Spina Bifida Hydrocephalus and Shunts

CSF is typically produced by the choroid plexus of the ventricles and circulates in one direction from the lateral ventricles to the third ventricle and through the aqueduct of Sylvius to the fourth ventricle.

From the fourth ventricle, CSF exits the brain through three separate openings: one in the midline (foramen of Magendie) and one on either side (foramina of Luschka). It enters the subarachnoid space at the foramen magnum, circulates down to the spine, and then circulates up again to the surface of the brain, where it is absorbed at the arachnoid granulations. These are sievelike structures where the CSF enters the venous circulation, leading to the sagittal sinus.

Several factors are implicated in the etiology of hydrocephalus in children with myelomeningocele, including the following: 

• Arnold-Chiari II malformation
• A degree of aqueductal stenosis
• Anomalous venous drainage in the posterior fossa caused by compression of the sigmoid sinuses
• Open myelomeningocele
• Presence of other central nervous system (CNS) malformations

Extensive deformity of the posterior fossa and its structures is associated with Arnold-Chiari II malformations. The brainstem has an abnormal disposition with respect to the midbrain and the tentorial hiatus, the posterior fossa has smaller capacity than usual, the fourth ventricle is displaced caudally, and the cerebellar tonsils through the foramen magnum are significantly prolapsed. These anatomic factors all contribute to the impairment of CSF circulation.

Although the development of hindbrain herniation during gestation in myelomeningocele was once thought to be the result of cord tethering “pulling” on the brain tissue, hindbrain herniation is now believed to be caused by the progressive caudal migration of the hindbrain in association with the low-pressure conditions created in the spine by the open myelomeningocele. Continued loss of CSF soon after birth exacerbates the hindbrain hernia and the associated hydrocephalus.

This can lead to acute neurologic deterioration caused by a combination of raised intracranial pressure (ICP) related to the ventriculomegaly and acute bulbar dysfunction caused by compression of the brainstem in the foramen magnum region. The neurologic state usually improves after ventricular shunting. In most patients, ventriculomegaly gradually develops within the first few weeks or months of life.

Hydrocephalus development can be temporally related to the closure of the myelomeningocele. In a small group of patients with open myelomeningocele, dramatic deterioration occurs after closure of the defect. The impaction of the hindbrain hernia plays a significant role in this acute deterioration.

In addition to the presumed effect of the low-pressure leak during gestation on the development of Arnold-Chiari II malformation, the abnormal and exposed spinal tissue is theorized to sustain damage in utero through trauma and exposure to neurotoxic substances in amniotic fluid. This is thought to lead to neurologic deficits that are worse than if the exposure did not occur.

The incidence of infantile hydrocephalus has been estimated at 3-5 cases per 1000 live births. The peak ages of presentation in this group include the first few weeks of life, age 4-8 years, and early adulthood. The latter two peaks represent delayed presentations of infantile hydrocephalus.

An estimated 750,000 people have hydrocephalus, and 160,000 ventriculoperitoneal shunts are implanted each year worldwide. About 56,600 children and adolescents younger than age 18 years have a shunt in place.

The incidence of myelomeningocele is in the range of 0.2-2 per 1000 live births. The overall incidence of myelomeningocele has significantly declined in the past two decades because of improved maternal nutrition during pregnancy, including the addition of folic acid, a wider availability of antenatal diagnosis, and therapeutic termination of pregnancy.

In a significant proportion of patients with open spina bifida, hydrocephalus is absent at birth but develops in the first few weeks or months of life. Hydrocephalus occurs in 15-25% of children with open myelomeningocele at birth; however, in most surgical series, the proportion of patients with myelomeningocele who require shunting reaches 80-90%.

In a retrospective chart review, shunt placement was shown to vary according to the level of the lesion, with a greater number of patients with thoracic lesions requiring shunts than those with lumbar or sacral lesions. Lesions at the levels of T12 and above have also been associated with increased incidence of brain abnormalities and lower scores on psychometric testing than lesions at L1 or below.

In the 1940s, before shunting was established, children with hydrocephalus had a poor prognosis. Most patients were not offered treatment, and only 20% of children who did not undergo surgery for hydrocephalus reached adulthood. Furthermore, children who survived had a 50% chance of having permanent brain damage. Outcomes improved after the introduction of valved shunt systems by Nulsen and Spitz in 1952 and after the development of silicone systems by Holter and Pudenz in the 1960s.

