value (HDAV; average of the five
highest daily values) and a Low Daily Average Value (LDAV;
average of the five lowest daily values). CT scans were ob-
tained in the first 24 to 48 hours to assess the degree of the
ventricular dilatation immediately after the procedure. CT
scanning was performed again before discharge and in the
event of raised ICP or the reappearance of symptoms. In
patients in which the ICP was high and became symptomatic
(headache, vomiting, lethargy) we preferred to perform lum-
bar tapping under ICP monitoring; under local anesthesia
obtained with Mepivacaine infiltration or external anesthetic
pomade (EMLA, Astra Zeneca, Italy), the subarachnoid space
was entered with a 20-gauge, 40- or 90-mm spinal needle with
the patient in the lateral position. After entering the spinal sac,
the stylet was carefully removed under ICP monitoring and
the lumbar CSF opening pressure was measured with a
Claude manometer. If the lumbar CSF pressure obtained was
equal to the ICP values displayed on the monitor, small
amounts of CSF were carefully removed through the spinal
needle and sent to the laboratory for routine investigations
(cytology, chemistry, bacteriology). CSF was removed slowly
until the ICP values displayed on the monitor were within
normal range (8–10 mmHg). At the end of the procedure, CSF
pressure was remeasured with the Claude manometer before
the spinal needle was removed to check equivalence with the
intraventricular pressure. If symptoms of high ICP persisted
or recurred in the following days and/or if a follow-up CT
scan showed increasing ventricular dilatation compared with
the first postoperative CT scan, lumbar puncture was repeated
up to three times.
This protocol was not used for patients harboring an intra-
cranial neoplasm (i.e., posterior fossa) that was not yet re-
moved. In these cases, symptoms were managed with medical
treatment. The ICP transducer was removed at the bedside
after normalization of the ICP values or, in the case of imme-
diate failure, at the time of VP shunt implant.
The trends of the DAVs were classified into four patterns
(Figure 1): Progressive decrease when, after the third day of
recording, the DAV constantly decreased (Fig. 1A); stable
when the DAV remained within a range of 3 mmHg from the
beginning to the end of the recording period (Fig. 1B); pro-
gressive increase when the DAV constantly increased after
Day 1 throughout the registration period (Fig. 1C); and sec-
ondary raise when, after an initial decrease, the DAV con-
stantly increased from Day 3 (Fig. 1D). The procedure was
classified as 1) complete success if hydrocephalus was cured
without further surgery over the follow-up period; 2) partial
success if indication was good, the procedure was effective,
but obstruction occurred and the patient had to be reoperated
for reopening of the stoma; 3) failure when a VP shunt was
implanted at any time during the follow-up period. All anal-
yses were performed using SPSS software (SPSS Inc., Munich,
Germany). A comparison of categorical data was made using
the ?2test. For all comparisons, P values less than 0.05 were
In all cases, the surgical procedure allowed the creation of a
4- to 6-mm opening in the floor of the third ventricle with
good visibility of the anatomical structures of the subarach-
noid space of the interpeduncular and prepontine cisterns. In
the earliest postoperative hours, ICP monitoring revealed low
values as a consequence of CSF loss during the procedure.
After 3 to 6 hours, ICP values began to rise. The trend of DAV,
HDAV, and LDAV for the entire group of patients (n ? 68) is
shown in Figure 2A.
FIGURE 1. Examples of patterns of ICP trends. A, progressive decrease;
B, stable; C, progressive increase; and D, secondary raise.
CINALLI ET AL.
128 | VOLUME 58 | NUMBER 1 | JANUARY 2006
The ICP changes after ETV were not homogeneous. In 31
recordings (normal ICP group, 45.6%), the DAV for the dura-
tion of the monitoring never exceeded 20 mmHg (Fig. 2B). On
the contrary, in 37 recordings (high ICP group, 54.4%), ICP
was persistently high with a DAV of 26.3 on Day 1, decreasing
very slowly in the subsequent days (24.4 mmHg on Day 2, 21.7
on Day 3, 20.4 on Day 4, 20.6 on Day 5, 18.2 on Day 6, 17.9 on
Day 7, and 15.3 on Day 8) (Fig. 2C). The increased value of ICP
remained high (?20 mmHg) for an average of 4.5 days (range,
2–9 d) in this group. Very high DAVs (?35 mmHg) were
recorded in nine patients and lasted for 1 or 2 days (usually
Day 1–Day 2).
No significant differences of distribution were observed be-
tween Group 1 (primary presentation), in which ICP recordings
showed normal ICP in 19 pa-
tients and high ICP in 25 pa-
tients, and Group 2 (shunt
malfunction), in which ICP re-
cordings showed normal ICP
in 11 patients (12 recordings)
and high ICP in nine patients
(12 recordings) (P ? 0.59). In
Group 2 (shunt malfunction),
the mean ICP was lower than
in the group of primary pre-
sentation on Day 1 (DAV, 15.6
mmHg vs 20.7 mmHg) and remained lower in the following
days. However, this difference was not statistically significant (P
ICP trend distribution between the two groups (Table 1)
showed a significantly higher incidence of “progressive de-
crease” or “stable” patterns in Group 1 patients (68%), and of
a “progressive increase” or “secondary rise” pattern in Group
2 patients (65%) (P ? 0.039).
In 48 patients, ETV was completely successful and hydro-
cephalus was cured without further surgery with a mean
follow-up period of 22.76 months (range, 1–56 mo) (Fig. 3).
