Introduction
Treating hydrocephalus has always been challenging, especially in children. One century
ago, there were no feasible options to treat hydrocephalus, and the natural history
of the disease was death by intracranial hypertension or progressive enlargement of
the cranial vault in infants.[1]
[2]
[3]
[4]
[5]
[6]
[7]
The Twentieth Century witnessed a dramatic change in the comprehension of the disease,
the surgical apparatus and the development of materials to enable shunting of the
cerebrospinal fluid (CSF) from the ventricles to other sterile body cavities. From
neuroendoscopy to shunt valve technologies, we could follow significant changes in
the paradigms of the treatment of hydrocephalus.[2]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
However, since the 1990s, a few improvements changed the learning curves and the outcome
of the hydrocephalus treatment. Currently, ventriculoperitoneal shunt (VPS) is the
mainstay for treatment, as well as endoscopic third ventriculostomy (ETV), which is
advocated in specific cases. Over the past decades, shunt revision rates became stable,
reaching over 70% of cases in a lifetime. Endoscopic third ventriculostomy also has
known limitations, especially in patients with myelomeningocele and Chiari type II.[5]
[6]
[7]
[28]
Ventriculosinus shunt (VSS), specifically when performed in a direction opposite to
the blood flow, as described by Ismail El Shafei in the 1980's, is a useful option
in the daily routine of neurosurgeons dealing with hydrodynamics. Although recent
experiences with VSS are scarce, we believe they should be discussed once more and
linked to the classic studies by El Shafei.[6]
[7]
[29]
[30]
[31]
The objective of this manuscript is to discuss new clinical data about retrograde
ventriculosinus shunt (RVSS) and link them to the pertinent literature.
Results
Since Toma et al[5] published a full review about VSS in 2010, only 5 manuscripts updated the subject,
1 by El-Shafei et al, and 4 by Oliveira et al and Pinto et al. In the present article,
we discuss the evolution of the technique and its current results.[6]
[8]
[10]
[22]
[25]
[28]
[32]
[33]
[34]
Review
Gartner, in 1896, suggested that the most physiological way to treat hydrocephalus
would be a connection of the ventricles with the venous blood system of the head and
neck. Portnoy proposed that the perfect system to treat hydrocephalus would be the
one that allowed for CSF drainage, could keep intracranial pressure (ICP) normal,
and would not cause complications.[3]
[5]
[14]
[20]
[26]
[27]
Ventriculoperitoneal shunt (VPS) is the preferred option to treat hydrocephalus, even
though it has a revision rate of more than 70% of patients throughout their lifetime,
and of 30% of patients in the first year. An important complication is the siphoning
phenomenon, which leads to over drainage and subdural effusions.[35]
[36]
Siphoning leads to over drainage of the intracranial CSF into the distal site when
the patient with VPS is in the orthostatic position. To solve this problem, anti-siphoning
devices (ASDs) were added to the VPS line, increasing the pressure of the drainage
system as the patient raises the head, thus decreasing the drainage of the CSF. Some
of them include gravitational devices, which change the pressure of the whole liquid
column according to a range of head positions, allowing different grades of additional
pressure. However, such solutions did not eliminate the hydrodynamic phenomenon of
siphoning, and significantly increased the cost of the prosthesis.[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[29]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
In the beginning of the XXI century, Aschoff stated that there were more than 300
shunt systems available; however, none could reach what was proposed by Gartner and
Portnoy. Several sites to perform the shunt were also tried, from the gallbladder
to the bladder, but the peritoneal cavity was the site in which the procedure reached
higher accuracy and effectiveness.[5]
In the 1940's, Ingraham induced hydrocephalus in dogs and performed treatment with
an anterograde ventriculojugular shunt; however, the distal end of the catheter obstructed
due to siphoning. When positioning the catheter against the blood flow, the catheter
flow would be laminar, and there would be no turbulence zone, which would prevent
blood clotting and catheter obstruction.[14]
This hemodynamic principle was demonstrated in the studies by El Shafei and El-Rifaii.
Another principle that was also demonstrated by these authors was the fact that the
internal jugular vein, which collapses in the orthostatic position, has the capacity
to work as a physiological anti-siphoning system. Thus, to use this natural property,
any ventriculovenous shunt catheter should have the distal extremity located proximally
to the internal jugular vein, in the opposite direction of the blood stream. Thus,
there is an improvement in hemodynamics with the removal of the regions of turbulence,
and the avoidance of back flow and thrombosis ([Fig. 1]).[8]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
Fig. 1 In the upstream region of the obstacle, there will be an impact zone, whereas in
the downstream region, there will be a weak zone.