The mortality associated with initial shunt insertion is approximately 0.1%; mortality due to shunt failure is in the range of 1-4%. Although shunting is necessary in children with myelomeningocele and hydrocephalus, those who require shunting have been shown to have a shorter lifespan than those who do not. This is likely due to the severity of the lesion and complications of shunting.

Most children with hydrocephalus currently reach adulthood if the shunt is appropriately maintained. In a 20-year follow-up survey of children who received shunting in the 1970s, more than half of them graduated from mainstream education.

In a retrospective study of the success rates of endoscopic third ventriculostomy (ETV) in 51 children with obstructive hydrocephalus,[4]  Duru et al found that outcomes were most favorable in patients younger than 6 months and in those whose hydrocephalus derived from aqueductal stenosis rather than other causes (eg, spina bifida).

The outcome of patients with spina bifida has also improved.[5] In a review of a cohort of patients treated in the 1970s for spina bifida aperta, 52% of the patients were alive 20 years after treatment. Most of the deaths occurred in the first year of life, mostly due to renal and respiratory problems associated with spina bifida; only a few were related to hydrocephalus. Chern et al concluded that surveillance imaging of children with spina bifida aperta and shunted hydrocephalus decreased emergencies during follow-up but had no clear effect on mortality and morbidity.[6]

In a review of children treated in the 1980s, only 27% died; most of them died in the first year of life from causes related to spina bifida rather than hydrocephalus.

In a survey of adults with spina bifida, 6% of patients died of shunt-related problems or died after craniovertebral decompression for Arnold-Chiari II malformation.

Infants with hydrocephalus develop an enlarging head with bulging fontanelle, enlarged scalp veins, macrocrania, suture diastasis, and positive Macewen (ie, cracked pot) sign. If the hydrocephalus is not treated, these infants develop sunset eyes, recurrent vomiting, and, later, respiratory arrest. Persistent leakage of cerebrospinal fluid (CSF) from the repaired spinal wound almost invariably indicates active hydrocephalus, even if the ventricular size is only modestly enlarged and the anterior fontanelle is not bulging.

The particular concern in children with myelomeningocele is the presence of hindbrain hernia in the context of the Arnold-Chiari II malformation, which can cause early clinical symptoms of bulbar palsy due to compression of the brainstem and can remain unnoticed by inexperienced observers or be confused with symptoms of shunt malfunction or untreated hydrocephalus.

Manifestations of brainstem dysfunction caused by hindbrain hernia and aggravated by ventricular dilatation include the following:

• Poor feeding
• Recurrent vomiting
• Poor sucking
• Generally subdued behavior with poor crying
• High-pitched cry or stridor caused by vocal cord paralysis - A predictor of poor outcome
• Episodes of apnea
• Extremity weakness in older children
• Recurrent aspiration - Often manifesting as recurrent pneumonia

Approximately 20% of children with myelomeningocele who also have an Arnold-Chiari II malformation develop brainstem symptoms. Myelomeningocele is also a contributor to mortality and morbidity in the first two decades of life.

Older children with closed fontanelles develop clinical signs of intracranial hypertension without progressive head enlargement. They develop headaches, blurred vision, decline in intellectual performance, and gradual drowsiness, which, if left untreated, lead to coma and death due to respiratory arrest.

Currently, most fetuses undergo scanning with ultrasonography (US) in utero. US permits good identification of any ventricular dilatation that indicates active hydrocephalus. In such cases, or in patients who present to pediatricians or family physicians with progressive head enlargement, US is typically performed first because it is widely available and does not expose the child to ionizing radiation.

In babies with open fontanelles and large heads, US reveals the enlarged ventricular system and any mass lesions or hemorrhage. However, the anatomic detail produced by US remains poor and serves only as a guide to further investigations.

Once hydrocephalus is suspected, either clinically or on the basis of US findings, the diagnosis must be confirmed with a more detailed investigation. In most parts of the world, computed tomography (CT) remains the most widely available neuroimaging investigation. Because image acquisition is rapid, the study does not require sedation of the child and, in most cases, is helpful for obtaining an accurate diagnosis.

CT provides a good image of the dilated ventricular system and any obstructive lesions (eg, brain tumors) or associated abnormalities (eg, arachnoid cysts). Anatomic detail is generally good, and in the vast majority of cases, proceeding to treatment on the basis of CT findings is regarded as safe. However, if endoscopic treatment is being considered, magnetic resonance imaging (MRI) should be performed.

MRI has played an increasing role in the management of hydrocephalus in children. Unfortunately, it still is not as widely available as CT.