Twenty-two patients were in the normal ICP group and 26
were in the high ICP group. Twenty-seven presented a pro-
gressive decrease pattern of the ICP trend, 14 presented a
secondary rise pattern, and seven presented a stable ICP pat-
tern. In 13 patients (19.1%), ETV was completely unsuccessful,
based on the persistence or early recurrence of symptoms of
high ICP, cine phase-contrast T2–weighted MRI sequences
showing CSF flow through the ETV, or T2-weighted, 2-mm
thick sagittal drive MRI sequences showing an anatomically
open stoma without underlying membranes and flow artefact
through it. In these patients, a VP shunt was inserted. They
were equally distributed in the normal ICP group (six out of
31, 19.4%) and in the high ICP group (seven out of 37, 18.9%),
with no significant differences compared with the success
group (P ? 0.984). Eight of the patients presented a “second-
ary rise” pattern of ICP trend, three presented a “progressive
increase” pattern, and two presented a “progressive decrease”
pattern (Table 2). The higher incidence of “secondary rise” and
“progressive increase” patterns in the failure group compared
with the success group was statistically significant (P ? 0.01).
Three patients (all in Group 2) presented with persistence or
early recurrence of high ICP symptoms, and cine phase-
contrast T2-weighted MRI sequences showing absence of CSF
flow through the ETV and/or T2-weighted, 2-mm thick sag-
ittal drive MRI sequences showing an anatomically closed
without flow artefact through it and dilatation of the third
ventricle with downward ballooning of the third ventricle
floor. These patients were diagnosed with secondary obstruc-
tion of the stoma and were reoperated. Two were reoperated
after 5 months and 16 days, respectively. The third was reop-
erated twice: 21 days and 8 months later (Patient 2). In these
three patients, the reoperation procedure was successful (in
TABLE 1. Intracranial pressure distributiona
ICP trend Patterns Group 1 (primary presentation)Group 2 (shunt malfunction)
aICP, intracranial pressure.
FIGURE 2. A, ICP trend after the
68 procedures (64 patients). Solid
line represents the mean of the
DAVs, the upper dotted line
HDAVs, and the lower dotted
line indicates the mean of the
LDAVs. B, trend of the DAVs
(solid line), HDAVs (upper dot-
ted line) and LDAVs (lower dot-
ted line) for the 31 recordings (29
patients) who had normal or low
ICP values after ETV. C, trend of
the DAVs (solid line), HDAVs
(upper dotted line), and LDAVs
(lower dotted line) for the 37 recordings (36 patients) who had high ICP
values after ETV.
INTRACRANIAL PRESSURE MONITORING AFTER ENDOSCOPIC THIRD VENTRICULOSTOMY
VOLUME 58 | NUMBER 1 | JANUARY 2006 | 129
one case, after two reoperations) and they are shunt-free with
an average follow-up period of 3.5 years. The ICP trend pat-
terns were homogeneous. After the four procedures that
ended up in ETV obstruction, the patterns show a secondary
rise in ICP, whereas after the three final procedures that
allowed resolution of hydrocephalus ICP trend showed a pro-
gressive decrease of ICP. In the three children younger than 1
year of age, ICP was immediately normal (4–9 mm Hg) on Day
1 and remained normal for the next 5 to 8 days.
The ICP varied little according to the four main etiological
groups. We considered the 20 shunt malfunction patients (re-
gardless of primary etiology) as a separate group under a
pathophysiological point of view, and grouped together the 12
patients affected by tumor of the posterior part of the third
ventricle (four pineal region, four tectal tumors, four thalamic
tumors). The other two main groups were posterior fossa
tumors and aqueductal stenosis. Two patients affected by
hydrocephalus induced by arachnoid cyst and Monro foramen
stenosis did not fit into these main groups.
The shunt malfunction group (20 patients, 24 procedures)
showed the lowest mean values throughout the recording
period: 12 recordings were part of the low ICP group and 12
were part of the high ICP group. The posterior fossa tumors
group (19 patients) showed the highest mean values, with a
trend to further elevation in the second half of the recording
period: nine were part of the low ICP group and 10 were part
of the high ICP group. The aqueductal stenosis group (11
patients) showed intermediate ICP values: four were part of
the low ICP group and seven were part of the high ICP group.
The periaqueductal tumor group (12 patients) showed inter-
mediate ICP values: four were part of the low ICP group and
eight were part of the high ICP group.
Patients who had high ICP and remained symptomatic or
showed a tendency toward increased ICP values or presented
increasing ventricular dilatation on CT scans received at least
one lumbar puncture to expedite the process of ICP normal-
ization and decrease the symptoms. Twelve patients received
one lumbar puncture, seven received two lumbar punctures,
and one received five lumbar punctures. In all cases, the CSF
opening pressure was equal to the ICP values displayed on the
monitor, allowing us to perform CSF subtraction in all cases.
At the end of the CSF subtraction, the CSF pressure was
measured again before removing the needle and, in all cases,
we found good equivalence between lumbar and intraventric-
ular pressure. All lumbar punctures were uneventful and well
tolerated by the patients, except sporadic cases of mild lumbar
pain for 3 to 4 days after the procedure. In most of the cases
lumbar punctures were effective in bringing about fast nor-
malization of the ICP and resolution of the symptoms (Fig. 4).
The average ICP values from the 6 hours before lumbar punc-
ture were compared to the average ICP values from the 6
hours after lumbar puncture. The details are shown in Figure
4. On average, the ICP decreased from 26.8 mmHg before
lumbar puncture to 14.5 mmHg after lumbar puncture. The
positive effects on the ICP lasted longer that can be explained
by simple subtraction of small amount of CSF (5–10 ml). The
response of ICP to lumbar puncture was not predictive of
successful outcome or failure.
TABLE 2. Intracranial pressure trend patternsa
ICP Trend PatternsSuccessful ETVFailed ETV
aICP, intracranial pressure; ETV, endoscopic third ventriculostomy.
FIGURE 3. Percent of shunt-free patients versus time. All failures
occurred within the first 6 months. Arrows indicate the patients who
were reoperated for occlusion of the stoma at 1 month (2 patients), 5
months and 8 months after ETV, respectively.