RVSS Technique and Variations
El-Shafei et al pioneered the technique by performing a puncture in the retrograde
direction of the blood flow in the superior sagittal sinus (SSS) using the pressure
impact of the blood flow in the dural sinuses to maintain the CSF pressure greater
than the dural sinus pressure, regardless of changes in posture or intrathoracic pressure
([Fig. 2]). These authors modified their technique later by adding a valve of very low pressure,
recognizing that, in the event of a reduction in intracranial pressure (ICP) due to
lumbar puncture or sudden change of heart output, some retrograde flow of venous blood
to the distal catheter could hinder the shunt. They also broke new ground by using
transcranial Doppler (TCD) data to indirectly evaluate cerebral hydrodynamics through
the cerebral blood flow velocities.[8]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
Fig. 2 RVSS design.
The surgical technique as described by El-Shafei et al involves two cranial burr holes
in the same arcuate incision in the scalp on the parietal region. One burr hole is
in the posterior parietal bone (Frazier point), and the other, in the middle third
of the sagittal sinus. A small opening is made in the dura mater of the parietal incision,
the lateral ventricle is punctured, and then, a small opening is made in the sagittal
sinus, and the catheter is inserted ∼ 2 cm against the direction of the blood flow
([Fig. 3]). El-Shafei et al published the results with the use of a single catheter and with
the association of a very low pressure gradient valve (just to add another mechanism
to avoid back flow).[8]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
Fig. 3 Phases of the surgical technique. Positioning and arcuate incision. Opening with
medial and lateral burr holes. Catheterization of lateral ventricle and superior sagittal
sinus. Hemostasis. Closure with pericranium and skin closure.
In the small but increasing experience of our group, we have adopted the classic technique
by El-Shafei et al with a valveless catheter. However, in adult patients, we have
changed landmarks for ventricular and superior sagittal sinus (SSS) puncture. We used
a frontal incision to enable the puncture in Kocher's point, and another incision
over the SSS to allow for1 burr hole in the middle third of the sagittal sinus, ∼
3 cm behind the bregma. The lateral ventricle is punctured, and then a small opening
is made in the SSS, and the catheter is inserted ∼ 2 cm against the direction of the
blood flow. The catheter used to perform the shunt is, by convenience, the PS Medical
valve (Medtronic, Dublin, Ireland) ventricular catheter, which is already in routine
use at our institution. The ventricular catheter is the only one applied without the
use of a valve system ([Fig. 4]).[10]
[21]
[22]
Fig. 4 Retrograde ventriculosinus shunt after failed ventriculoperitoneal shunt. Left ventricular
puncture followed by burr hole over the SSS and puncture.
Another potential option designed by our group, but with limited application until
now, is the fragmentation of the catheter in two parts: one catheter for the ventricular
puncture, and another for the SSS puncture. Then, the margin of one catheter has a
connector to allow the connection of both catheters. The last variation with no clinical
applications described in the literature is a combination of two catheters of different
diameters connected as described before ([Fig. 5]). Using this rationale, it would be possible to increase or decrease flow resistance
and, therefore, the risk of back flow. However, these ideas should be further discussed
in experimental and clinical bases.[10]
[21]
[22]
Fig. 5 Option for retrograde ventriculosinus shunt catheter, with two catheters of different
diameters connected.
Experimental Basis for RVSS
Van Canneyt et al performed an experimental study to analyze RVSS. Firstly, an experimental
model of the cerebral ventricles, the arachnoid villi, the cortical veins, and the
SSS was built. Secondly, a numerical model of the cortical veins and the SSS was built.
No over drainage was found in the antegrade or the retrograde positions of the shunt.
Blood reflow was only found while mimicking lumbar puncture or changes in position
with the experimental model (lowering the intracranial pressure or increasing the
sinus pressure rapidly). These authors suggested that the main advantages of this
method include short length of derivation and prevention of siphoning. After an experimental
analysis, they concluded that when the catheter is placed in an anterograde manner,
the pressure at the catheter tip is 3.3 Pa, which is lower than that of the SSS. In
the backward position, the catheter tip pressure was 16.7 Pa, serving as a protection
against overdrainage.[6]
[12]
[13]
[14]
[15]
[25]
[26]
Pinto et al proposed and designed an animal experimental model for RVSS using mongrel
dogs and inducing hydrocephalus with kaolin. They stated that the RVSS technique was
feasible and even simple, not requiring a large learning curve.[25]
In the dogs with the RVSS model, they could successfully induce hydrocephalus and
its clinical features. The surgical procedure finished with no remarks, however they
failed to predict that the external diameter of the catheter used (2.5 mm) would prevent
the normal blood flow through the SSS, which also has a 2.5 mm caliber, resulting
in cerebral venous infarction due to venous blood stasis and drainage blocking of
the hypertensive CSF accumulated in the ventricles ([Fig. 6]). Thus, they proposed changing the external diameter of the catheter to adapt to
canine anatomy and proportions.[25]
Fig. 6 Catheterization of the superior sagittal sinus (SSS) in a canine model of retrograde
ventriculosinus shunt. The catheter had almost the same diameter as the SSS, so there
was venous blockage and infarction.