The need for sedation (or even general anesthesia) to obtain good images is a major consideration, especially in very young infants, because acquisition takes several minutes or more and any movement will cause severe deterioration of picture quality. This is particularly problematic in children with associated congenital conditions that cause poor respiratory drive.

MRI shows structures of the brain with superior anatomic detail and assists the surgeon in choosing the best treatment. MRI is especially necessary in performing endoscopic treatment, where the surgeon must verify the presence of aqueductal stenosis or any intracranial cysts that may be fenestrated, and in determining the relations of important anatomic structures (eg, the floor of the third ventricle to the basilar artery) in children with aqueductal stenosis.

The use of phase-contrast cardiac-gated sequences can provide information on the flow of cerebrospinal fluid (CSF) through the aqueduct, and ventriculostomy can provide information on the patency of the stoma. (See the image below.)

Children born with open myelomeningocele have a very high (close to 100%) incidence of hindbrain hernia (see the image below).

Marked hydrocephalus is present in as many as 15-20% of these patients, and a further 20% have moderate ventriculomegaly at birth. A large proportion of patients (as many as 90%) eventually develop clinical hydrocephalus. The lateral ventricles have a characteristic appearance in almost all patients with spina bifida (see the image below): The occipital horns are more dilated than the frontal horns, and the long axes of the lateral ventricles tend to be parallel.

A contributing factor may be the partial or complete absence of the falx and the absence of the septum pellucidum in a very large proportion of these patients. MRI produces a good image of the hindbrain hernia, the small posterior fossa, and the midbrain deformity with kinking of the aqueduct.

In a study of fetuses with ventriculomegaly, Pier et al found that whereas neither two-dimensional (2D) nor three-dimensional (3D) measurements could predict whether a fetus would have a normal outcome, live birth in fetuses with ventriculomegaly was linked to ventricular volume and diameter, though not parenchymal volume.[7] The investigators evaluated the postnatal outcomes of 307 fetuses with ventriculomegaly using 2D MRI measurements of lateral ventricular width and 3D measurements of lateral ventricular and supratentorial parenchymal volumes.

Indications

Children with a clinical picture of active hydrocephalus and significant ventriculomegaly (often with evidence of periventricular lucency indicating raised cerebrospinal fluid [CSF] pressure in the ventricular system) require treatment early in life. However, children with mild or moderate ventriculomegaly and a head circumference within the reference range may not require initial treatment. Watchful waiting may be adopted during the first few months of life. Head circumference monitoring and repeated ultrasonography (US), computed tomography (CT), or magnetic resonance imaging (MRI) are helpful in deciding whether shunting is required.

Age, prematurity status, and weight are significant considerations for the timing of treatment in children. In general, shunting is avoided, if possible, in children younger than 6 months because they have an increased risk of infection. For the same reason, shunting should be deferred, if possible, in premature babies who weigh less than 1.5 kg until they have gained weight. (These considerations often arise in children with posthemorrhagic hydrocephalus because they are often premature and small for their age.) Over the past few decades, delayed treatment of hydrocephalus has become more common, and the number of hydrocephalic myelomeningocele patients not treated on initial inpatient stay has increased.[8]

If feasible, shunts should be placed when the myelomeningocele is closed, because this appears to protect against CSF leakage from the spinal wound, which can lead to shunt infection, and improves the chances for better development by reducing intracranial hypertension early. One of the signs of ongoing hydrocephalus after closure of myelomeningocele is persistent CSF leakage.

Children with open myelomeningocele in whom closure of the defect has been delayed may already have CSF infection. In such circumstances, CSF microbiologic testing should be performed; if CSF infection is present, external ventricular drainage should be performed for 7-10 days in conjunction with antibiotic treatment, until the CSF infection is controlled and a shunt can be inserted.

Although children with intraventricular hemorrhage (IVH) may have active hydrocephalus early in life, shunting is often difficult to consider because the CSF is heavily blood-stained, the protein content is too high (>1 g/dL), or both. In such cases, the traditional approach is to insert an external ventricular drain until the CSF clears and a shunt can be inserted. Injection of tissue plasminogen activator (TPA) into the ventricles has been attempted in an effort to accelerate blood clearance from the CSF, with moderate success. This therapeutic maneuver has not been universally accepted.