FIGURE 4. Average ICP values during the 6 hours before (?) and after
(f) lumbar puncture. Patients in whom ETV failed to control hydrocepha-
lus, requiring VP shunt insertion (*) or who presented with obstructed
stoma requiring reoperation ETV (5) are indicated.
CINALLI ET AL.
130 | VOLUME 58 | NUMBER 1 | JANUARY 2006
A 12-year-old girl was admitted with headache. An MRI scan
showed a huge dilatation of the lateral and third ventricles with
enlarged pseudocystic suprapineal recess. ETV was uneventfully per-
formed. At the end of the procedure, a Codman intraventricular
pressure transducer was positioned and the ICP was continuously
monitored (Fig. 5). After the operation, the symptoms immediately
improved. Her ICP remained very high over the first 24 hours, with
progressive and slow reduction in the next days until normalization
on Day 7. On Day 7, all the symptoms resolved, the transducer was
removed, and the child was discharged. MRI scans performed at 1
month confirmed nontumoral aqueductal stenosis and showed the
flow artefact in the floor of the third ventricle. At a 1-year follow up
examination, the child was free of symptoms.
A 2-year-old boy, shunted for hydrocephalus and aqueductal ste-
nosis at the age of 1 month, presented with repeated vomiting. CT
scans showed triventricular dilatation with transependymal resorp-
tion. ETV was uneventfully performed and the shunt was removed.
ICP monitoring showed very high values from the first postoperative
day, with a progressive increase of average values. The onset of high
ICP symptoms on Day 5, in spite of a normal CT scan (Fig. 6A)
prompted a lumbar puncture that induced immediate and persistent
normalization of ICP values. Three weeks later, the patient was read-
mitted for the same reasons. CT scans showed re-enlargement of the
supratentorial ventricular system. The stoma was obstructed by scar
tissue at surgery and a new ETV was uneventfully performed. Post-
operative ICP monitoring showed high values in the first 24 hours.
Lumbar puncture was performed because of symptoms of high ICP.
ICP values remained high, and the patient remained asymptomatic
until Day 5 when the onset of headache prompted a new lumbar
puncture that resolved the symptoms and normalized ICP values.
Seven months later, the patient was again readmitted for headache.
CT scans showed triventricular dilatation, and at surgery the stoma
was again found obstructed and a new ETV was uneventfully per-
formed. Twenty-four hours after surgery, the patient was severely
symptomatic with high ICP values (Fig. 6B). Lumbar puncture re-
solved the symptoms rapidly, the ICP decreased slowly to normal
values over 3 days and the patient was discharged. He is now shunt-
and symptom-free with a normal neurological examination and nor-
mal psychomotor development at a 40-month follow up examination.
MRI scans show normal ventricles and well-represented flow void on
An 11-year-old girl, shunted at the age of 7 months, presented with
headache, vomiting, and lethargy. CT scans of the head showed
triventricular hydrocephalus with periventricular lucency and a diag-
nosis of shunt malfunction was made (Fig. 7A). ETV was uneventfully
performed and the ICP was monitored in the postoperative days.
Symptoms immediately resolved and ICP was under 20 mmHg in the
first 2 days. CT scans showed decreased ventricular dilatation (Fig.
7B). On the third day, the ICP began to rise (Fig. 7C) and the patient
presented with headache and vomiting. A CT scan performed on Day
4 showed enlargement of the ventricles and disappeared pericerebral
sulci (Fig. 7D). A lumbar tapping performed on Day 4 immediately
resolved the symptoms, but the ICP rapidly rose again. After a second
lumbar tapping on Day 5, ICP stabilized around 20 mmHg and slowly
decreased over the next days. On Day 8, a CT scan showed the
reduction in size of the ventricles (Fig. 7E). On Day 10, the transducer
was removed and the patient discharged asymptomatic. She is still
well after 3 years of follow-up.
A 9-year-old boy presented with headache, vomiting, and ataxic
gait. CT and MRI scans revealed a fourth ventricle tumor with hydro-
cephalus. The tumor was completely removed using a suboccipital
approach. Eleven days later, a CT scan revealed triventricular hydro-
cephalus and an ETV was performed. The postoperative period was
characterized by significant headache and high ICP values (range,
15–45 mmHg; mean, 25–35 mmHg). Lumbar puncture performed on
Day 4 allowed only transient relief of the symptoms and decrease of
ICP values (Fig. 8), followed by significant re-elevation of ICP values
(range, 45–85 mmHg; mean, 60 mmHg) with severe symptoms of
FIGURE 5. Patient 1. ICP monitoring after ETV. Note the slow, progres-
sive decrease of ICP values in the first postoperative week.
FIGURE 6. Patient 2. A, ICP monitoring after the first ETV. The onset
of high ICP symptoms on Day 5 prompted a lumbar puncture (arrow)
that induced immediate and persistent normalization of ICP values. B,
ICP monitoring after the third ETV. Twenty-four hours after surgery, the
patient was severely symptomatic with high ICP values. Lumbar puncture
resolved the symptoms rapidly (arrow), the ICP decreased slowly to nor-
mal values over 3 days, and the patient was discharged.
INTRACRANIAL PRESSURE MONITORING AFTER ENDOSCOPIC THIRD VENTRICULOSTOMY
VOLUME 58 | NUMBER 1 | JANUARY 2006 | 131
increased ICP within 6 hours, prompting the emergency implant of a
VP shunt. Postoperative MRI scans revealed diffuse subarachnoid
seeding of the tumor, in both the intracranial and intraspinal spaces.