Clinical Basis for RVSS
El Shafei is surely the most important name in the experimental and clinical development
of RVVSs. The first article he published (along with another author) in 2001 included
56 patients and was followed by a second article with other authors in 2005, which
included 54 patients, producing a total of 110 patients reviewed retrospectively;
99 patients had valveless catheters, and 11 had catheters with low pressure valves.[8]
[9]
[10]
[11]
[12]
[13]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
In 2010, they published the last experience. During the past 40 years, 229 RVVSs were
implanted; 219 patients (95.63%) benefited from the shunt, and only 2 of them needed
a single shunt revision. The manifestations of increased intracranial pressure disappeared,
different degrees of regression of the ventriculomegaly occurred in patients with
closed craniums, but not in infants and young children with open craniums, and there
were no problems related to improper CSF drainage or venous thrombosis. The follow-up
period is of up to 40 years (mean 9 years 8 months).[8]
[9]
[10]
[11]
[12]
[13]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
El-Shafei showed that, in children, the clinical improvement occurred immediately
after shunt insertion with normotensive fontanelle and disappearance of the hairy
scalp engorged veins, with gradual relief of eye signals. The psychomotor development
of the patients was satisfactory.[8]
[9]
[10]
[11]
[12]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
Paul C. Sharkey, in 1965, published his experience with 4 pediatric patients. Fourteen
years later, in 1979, Cecil J. Hash published a series of 36 procedures. And Wen reported
15 cases of obstructive hydrocephalus in 1965, and then changed his technique in 52
patients in the pediatric age group, and published the results in 1982.[5]
Børgesen, Gjerris and Agerlin also applied the technique. The first series was published
in 2001. The second article was published in 2004 with two extra authors, and included
the follow-up results of 45 patients from Denmark, Italy and Norway.[1]
[5]
[45]
Most series treated all kinds of hydrocephalus in children and adults of all age groups.
However, cases of spina bifida were excluded. All authors reported clinical and radiological
improvement. The clinical findings, such as headache, nausea, vomiting and papilledema
changed after the surgery.[5]
The follow-up period and shunt revisions were variable. The revision rate varied between
11.8% and 37.5%. Most patients had a shorter follow-up period. In healthy young children,
brain growth causes an increase in the thickness of the cerebral mantle and head circumference,
while the ventricular size remains stable, that is, in normal situations, the strength
of the brain wraps is lower than the CSF pressure, allowing centrifugal growth.[1]
The last review on the subject showed that among 265 patients, there were no reports
of sinus thrombosis, air embolism, uncontrollable operative bleeding or nephritis
associated with the shunt.[5]
RVSS versus VPS in Hydrocephalus after Myelomeningocele Repair
Myelomeningocele is a neural tube defect in neurulation embryonic stage by the third
week of life. Its incidence may reach 1–2 per 1,000 live births. Even with all advances
in intrauterine and postnatal management, up to 90% of these patients may present
hydrocephalus in different degrees, requiring treatment ([Fig. 7]). Ventriculoperitoneal shunt and ETV are classic options, but revisions rates are
high in VPS; ETV still has a limited success in cases of Chiari II, especially in
younger infants.[7]
[16]
[23]
[46]
Fig. 7 Post-operative aspect of a surgically repaired myelomeningocele.
Oliveira et al compared RVSS and VPS in a randomized trial with preliminary results.