Shunting in older children and young adults

An issue that merits attention is the need to decide whether shunting is indicated in older children or young adults who have myelomeningocele and untreated ventriculomegaly. Some patients have the typically shaped ventricles of hydrocephalus that are caused by spina bifida but do not appear to have tension, with no periventricular flow of CSF as seen on T2-weighted MRI and no symptoms (eg, headache, drowsiness, diplopia, bulbar features) that suggest active hydrocephalus. If these patients never receive shunting, they should be serially monitored with intelligence and psychometric testing.

If a patient has no clinical symptoms and psychometric test results indicate stability, shunting based solely on radiologic appearance should be discouraged on the grounds that the risks outweigh the benefits, and serial follow-up should be continued.

Contraindications

Relative contraindications for treatment of infantile hydrocephalus include severe central nervous system (CNS) malformations that are regarded as incompatible with normal development, such as some of the congenital neurodevelopmental syndromes associated with severe malformation of a large part of the cerebral substance (ie, anencephaly). In these cases, neonatologists prefer to counsel parents against treatment of hydrocephalus. The same considerations are applied to severe cases of IVH in which radiologic investigations clearly indicate that the hemorrhage has damaged significant large parts of the brain.

Intervention on existing shunt

Any intervention on the shunt should be considered cautiously in patients who already have received shunting. Because shunts may be disconnected or may appear to have long since stopped working, it is tempting to regard the situation as compensated hydrocephalus and thus elect to remove the shunt, especially if it is causing local discomfort in the neck. However, shunts that have been implanted for years acquire a tube of strong fibrous tissue that surrounds them along their entire length.

Although the tube may appear fractured on radiographs, CSF is bridging the gap, guided by the encircling fibrous tube. Such shunts are actually functioning, and any attempt to remove them without instituting any alternative means of CSF drainage (eg, third ventriculostomy) may cause neurologic deterioration.

However, if the patient has subtle symptoms or a declining intelligence on psychometric testing, treatment should be offered. In such cases, patients who have not received shunting should receive them; in patients with shunts, the shunt systems should be evaluated and explored and revised if necessary.

In general, medical therapy of hydrocephalus is far inferior to surgical management of the condition. Medical therapy has been used with limited success in an attempt to avoid shunting in patients with posthemorrhagic hydrocephalus.

Antenatal administration of steroids may modestly reduce the incidence of IVH in premature infants. Postnatal use of indomethacin has been shown to reduce severe IVH in some studies, though improvement in cognitive functioning has not been documented.

Diuretic therapy, such as the use of acetazolamide and furosemide, was invalidated as a therapy in a study of 177 infants. In this study, the medications did not affect rates of shunt placement and may have impaired neurologic outcome. Fibrinolytic therapy is under investigation.

Serial lumbar puncture has been frequently performed after IVH to prevent or manage developing hydrocephalus, but no clear guidelines have indicated when to initiate treatment, and no explicit evidence of effectiveness has been reported. Early use of lumbar puncture (prior to evidence of head expansion) has been shown by meta-analysis to have no benefit.

Preparation for surgery

Choice of shunt placement site

In most cases, surgical treatment of hydrocephalus consists of ventricular shunt insertion. The shunt is an artificial device, made mostly of plastic (though some parts may be metal), that includes a catheter inserted in the ventricle of the brain (with a one-way valve allowing unidirectional flow of CSF out of the brain) and a distal catheter that drains CSF to an extracranial location in the body.

Although in most cases, the preferred distal site is the peritoneum, other sites are available (eg, right atrium, pleura, gall bladder, ureter, bladder, and sagittal sinus) in patients with coexisting abdominal problems. In current practice, however, the overwhelming majority of shunts are ventriculoperitoneal.

Selection of valve design

All shunts are designed to maintain normal intracranial pressure (ICP). More than a dozen different commercial shunts are currently available. The design of the valve is controversial. Essentially, two types of shunts are available: the pressure-regulating shunt and the flow-regulating shunt (as well as two brands of programmable shunt valves).

The pressure-regulating shunts are designed to maintain a difference of pressure between their inlet and their outlet and to allow flow of CSF once the preset pressure has been reached. The flow-regulating shunts are designed to allow a constant flow of CSF, simulating the normal flow.

Despite different designs, large randomized trials have been unable to demonstrate differences between the various types. Different types of valves are seemingly associated with different types of complications. For example, the pressure-regulating valves are more prone to cause overdrainage complications, whereas the flow-regulating valves are more prone to cause valve obstruction.

A significant effort in research and development is directed at shunt valve design. Although advances have been achieved, there is still room for improvement. The 30% failure rate in the first year is regarded as high, but no means of reducing it has been found.