In the past few years, ETV has been established as the
procedure of choice for the treatment of obstructive hydro-
cephalus. As early as 1932, Dandy (8) cautioned that, before
performing a third ventriculostomy for obstructive hydro-
cephalus, it was necessary to establish whether the distal
subarachnoid pathways were open or closed. However, the
difficulty to preoperatively detect the presence of coexisting
obstruction in the basilar cisterns or in the subarachnoid space
of the surface may account for the 25 to 40% failure rate
reported in the literature (5, 12, 18). Further complicating this
scenario is the observation that obstruction of the ventricular
system may cause a secondary reversible obstruction in the
subarachnoid space (26). Experimental occlusion of the aque-
duct in monkeys has been found to cause compression of the
brain against the undersurface of the cranium, resulting in
progressive obliteration of the convexity subarachnoid space,
and, in a second phase, impaction of the temporal lobes
through the incisura that resulted in obstruction of the basilar
cisterns. However, no histological evidence of irreversible ob-
struction was observed (25). Milhorat et al. (26) reported two
children affected by aqueductal stenosis, in whom complete
cisternal blocks, demonstrated by isotope cisternography,
were eliminated by temporary ventricular drainage. The relief
of intraventricular pressure re-expanded the subarachnoid
space in these cases. According to the authors, the complete-
ness of the re-expansion should depend upon multiple factors,
including the duration and the completeness of the obstruc-
tion, the effectiveness of previous shunts, and the presence or
absence of associated meningeal infections or other inflamma-
These observations may account for the lack of a preopera-
tive test that is able to predict whether the subarachnoid
spaces are competent. Also, invasive and dynamic techniques,
including isotope studies (14, 15, 22, 32, 35), CSF infusion tests
(10, 20, 23), and CT ventriculography (28, 35), may fail to
adequately investigate the total absorptive capacity, because
they explore the distal subarachnoid pathways in condition of
high intraventricular pressure. This condition may change
after CSF drainage or third ventriculostomy, from which res-
olution of raised ICP and/or equilibration of intraventricular
and extraventricular CSF pressure occur (23). The CSF resorp-
tion pathways may reopen even several days after the proce-
dure (14). Isotope studies (22, 29) have indicated that absorp-
tive capacity and CSF circulation through the subarachnoid
space may continue to improve for several months after ETV.
FIGURE 7. Patient 3. A, CT scan showing shunt malfunction with
triventricular dilatation. B, CT scan obtained on the day after ETV. Ven-
tricular dilatation has decreased and safety EVD is visible in the right
frontal horn. C, ICP monitoring after ETV showing an increase in ICP.
D, CT scan performed 4 days after ETV showing re-enlargement of the
ventricular system. A lumbar tapping (arrow 1) performed on Day 4
immediately resolved the symptoms, but the ICP rapidly rose again. After
a second lumbar tapping on Day 5 (arrow 2), the ICP stabilized around
20 mmHg and slowly decreased in the next days. E, CT scan performed 8
days after ETV showing reduction of the ventricular dilatation.
FIGURE 8. Patient 4. ICP monitoring after ETV showing persistent and
significant ICP increase, resistant to lumbar puncture. Lumbar puncture
performed on Day 4 (arrow) allowed only transient relief of the symptoms
and decrease of ICP values.
CINALLI ET AL.
132 | VOLUME 58 | NUMBER 1 | JANUARY 2006
In 1986, Hirsch et al. (14), recording ICP after percutaneous
ventriculocisternostomy (performed under radioscopic con-
trol), first noted that ICP could remain high and return to
normal only progressively even after successful procedures.
They attributed this to the slowly progressive opening of the
subarachnoid spaces. Hoffman et al. (15) offered the same
interpretation to explain the phenomenon that, after third
ventriculostomy in infants, the fontanelle remained full in the
first postoperative days. They first proposed to perform one or
more lumbar punctures to stimulate CSF circulation.
Only sporadic reports on ICP recording after ETV are de-
scribed in literature (1, 4, 11, 16, 21, 22, 29, 30, 33, 34) because
the placement of an ICP transducer is not considered neces-
sary by most in the management of the postoperative period
after ETV. In these series, ICP has been found to remain
elevated immediately after ETV, and to begin to decrease
between 4 and 8 days. This has been interpreted as a conse-
quence of higher CSF outflow resistance to the increased
amount of CSF entering the subarachnoid spaces through the
ETV. The origin of this phenomenon still remains under de-
bate. Possible hypotheses are the progressive reopening of the
subarachnoid spaces of the convexity, a slow increase in per-
meability of the pacchionian granulations, and possible effu-
sion of the CSF into the spinal subdural space (3). However,
ICP elevation after the procedure is not observed in all cases.
Nishiyama et al. (29), in their series of 15 patients treated with
ETV for shunt malfunction, analyzed the amount of CSF
drained through an external drainage positioned at 30 cm high
and showed that adaptation can be roughly summarized into
two patterns. In the first (7 patients), the amount of CSF
drained is insignificant (? 20 ml/d) and rapidly decreases
within 2 days. In the second (8 patients), the amount of CSF
drained is larger (150–250 ml/d) and decreases more slowly
over several days.
We also have recorded in our series two patterns of ICP
during the “adaptation period.” In the first group (53%), ICP
was immediately normal (? 20 mmHg), and remained normal
for all the duration of the monitoring. Most of patients with
shunt malfunction (78%) and half (48%) of children with pri-
mary presentation belonged to this group. In the second group
(47%), ICP was immediately high or became high until the
second postoperative day, and decreased slowly in the subse-
quent days; elevated values of ICP had been recorded until the
ninth postoperative day.