A total of 9 patients were included in the study for a period of 2 years. Four patients
were allocated to the control group (VPS), and 5 patients to the RVSS group. Transcranial
Doppler was used to analyze indirect hydrodynamic patterns, as reported previously
by El-Shafei.[27]
[47]
[48]
The follow-up included evaluations of the cephalic perimeter (CP), bregmatic fontanelle,
TCD, computed tomography, neuropsychomotor development, complications and shunt revision,
and outcomes focused on patients and caregivers.[22]
The cephalic perimeter, the tomography and the fontanelle were characteristic in each
group. In the RVSS group, there was no abrupt reduction in head circumference after
surgery. The CP assumed a physiologic contour, but remained high. In control group,
the CP reduced abruptly after surgery, with subsequent regrowth. Similarly, the fontanelle
in the RVSS group remained normotensive, meaning cranial normotension. In the control
group, they became normotensive or hypotensive, especially in older children, and
could imply a siphoning component in older children with a more upright posture ([Figs. 8] and [9]).[22]
Fig. 8 Pre-operative evaluation before retrograde ventriculosinus shunt. Normal ventricular
size at the age in months. Hydrocephalus at the age of 6 months. Normal sinus anatomy.
Fig. 9 Postoperative retrograde ventriculosinus shunt evaluation. There is no clear decrease
in ventricular size.
The TCD showed significant improvements in cerebral hemodynamics parameters after
treatment in both groups, with decreased velocity of intracranial vessels and a decrease
in the pulsatility and resistance indexes ([Figs. 10] and [11]). There were no differences between VPS and RVSS. The TCD associated with clinical
and physical examinations proved to be a safe and appropriate tool to assess dysfunction
of RVSS.[6]
[7]
[22]
[30]
[31]
Fig. 10 Duplex evaluating ventricular size. There is no clear ventricular size decrease before
(left) and after (right) retrograde ventriculosinus shunt.
Fig. 11 Transcranial Doppler (TCD) of a retrograde ventriculosinus shunt (RVSS) patient.
Pre-operative and postoperative TCD. Postoperative mean velocities in middle cerebral
artery (MCA) decreased after RVSS, and there was a decrease in pulsatility index (PI).
In the RVSS group, there was need for one revision and one conversion to VPS. In VPS
group there was no revision or complications. Psychomotor development after surgery
was similar in both groups. However, further evaluation is needed with specific neuropsychological
tests which can detect subtle changes.[22]
RVSS and Failed VPS
Peritoneal cavity alterations are one of the most frequent causes of VPS failure,
due to the formation of primary or secondary adhesions, fibrosis, ascites, cysts,
pseudocysts and peritonitis.[10] Although there are other second options, like ventriculopleural and ventriculoatrial,
they may be useful, but they are potentially harmful.[10]
Oliveira et al described three cases of RVSS applied in situations when VPS is not
feasible, or the peritoneum is not useful. In one case, there was refractory ascites;
in another, a peritonitis, and in the last, pseudocyst formation. In all three cases,
the patients were adults and tolerated RVSS to treat the underlying hydrocephalus.
One patient needed revision surgery to reposition the catheter, which was not inside
the SSS. Although applied in adults, this experience also sheds light over the potential
use of RVSS in children with a hostile peritoneum.[10]
Discussion
Retrograde ventriculosinus shunt is a safe and more physiological option that requires
the use of less prosthetic material. It is feasible and applicable, as long as there
is pervious SSS to allow drainage of excessive CSF. It uses a natural anti-siphon
device (internal jugular vein), and enables CSF drainage inside the venous system.
Manipulation of the SSS is safer than usually believed, and one important matter is
that the dural opening for the ventricular puncture should be just enough to catheterize
the ventricle and avoid creating CSF leakage.[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
The main advantages are the use of less prosthetic material, the safety and efficacy,
respecting hydrodynamics and avoiding over drainage. The main disadvantages are: the
necessary learning curve and the possibility of retrograde flow if a lumbar puncture
is performed (which can be solved if a low-pressure valve is added to the shunt system).
Complications like dysfunction and infection are not common, and overdrainage is an
unlikely event.[8]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
In RVSS, although there is no immediate and marked ventricular size decrease (which
is common in VPS), there is stable ventricular size with improved hemodynamics and
hydrodynamics revealed by a decrease in the pulsatility index (PI) after surgery,
which is maintained in the late follow-up. In patients with increased PI there was
also clinical impairment, which was indicative of the need of a revision surgery.[8]
[29]
[32]
[33]
[34]
[40]
[41]
[42]
[43]
[44]
[49]
Especially in children, RVSS generates a normotensive state after shunting, allowing
centrifugal head growth, once there is no intracranial hypotension due to over drainage,
which may reflect in long-term better psychomotor development. Additionally, there
is no need to operate the abdomen, avoiding complications of disabsortive peritoneum.[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[49]
In our sample, TCD was used successfully for diagnostic and follow-up evaluation of
hemodynamics and hydrodynamics in the pre- and postoperative phases of RVSS. It was
technically feasible in all patients, correlated closely with other clinical and imaging
parameters, and was sensitive for the identification of system malfunction.[49]