Operative details

Anesthetic factors must be considered, particularly those relative to respiratory function and reserve. Many of these patients are premature neonates who have poor respiratory reserve and may be experiencing physiologic jaundice when surgery is required.

Because the magnitude of the surgery is not extensive, the circulating blood volume is usually not problematic, and a blood transfusion is likely to be unnecessary during shunt insertion or endoscopic ventriculostomy, unless a preexisting problem is present.

The importance of avoiding hypothermia and excessive blood loss in pediatric patients cannot be overemphasized, especially in neonates. Heat and blood loss account for significant morbidity and mortality in pediatric operative correction. Having a dedicated pediatric anesthesia team and having a capable operating room team are equally important. In addition, limiting operating room traffic during these procedures can be helpful in decreasing infections.

In most cases, shunt insertion involves making a posterior parietal or frontal bur hole through a small linear or curved skin incision. The peritoneal cavity is entered through a small linear incision either in the upper midline epigastric region or in the right upper quadrant. The distal tube is passed from the cranial to the abdominal wound by using a purpose-designed tubular dissector advanced in the subcutaneous fat.

In typical cases, if a posterior parietal bur hole is used, the shunt valve is situated behind the ear, avoiding the need for a step incision, and is usually easily palpable by the patient and parents. Administering a prophylactic antibiotic, usually cephalosporin or vancomycin, is common at the commencement of the operation to lessen the likelihood of shunt infection.

In children born with open spina bifida who undergo simultaneous shunting and myelomeningocele closure, additional precautions should be taken to maintain the sterility of the surgical fields. Most neurosurgeons prefer to close the myelomeningocele with the child prone and to subsequently turn the child on his or her back for the shunt placement while adequately protecting the newly repaired spinal wound with ample padding.

Finally, 20-40% of children with myelomeningocele have an allergy to latex. Care must be exercised to avoid contact with latex products during surgery and postoperative hospitalization to minimize the risk of anaphylactic reactions.

Endoscopic third ventriculostomy (ETV), first performed by Walter Dandy in the 1910s with moderate success, has experienced a resurgence. The endoscopic equipment has improved, which has resulted in increased use of the procedure. ETV has a success rate of 70% or higher when used in patients with aqueductal stenosis and is regarded by many as the procedure of choice in these patients.[4] Endoscopic cyst fenestration can be used in the presence of arachnoid cysts in various locations (ie, suprasellar, interhemispheric, posterior fossa), with variable success.[9]

ETV has also been performed to treat hydrocephalus in children with myelomeningocele. However, the reported success rates are only about 30-40%. One possible explanation for the low success rate of ETV is that most patients are infants or neonates when they receive initial treatment and do not have fully developed subarachnoid spaces.

A frontal ventricular catheter attached to a blind reservoir or an Ommaya reservoir can be left in place and can be converted to a ventriculoperitoneal shunt if the third ventriculostomy fails.

ETV can be performed in children who have already received shunting and who present with shunt malfunction at an older age.[10] The reported success rate is approximately 50%. In such patients, an external ventricular drain should be used for the first few days following ETV (especially if the shunt has been removed) to allow emergency decompression if the third ventriculostomy does not function adequately and the patient's condition rapidly deteriorates.

ETV may be more effective if it is combined with choroid plexus cauterization. Improved outcomes were reported in a study of select patients younger than age 1 year. However, cauterization is not routinely performed and has not been established as essential.

Kulkarni et al developed and validated a model to predict the probability of success of ETV in the treatment of hydrocephalus.[11]  Analysis of 618 ETVs performed consecutively on children at 12 international institutions identified predictors of success at 6 months. A multivariable logistic regression model was developed on 70% of the dataset and was validated on 30% (validation set). The model contained patient age, cause of hydrocephalus, and previous CSF shunt.

ETVs were successful and the model maintained good fit, discrimination, and calibration on 64.4% of the validation set.[11] Kulkarni et al concluded that this model can be used to identify children in whom ETV is most likely to succeed and who can thus be spared the long-term complications of CSF shunting.

Operative details

ETV is traditionally performed through a frontal bur hole situated just anteriorly to the coronal suture. A rigid or flexible endoscope is preferred. The third ventricle floor is perforated by using a purpose-designed monopolar diathermy with a retractable tip or another similar purpose-designed dissector. After formation, the stoma is commonly dilated using some kind of purpose-designed balloon dilator.