In this second group (high ICP), some patients became
symptomatic with headaches and vomiting. In some cases, the
symptoms were so intense that a suspicion of failure was
raised. CT scans performed in this period showed increasing
dilatation of the ventricular system in some patients, adding
further evidence of failure. This phenomenon has been also
observed by other authors (1, 16) and could be responsible for
wrong diagnosis of failure, leading to the implant of unnec-
essary VP shunts in patients who do not require them. In fact,
for most of these symptomatic patients, this period was tran-
sient and complete resolution was observed in most cases
either spontaneously or after lumbar puncture.
How to adequately manage symptomatic intracranial hy-
pertension during this “adaptation period” is still under de-
bate. Some authors (1, 2, 16, 29, 30) have proposed leaving an
external ventricular drain in place during the procedure to
allow intermittent CSF drainage during the periods of patho-
logical ICP elevation. According to these authors, this would
allow transient drop in intraventricular pressure, allowing
re-expansion of the intracranial subarachnoid spaces and fa-
cilitating CSF circulation toward the convexity. After 5 years
of systematic ICP monitoring in all children older than 1 year
who underwent ETV for obstructive hydrocephalus, we have
ascertained the benign course of this “adaptation period,” its
possible association with re-enlargment of the ventricles in the
immediate (? 24 h) postoperative CT scan and the positive
response to CSF subtraction by the mean of postoperative
lumbar punctures. On the basis of this experience, we now
think postoperative ICP monitoring, although useful, can be
replaced by careful clinical observation, adequate serial post-
operative imaging (including early CT or cine phase-contrast
T2-weighted MRI sequences) in patients operated on for pri-
mary hydrocephalus. On the contrary, we continue to use ICP
monitoring and safety EVD in the first postoperative days
after ETV performed in patients with shunt malfunction who
had their shunt removed. In fact, these patients can present
severe intracranial hypertension symptoms after removal of
the VP shunt and a longer period of adaptation if compared
with primary hydrocephalic patients. Differential diagnosis
between the adaptation period and failure can be very diffi-
cult, and ICP trend interpretation has always proven to be
effective in our experience in the management of these pa-
In our experience, lumbar puncture under ICP monitoring,
has been a valid diagnostic test to make the difference be-
tween “symptomatic adaptation period” and early failure.
This maneuver allows the withdrawal of the amount of CSF
necessary to restore a normal ICP, as already suggested and as
reported by Matson (24) in 1969 and by others (15, 17, 31). In
fact, after lumbar puncture, transient ICP decrease was ob-
served in almost all cases with resolution of the high ICP
symptoms (Fig. 4). If ICP returned to pre-puncture values with
recurrence of high ICP symptoms, lumbar tapping was re-
peated 24 to 48 hours later. A third tap was rarely necessary.
In the event of persistently high ICP values and symptoms,
MRI scans were performed to verify the patency of the stoma.
If flow artefact was clearly visible on midline sagittal MRI
T2-weighted turbo spin echo or cine phase-contrast dynamic
sequences, the patient was classified as a failure and im-
planted with a VP shunt (7). In most of the cases, however, ICP
values and symptoms remained normal or normalized after
further lumbar punctures. These patients did not require fur-
ther surgery and were classified as successful. The positive
effects on the ICP last longer than can be explained by simple
subtraction of the small CSF volume, which is usually 5 to 10
ml. One possible explanation is the increased compliance of
the spinal subarachnoid spaces and the increased pressure
gradient between third ventricle and posterior fossa subarach-
INTRACRANIAL PRESSURE MONITORING AFTER ENDOSCOPIC THIRD VENTRICULOSTOMY
VOLUME 58 | NUMBER 1 | JANUARY 2006 | 133
noid space (subarachnoid space pressure ? intraventricular
pressure) after CSF subtraction, that would be prolonged in
time because of the CSF leak in the peridural space and in the
muscles through the dural hole opened by the spinal needle.
We have described continuous ICP monitoring after ETV in
children affected by obstructive hydrocephalus of various eti-
ologies. We identified patients in whom ICP normalized in the
first hours after the procedure and a group of patients in
whom ICP remained high for the first two to nine days. High
values of ICP in the first postoperative days, whether associ-
ated with symptoms and signs of intracranial hypertension
and/or increased ventricular dilatation after ETV or not, may
not be predictive of failure of the procedure but can be related
to the transient phase of increased ICP in more than half (54%)
of the cases of ETV in children. Small CSF subtraction from the
lumbar sac can help in the differential diagnosis and a cycle of
one to three lumbar taps should always be performed before
concluding that a VP shunt implant is necessary.
1. Bellotti A, Rapana ` A, Iaccarino C, Schonauer M: Intracranial pressure mon-
itoring after endoscopic third ventriculostomy: An effective method to
manage the ‘adaptation period’. Clin Neurol Neurosurg 103:223–227, 2001.
2. Bo ¨schert J, Hellwig D, Krauss JK: Endoscopic third ventriculostomy for
shunt dysfunction in occlusive hydrocephalus: Long-term follow up and
review. J Neurosurg 98:1032–1039, 2003.
3. Cartmill M, Vloeberghs M: The fate of the cerebrospinal fluid after
neuroendoscopic third ventriculostomy. Childs Nerv Syst 16:879–881, 2000.
4. Cinalli G, Ruggiero C, Aliberti F, Maggi G: ICP monitoring following endoscopic
third ventriculostomy in children. Childs Nerv Syst 18:548, 2002.
5. Cinalli G , Sainte-Rose C, Chumas P, Zerah M, Brunelle F, Lot G, Pierre-
Kahn A, Renier D: Failure of third ventriculostomy in the treatment of
aqueductal stenosis in children. J Neurosurg 90:448–454, 1999.
6. Cinalli G: Endoscopic third ventriculostomy, in Cinalli G, Maixner W,
Sainte-Rose C (eds): Pediatric Hydrocephalus. Milan, Springer Verlag, 2004.