Perforation of the third ventricle floor is the most delicate and important phase because perforation of the adjacent basilar artery is a risk. ETV can be particularly difficult in children with myelomeningocele because the ventricular anatomy is often abnormal, the third ventricle floor is thicker and more difficult to penetrate, the size of the third ventricle is smaller in these children than in those with aqueductal stenosis, or the septum pellucidum is absent, which can lead to disorientation in the inexperienced operator.

In general, inexperienced operators should avoid ETV in children with hydrocephalus caused by myelomeningocele. Apart from damage to the basilar artery, another potential source of intraoperative difficulty is damage to the choroid plexus, which can lead to hemorrhage that clouds the operative field. Most nonarterial bleeding stops with gentle warm irrigation.

Failure to perforate the Liliequist membrane may also result in ETV failure. Preoperative MRI is very important because it reveals the bowing of the third ventricle floor and its relationship to the basilar artery. Bowing of the third ventricle floor correlates with a pressure gradient between the ventricular system and the extraventricular CSF spaces. If the third ventricle floor is not bowed, the success rate of ETV is significantly decreased.

In cases of shunt revision or shunt removal after successful ventriculostomy, rupture of the choroid plexus during retrieval of the ventricular catheter is common and can lead to life-threatening hemorrhage. Different techniques can be used to avoid this complication; the most common of these techniques involves insertion of a stylet into the catheter lumen, which allows coagulation with the diathermy before the catheter is retrieved. However, if the ventricular catheter is not easily removed, it should be left in place and an additional catheter inserted. Image guidance can also be helpful in ventricular catheter placement, especially in patients with loculated hydrocephalus and cannulating complex cysts.

Several surgical centers have attempted to lessen the effect of Arnold-Chiari II malformation, neurologic impairment, and need for shunting[3] through the use of in-utero surgical repair of the myelomeningocele. The first such repair was performed in 1997, after several animal models demonstrated a benefit from the procedure. Studies of the outcomes of these surgical procedures have been somewhat promising, though not conclusive.[12, 13, 14]

A reduction in shunt-dependent hydrocephalus has been shown in children who underwent surgical correction at less than 25 weeks’ gestation, who had ventricular measurements of less than 17 mm at time of surgery, and who had an anatomic level lower than L3. An approximately 50% reduction in the need for shunts was reported in this select group. However, long-term follow-up was lacking.

One study demonstrated a decrease in hindbrain herniation with improvement on serial fetal scans. However, another study failed to demonstrate any difference in the progression of ventriculomegaly between patients who underwent intrauterine repair and control subjects.

Several small studies investigated leg and neurologic function and did not provide clear evidence that in-utero repair improves either function. A study involving 30 children (43% of whom required a shunt) investigated neurodevelopmental outcome at 2 years following treatment. The study revealed that 87% of the children had normal or mildly delayed cognitive language and personal-social skills.

In-utero repair is certainly not without risk to both the mother and the fetus. In fetuses, a 4% risk of mortality and an 11% risk of morbidity (primarily from prematurity) is associated with the surgery. In mothers, uterine dehiscence, uterine rupture, and hysterectomy are also risks. Development of minimally invasive techniques may improve these outcomes, but they have not yet been fully established as successful options.

As a consequence of the lack of clarity regarding the risks and benefits, a randomized, controlled trial (the Management of Myelomeningocele Study [MOMS]) funded by the National Institute of Child Health and Human Development was initiated at three centers in the United States in an attempt to achieve a definitive evaluation of the risks and benefits of in-utero repair of myelomeningocele.[15]

In a 2011 report from the MOMS investigators, myelomeningocele repair before 26 weeks’ gestation was associated with a decreased risk of death or shunting before age 12 months.[15]  In addition, mental and motor function scores were improved at age 30 months. Improvement was also noted in secondary outcomes, including hindbrain herniation by 12 months and walking by 30 months. This study was stopped for efficacy because statistical significance was established. However, the risk of preterm delivery was increased, and uterine dehiscence at delivery was associated with the antenatal surgery. Long-term outcomes have not been established.

Additional findings from the MOMS trial, detailed in subsequent reports, included the following:

• Larger (≥15 mm) ventricles at initial screening are associated with an increased need for shunting among those undergoing fetal surgery for myelomeningocele ^[16]
• Antenatal surgery does not significantly decrease the need for clean intermittent catheterization (CIC) by 30 months of age but is associated with a lower frequency of bladder trabeculation and open bladder neck ^[17]

The CECAM (Cirurgia Endoscópica para Correção Antenatal da Meningomielocele) investigators reported good results from fetoscopic treatment of open spina bifida with their particular surgical technique, which included placement of a biocellulose patch over the lesion and simple closure of the skin.[18]  Phase II studies are required for further assessment of this approach.

In a study of 21 consecutive fetuses who underwent intrauterine surgery to repair open spina bifida, Etchegaray et al found that whereas this approach was associated with an increased risk of preterm delivery and premature rupture of membranes, it nevertheless significantly reduced the need for postnatal treatment of hydrocephalus and improved short-term motor outcomes.[19] Their findings were similar to those from the MOMS trial.

After successful completion of shunt insertion or revision in patients in whom a differential-pressure valve has been implanted, elevation into the upright position is commonly avoided to prevent shunt overdrainage and the formation of a subdural hematoma (SDH). In contrast, patients in whom a flow-regulating shunt has been implanted are commonly elevated to 30° or more the day after the implantation to promote CSF drainage. The efficacy of these practices is unclear. Feeding of very young babies can commence soon after surgery.

Patients who undergo shunt revision tend to recover quickly, and they are usually discharged home a few days after surgery.

Shunt insertion

The patient and the doctor must have an ongoing commitment to manage the complications associated with shunting. Most shunt obstructions are related to obstruction of the ventricular catheter by glioependymal tissue, which grows into the lumen from the ventricular wall through the draining holes.

In many shunt revision operations, as many as 30% of IVHs occur during removal of the old catheter because of rupture of the choroid plexus, which has grown into the shunt lumen.

Symptoms of acute shunt obstruction include headache, nausea and vomiting, papilledema, cranial nerve VI palsy, change in personality, and the setting-sun sign (lack of upward gaze) in infants. Chronic failure may be heralded by accelerated head growth, loss of milestones, papilledema, optic atrophy, and change in seizure frequency.

A significant late complication is fracture or destruction of the shunt tube due to material degradation and fatigue. (See the image below.)

Common locations for distal tube fracture include the occipitocervical junction, the root of the neck, and the junction between the inferior border of the ribs and the abdominal wall. These are points of maximal mechanical stress where the material is degraded most.

Shunt complications can be divided into the following three categories:

• Mechanical
• Infective
• Overdrainage-related

Mechanical complications

As many as 80% of shunts develop mechanical complications at some stage, and one third to one half of these complications occur within the first year of shunt placement. An additional 15% of shunts fail in the second year, and 1-7% of shunts per year fail after the second year. On average, each patient is likely to undergo two or three operations throughout childhood for shunt revision.

Infective complications

Infective complications occur in 5-10% of all shunt operations (though some have cited lower figures[20] ) and are more common in younger patients, especially in those younger than age 6 months. Most shunt infections manifest in the first 3 months after insertion, and almost all present within the first 6 months. Staphylococcus is the most common offending organism. The use of shunts impregnated with antistaphylococcal antibiotics may reduce the incidence of shunt infection.[21]

Symptoms of shunt infection include redness and swelling along the surgical incision site, tenderness over the reservoir, swelling or drainage, nuchal rigidity, or abdominal pain. During the 1970s, when ventriculoatrial shunts were commonly used, an appreciable number of patients experienced bacterial endocarditis and shunt nephritis caused by direct bacteremia due to bacterial colonization of the shunt lumen. The change from ventriculoatrial to ventriculoperitoneal shunts substantially decreased the complications of shunting.

In children with hydrocephalus, detailed neuropsychological testing has revealed that performance IQ is poorer than verbal IQ. Repeated shunt infections have been associated with poor outcome.

Some children with myelomeningocele develop intellectual impairment.[22, 23] This may be related to hydrocephalus and shunt infection and malfunction. Approximately 80% of children with myelomeningocele have intelligence within the reference range but have specific learning disabilities, including difficulty with pragmatic communication, short-term memory, executive functioning, and reading comprehension.

These children often have what is described as a “cocktail personality” and can be loquacious without any substantial context to their conversations. They also tend to be weaker in the area of perceptual and motor skills. Children with myelomeningocele without hydrocephalus typically have a normal intelligence.

Overdrainage-related complications

Overdrainage of CSF is another significant shunt complication that is difficult to counteract. Early overdrainage leads to the formation of SDHs, which are difficult to treat, and ligation of the shunt is sometimes necessary. Late chronic overdrainage leads to the development of slit ventricles and mostly affects patients with differential pressure valves, which drain excess CSF when the patients assume the upright position, because of the siphoning effect of the column of fluid in the distal tube. Chronic overdrainage leads to collapse of the ventricles and intermittent shunt obstruction.