7. Cinalli G, Spennato P, Cianciulli E: Analysis of the outcome of endoscopic
third ventriculostomy. Adv Tech Stand Neurosurg (in press).
8. Dandy WE: The brain, in Lewis D (ed): Practice of Surgery. Hagerston, W.F.
Prior Co., 1932.
9. Drake JM: Ventriculostomy for treatment of hydrocephalus. Neurosurg Clin
N Am 4:657–666, 1993.
10. Ekstedt J: CSF hydrodynamic studies in man: 1. Method of constant pressure
CSF infusion. J Neurol Neurosurg Psychiatry 40:105–119, 1977.
11. Frim DM, Goumnerova L: Telemetric intraventricular pressure measure-
ment after third ventriculostomy in a patient with noncommunicating hy-
drocephalus. Neurosurgery 41:1425–1428, 1997.
12. Fritsch MJ, Mehdorn M: Endoscopic intraventricular surgery for treatment
of hydrocephalus and loculated CSF space in children less than one year of
age. Pediatr Neurosurg 36:183–188, 2002.
outcome and CSF flow patterns. Pediatr Neurosurg 27:149–152, 1997.
14. Hirsch JF, Hirsch E, Sainte-Rose C, Renier D, Pierre-Kahn A: Stenosis of the
aqueduct of Sylvius. Etiology and treatment. J Neurosurg Sci 30:29–36, 1986.
15. Hoffman HJ, Harwood-Nash D, Gilday DL, Craven MA: Percutaneous third
ventriculostomy in the management of noncommunicating hydrocephalus.
Neurosurgery 7:313–321, 1980.
16. Hopf NJ, Grunert P, Fries G, Resch KD, Perneczky A: Endoscopic third
ventriculostomy: Outcome analysis of 100 consecutive procedures. Neuro-
surgery 44:795–804, 1999.
17. Jaksche H, Loew F: Burr hole third ventriculostomy: An unpopular but
effective procedure for treatment of certain forms of occlusive hydroceph-
alus. Acta Neurochir 79:48–51, 1986.
18. Javadpour M , Mallucci C, Brodbelt A, Golash A, May P: The impact of
endoscopic third ventriculostomy on the management of newly diagnosed
hydrocephalus in infants. Pediatr Neurosurg 35:131–135, 2001.
19. Jones RF, Stening WA, Brydon M: Endoscopic third ventriculostomy.
Neurosurgery 26:86–91, 1990.
20. Katzman R, Hussey F: A simple constant infusion manometric test for measure-
ment of CSF absorption: I. Rationale and method. Neurology 20:534–544, 1970.
21. Kehler U, Gliemroth J, Knopp U, Arnold H: The role of third ventriculostomy in
previously shunted hydrocephalus, in Hellwig D, Bauer B (eds): Minimally invasive
techniques for neurosurgery. Berlin, Springer Verlag, 1998, pp 77–80.
22. Kelly PJ: Stereotactic third ventriculostomy in patients with nontumoral
adolescent/adult onset aqueductal stenosis and symptomatic hydrocepha-
lus. J Neurosurg 75:865–873, 1991.
23. Magnaes B: Cerebrospinal fluid hydrodynamics in adult patients with be-
nign non-communicating hydrocephalus: One-hour test shunting and bal-
anced cerebrospinal fluid infusion test to select patients for intracranial
bypass operation. Neurosurgery 11:769–775, 1982.
24. Matson DD: Neurosurgery of infancy and childhood. Springfield, Charles C
Thomas, 1969, ed 2, p 240.
25. Milhorat TH, Clark RG, Hammock MK: Experimental hydrocephalus. Part
2: Gross pathological findings in acute and subacute obstructive hydroceph-
alus in the dogs and monkeys. J Neurosurg 32:390–399, 1970.
26. Milhorat TH, Hammock MK, Di Chiro G: The subarachnoid space in con-
genital obstructive hydrocephalus. Part 1: Cisternographic findings.
J Neurosurg 35:1–6, 1971.
27. Musolino A, Soria V, Munari C, Devaux B, Merienne L, Constans JP, Chodkiewicz
triventricular hydrocephalus. Apropos of 23 cases. Neurochirurgie 34:361–373,
28. Nelson JR, Goodman SJ: An evaluation of the cerebro-spinal fluid infusion
test for hydrocephalus. Neurology 21:1037–1053, 1971.
29. Nishiyama K, Mori H , Tanaka R: Changes in cerebrospinal fluid hydrody-
namics following endoscopic third ventriculostomy for shunt-dependent
noncommunicating hydrocephalus. J Neurosurg 98:1027–1031, 2003.
30. Oi S, Shibata M, Tominaga J, Honda Y, Shinoda M, Takei F, Tsugane R,
Matsuzawa K, Sato O: Efficacy of neuroendoscopic procedures in minimally
invasive preferential management of pineal region tumors: A prospective
study. J Neurosurg 93:245–253, 2000.
31. Oka K, Yamamoto M, Ikeda K, Tomonaga M: Flexible endoneurosurgical
therapy for aqueductal stenosis. Neurosurgery 33:236–243, 1993.
32. Pierre-Kahn A, Renier D, Bombois B, Askienay S, Moreau R, Hirsch J: Role
of the ventriculocisternostomy in the treatment of non-communicating hy-
drocephalus. Neurochirurgie 21:557–569, 1975.
33. Sainte-Rose C: Third ventriculostomy, in Manwaring KH, Crone KR (eds):
Neuroendoscopy. New York, Mary Ann Liebert, 1992, pp 47–62.
34. Sainte-Rose C, Chumas P: Endoscopic third ventriculostomy. Techniques in
Neurosurgery 1:176–184, 1995.
35. Schwartz TH, Ho B, Prestigiacomo CJ, Bruce JN, Feldstein NA, Goodman
RR: Ventricular volume following endoscopic third ventriculostomy.