Siphoning is due to a pressure gradient between the proximal and distal ends and is a factor of the height multiplied by the mass of CSF multiplied by the acceleration of gravity. Siphoning usually becomes apparent when patients come out of recumbency. Antisiphon devices of different types have been developed to overcome overdrainage and are incorporated in many shunt systems, with variable success. Over the years, technical aspects of the shunt valves and the development of flow-regulating valves have improved the frequency of adverse effects related to overdrainage and mechanical complications.

Secondary craniosynostosis can develop as a consequence of chronic shunt overdrainage. Children with milder forms who have had differential pressure valves for years develop a thick skull (ie, hyperostosis cranii ex vacuo). Secondary craniosynostosis is more common in children with myelomeningocele, and the presence of myelomeningocele is believed to result in a state of reduced CSF content in the entire neuraxis, leading to a reduced drive for brain development and early suture closure.

Additional concerns

Some adolescents with myelomeningocele and shunted hydrocephalus develop focal discomfort at the shunt valve or along the distal catheter in the posterior triangle of the neck. This had been termed shuntalgia and is characterized by tenderness with a palpable, firm, fibrotic sheath of scar tissue in the area of pain. This may be related to the adolescent growth spurt, and although it is resistant to nonsteroidal anti-inflammatory drugs (NSAIDs), it is usually self-limiting.

Endoscopic third ventriculostomy

Endoscopic treatment of hydrocephalus carries a risk of complications similar to those of intraoperative and immediate postoperative shunt insertion (ie, a 10% risk of infection or hemorrhage, basilar artery injury, and hypothalamic or pituitary dysfunction) but does not carry the long-term problems and complications of shunts and is not associated with overdrainage.[24]  However, the rare complication of late rapid deterioration can occur up to 7 years after the procedure and is not well understood, though it is thought to be a result of stoma scarring. This can result in coma and possible death due to obstructive hydrocephalus.

Leaving an Ommaya reservoir in place can be helpful, especially in patients who are in areas where access to neurosurgical care is limited. Tapping the Ommaya reservoir and aspirating CSF can serve as a bridge to definitive treatment. An element of underdrainage is present even in successful cases because the ventricles remain larger than they are in shunted patients. The mortality of ETV is approximately 1%.

Performing a CT scan and shunt series before discharge is customary to verify the position of the ventricular tube and to serve as future reference in case of possible shunt malfunction. Postoperative follow-up a few weeks after shunt placement is usually necessary to ensure that the wound is healing well and that the head circumference is decreasing accordingly.

The issue of repeat scanning in the months after shunt insertion or ventriculostomy remains controversial. Certainly, satisfactory shunt function should be verified with at least one scan during the first year. In patients who undergo ETV, cine phase-contrast MRI is mandatory to verify the patency of the third ventricolostomy. Some surgeons customarily perform yearly scanning, but the use of such routines is not universally accepted.

Currently, shunt malfunctions cannot be detected before they manifest clinically. Attempts to visualize CSF flow by using US or other imaging techniques have not been met with universal acceptance, because they are associated with some false-negative and false-positive results.

Patients with spina bifida require ongoing follow-up for life because problems unrelated to hydrocephalus appear at various stages and necessitate treatment. These patients have continuing urologic problems due to neuropathic bladder. Neurogenic bowel can cause debilitating social problems. Although neurogenic bowel and bladder are initially managed medically with medication, retrograde scheduled enemas, and routine CIC, some patients and physicians opt for surgical treatment, such as the Malone procedure for antegrade colonic enemas (MACE) and the Mitrofanoff procedure for ease of catheterization.

Retethering of the cord occurs in 15-20% patients with myelomeningocele and is characterized by progressive weakness of one or both legs, onset or progression of scoliosis, change in gait, change in bowel or bladder control or function, or lower back pain. This condition may be confused with shunt malfunction.

Orthopedic problems, such as scoliosis and foot deformities, also warrant careful follow-up, because they are likely to necessitate surgical treatment. A consideration particular to children with myelomeningocele is the relationship between hydrocephalus and scoliosis, which is present in a very high proportion of these patients. Scoliosis deteriorates in the presence of untreated hydrocephalus and improves following successful shunting. Active hydrocephalus is postulated to exacerbate the compressive effect of the hindbrain hernia on the descending pathways at the craniovertebral junction, inducing neuromuscular imbalance.