J Neurosurg 91:20–25, 1999.
dren’s hospitals. The success rates described are among the highest
published for ETV. A uniform philosophy underlies the study: A high
value is given to life without a shunt and shunt avoidance justifies
prolonged stays in intensive care, multiple surgical procedures, and
invasive procedures. These decisions are individual. Most pediatric neu-
rosurgeons are not as patient or as committed to shunt independence to
accept intracranial pressure (ICP) in the range of 30 mm Hg or higher or
to accept one to two weeks of adaptation to avoid shunting.
hese authors report a very large series of patients undergoing endo-
scopic third ventriculostomy (ETV) at two very large and busy chil-
CINALLI ET AL.
134 | VOLUME 58 | NUMBER 1 | JANUARY 2006
The most important information presented relates to the use of
lumbar puncture to encourage cerebrospinal fluid (CSF) flow through
the ostium of the ETV and toward natural absorption. Opening the
external drain could have resulted in a continuing closure of the
ostium so that lumbar puncture could lower ICP transiently during
the period of adaptation. I have not yet used this technique but will
definitely attempt this form of management in patients who are at
high risk for problems with shunts. This is especially important for
patients who have had abdominal complications and would therefore
need ventricular shunting to the pleura or atrium.
The authors evaluate the patency of the ostium of the ETV using
cine MRI flow studies. In patients with EVDs, I have found it helpful
to inject iohexol into the ventricle and then to obtain a CT scan to
determine patency. Doing so also gives some understanding of the
patency and anatomy of the basal cisterns and cortical subarachnoid
spaces. Personally, I would perform such a study before performing
lumbar puncture on patients with large ventricles, a history of previ-
ous obstruction, and signs of intracranial hypertension.
There are three important and controversial aspects to the perfor-
mance of ETV: indications for performance, peri-operative management,
and definition of failure. This article has a great deal to say about all three
of these issues. Basically, the authors seem to be advocating the use of
ETV for all patients with triventricular hydrocephalus who have closed
fontanels, regardless of the underlying etiology of the hydrocephalus.
They clearly are trying to offer the procedure to the maximum number of
patients. As such, this work provides evidence that their low threshold
for patient selection is reasonable.
Regarding peri-operative management, all patients were managed
with EVD and an ICP monitoring transducer. This is the safest man-
agement scheme in the post-operative period because of the ready and
reliable drainage of ventricular CSF even though the drain was rarely
or never opened. Increasing numbers of reports of late failure of EVD
and of sudden death at the time of failure have led some authors to
recommend the routine placement of implanted reservoirs after EVD.
By doing so, ICP can be easily measured in confusing cases and CSF
can be withdrawn in patients with high ICP and deteriorating neuro-
logic condition (1, 2, 3). Other authors have found no need to leave
any device behind when ETV has been successful.
At the moment all three of these strategies are options and have
pros and cons. In my opinion in patients with severe hydrocephalus
treated initially with ETV, it is reasonable to count on clinical judg-
ment. I usually leave an EVD in place for a day or two, but I do not
believe it is altogether necessary. In patients undergoing ETV at the
time of shunt failure or by programmed shunt removal, I recommend
leaving a reservoir behind to assess the patient who inevitably returns
with headaches. In the immediate post-operative period, a needle can
be affixed into the reservoir transcutaneously to monitor ICP.
Finally, regarding the most important question, what is the defini-
tion of success or failure of the treatment of hydrocephalus? Is the
absence of incapacitating intracranial hypertension in the presence of
continuing ventriculomegaly an acceptable definition of successful
treatment of hydrocephalus? Is the patient free of continuing loss of
neurologic and cognitive function? These questions have few or no
answers. Only carefully performed long-term late outcome studies
involving neuropsychological assessments can answer this question.
Harold L. Rekate
1. Aquilina K, Edwards RJ, Pople IK: Routine placement of a ventricular reser-
voir at endoscopic third ventriculostomy. Neurosurgery 53:91–96, 2003.
2. Hader WJ, Drake J, Cochrane D, Sparrow O, Johnson ES, Kestle J: Death after
late failure of third ventriculostomy in children: Report of three cases.
J Neurosurg 97:211–215, 2002.
3. Tuli S, Alshail E, Drake J: Third ventriculostomy versus cerebrospinal fluid
shunt as a first procedure in pediatric hydrocephalus. Pediatr Neurosurg
cephalus who underwent 68 endoscopic third ventriculostomy proce-
dures. In children with posterior fossa tumors, ETV was performed
before removal in all cases except one. They have noted in the past
that hydrocephalus following tumor removal is most likely related to
adhesions at the level of the outlet of the 4th ventricle (obstructive)
and can be successfully treated with ETV in most cases (1).
In a first group (46 children), ETV was performed as primary treatment
and a simple transducer was implanted. In a second group (22 cases), the
patients presented with shunt malfunction and underwent ETV and shunt
removal and a transducer in addition to an EVD was implanted. Another
three patients were re-operated for obstruction of the stoma: two were
re-operated once and 1 twice. The highest ICP values were observed in
patients with posterior fossa tumors and the lowest values were in patients
presenting with shunt malfunction in which the shunt was removed. In 31
ICP was persistently high on day one (mean 29.7) and decreased in the
these 37 patients had resolution of their high post-operative ICP with serial
lumbar punctures. Post-operative ICP was not statistically significant higher
in the patients in which ETV failed.
ETV in an unselected group of consecutive pediatric patients and not to
the care of more than half (54%) of the cases of ETV in children due to
transient phases of increased ICP. What is interesting is the description of a
potentially misleading period of transient increases in ICP following ETV.
They describe a simple approach of serial lumbar punctures to differentiate
between failure of the ETV and a transient “adaptation period” which they
suggest is not indicative of the need for shunting.
Proof of stoma patency was verified by correspondence of opening
pressure measurements during serial LP’s and monitored ICP. They
reported 16 ETV failures overall, 13 of which required shunting. No
significant correlation of higher incidence of subsequent ETV failure in
patients requiring LP’s was found. In these cases a positive diagnosis of
stoma obstruction was confirmed using cine PC MRI in dynamic se-
quences and/or on anatomical 2 mm thickness T2 DRIVE sequences. If
the stoma was noted to be patent, the patient was shunted.
They conclude that post-operative elevations of ICP can account for
persistent symptoms of intracranial hypertension and ventricular di-
latation on CT scan following third ventriculostomy and that a cycle
of one to three lumbar punctures should always be performed in
patients who remain symptomatic and show increasing ventricular
dilatation following ETV, before ETV is assumed to have failed and an
extracranial CSF shunt is implanted.
he authors evaluated changes in ICP following endoscopic third
ventriculostomy in 64 children with non-communicating hydro-
Michael L. Levy
San Diego, California
1. Sainte-Rose C, Cinalli G, Roux FE, et al: Management of hydrocephalus in
pediatric patients with posterior fossa tumors: the role of endoscopic third
ventriculostomy. J Neurosurg 95:791–797, 2001.
INTRACRANIAL PRESSURE MONITORING AFTER ENDOSCOPIC THIRD VENTRICULOSTOMY
VOLUME 58 | NUMBER 1 | JANUARY 2006 | 135
underwent continuous ICP monitoring for several days thereafter. In
roughly one-half of the patients, the ICP remained high for several
days following the procedure, but then eventually resolved in most
cases, sometimes following one or more spinal taps. In 20%, the ETV
eventually failed, necessitating placement of a shunt. The immediate
ICP values did not correlate with eventual failure. It is concluded that
there may be protracted intracranial hypertension following ETV
which does not necessarily represent failure, and may reflect opening
of the subarachnoid space.
This phenomenon has been noted by others. It is unnerving, and
raises the question as to how long one must monitor these patients
before declaring failure. The authors have kept patients hospitalized
for an average of seven days. And since the first few days of moni-
toring did not predict outcome, it is difficult to know how to use the
information that was obtained.
The use of spinal taps was advocated by many neurosurgeons in
the past following removal of posterior fossa tumors to promote flow,
and avoid shunts. Its value was never really proven, and it is not clear
if it really made any difference here.
It has been our practice to place a ventricular reservoir when doing
ETV. This has the advantage that there is presumably less risk of
infection than with EVD, and it is left in to assess pressure months or
years after the ETV, and may serve as a method of decompressing the
ventricles in an emergency.
his is the report of a series of children who underwent endoscopic
third ventriculostomy (ETV) for a variety of conditions, and who
Leslie N. Sutton
During the period 1999-2004 they performed 126 ETVs. Sixty-four
children (68 procedures) were managed according to a protocol that
included continuous post-operative ICP monitoring and lumbar punc-
tures as necessary to relieve symptoms of intracranial hypertension. In
44 of the 64 cases, ETV was performed as the primary treatment for
non-communicating hydrocephalus. In the remaining 20 cases, it was
done at the time of shunt malfunction in an effort to eliminate the
shunt. The stated purpose of the present study was “to analyze
changes in ICP following endoscopic third ventriculostomy per-
formed in children affected by non-communicating hydrocephalus.”
he authors have drawn on their extensive experience with ETV to
investigate the patterns of ICP change following that procedure.
In addition, the potential utility of post-operative lumbar punctures
for control of transient increases in ICP is discussed.
The ICP changes observed after ETV were quite variable, corre-
sponding in general to one of several patterns. The ICP might remain
normal, be high initially with gradual normalization, or show a sec-
ondary rise in the days after surgery. What one would like to know is
whether there is any reliable predictive value in these observations
vis-a `-vis success or failure of the ETV procedure. The operation was
ultimately successful in 48 (75%) of the 64 children. In the group for
which the procedure was successful, 56% had a pattern of decreasing
ICP and 29% showed a secondary rise. In contrast, only 15% of the
failures had a decreasing pattern of ICP while 62% showed a second-
ary rise. These observations correspond to what one might expect, but
are not specific enough to have real predictive value. In fact one might
conclude that the success or failure of ETV in an individual case does
not depend on the initial level of ICP or the direction of change within
the first week or so after operation.
As a result of this investigation, the authors no longer consider
post-operative ICP monitoring necessary, except in cases where a
shunt has been removed. They conclude that, “On the basis of this
experience we now believe that post-operative ICP monitoring, al-
though useful, can be replaced by careful clinical observation, [and]
adequate serial post-operative imaging . . . in patients operated on for
primary hydrocephalus.” Although one might intuitively agree that
ICP monitoring is potentially useful, the results of this investigation
don’t in my opinion support that statement. In fact, the authors’
indication for performing an LP to relieve pressure is related to
symptoms rather than the measured level of ICP elevation.
Finally, the authors conclude that lumbar puncture is a useful
temporizing measure for controlling symptoms referable to post-
operative intracranial hypertension. Their experience bears this out.
The predictive value of LP is less clear. They state that, “In our
experience, lumbar puncture under ICP monitoring, has been a valid
diagnostic test to make the difference between ‘symptomatic adapta-
tion period’ and early failure.” If they mean that diagnostic LP for
pressure measurement is useful, one might agree. But then why is
concurrent ICP monitoring necessary? Their comment should be con-
sidered in the context of an earlier statement in the Results section
that: “The response of ICP to lumbar puncture was not predictive of
successful outcome or failure.”
Paul H. Chapman
CINALLI ET AL.
136 | VOLUME 58 | NUMBER 1 | JANUARY 2006