Introduction

Chiari malformation is a pathologic entity first described at the end of the nineteenth century by Hans Chiari [1], who was studying the pathology of babies dying with hydrocephalus. He described four different types (Chiari I–IV), with some authors also using an intermediate type Chiari 1.5 between Chiari I and Chiari II [2]. The Chiari malformation has an unfortunate name that often leads to confusion and controversy. The utilization of the term “malformation” usually refers to a pathologic entity with a single definable cause, which begins expression before birth, and leads to predictable effects on the clinical presentation of the patients and the pathology of the patient’s brain. The development of magnetic resonance imaging (MRI) has led to an improved understanding of cerebellar herniation. The original studies of Hans Chiari dealt with four types of cerebellar malformation in these autopsy studies. These four types represent completely different pathologies. The Chiari III malformation (occipito-cervical encephalocele) and Chiari IV malformation (cerebellar agenesis) are extremely rare and are unlikely to be seen in the adult population. On the other hand, Chiari I (chronic tonsilar herniation, which has been called the “adult Chiari malformation”) and Chiari II (essentially always associated with spina bifida) are now frequently cared for in adult settings and are the subject of this review.

What is the Chiari I malformation?

Chiari I or adult Chiari malformation is an unfortunate term as it presupposes that it is a single entity with a single cause and that it is developmental, beginning before the skull and brain are formed. Neither of these conditions is correct in a large percentage of cases. The use of terms such as hindbrain hernia or Chiari anomaly [3,4] are more appropriate but the use of the term Chiari malformation is so engrained in the parlance of neurosurgery and the literature that it is something we have to live with. Those who study and care for patients with hindbrain herniation should develop a consensus on the definition and classification of the problem to assure that we are all speaking about the same entity. In general, Chiari I malformation is a crowding of the posterior fossa with tissue primarily derived from the inferior-most part of the cerebellar hemispheres, generally called the cerebellar tonsils. The patients share an absence of the cisterna magna and by neuroradiological convention the tips of the inferior cerebellar hemispheres project below the foramen magnum by at least 5 mm.

Hindbrain herniation may occur from any of four pathogenetic mechanisms. These mechanisms are based on the physical principles of the Monro–Kellie hypothesis and are:

1. 
   

   (1) The skull, particularly in the posterior fossa, is too small to accommodate the brain contents, which causes the herniation through the foramen magnum. Since the squamous portion of the occipital bone is made of “membranous bone” this development must occur quite early in development. This mechanism is most convincingly expressed in the case of Crouzon’s and Pfeiffer’s syndromes and is most dramatic in patients with Kleebaatschädl (cloverleaf skull).

   
   
2. 
   

   (2) There is a pressure differential across the foramen magnum caused by increased pressure intracranially. This can be due to a mass within the skull or to too much fluid intracranially. In hydrocephalus this is increased cerebrospinal fluid (CSF) within the ventricles. In pseudotumor cerebri it is a combination of too much CSF in the cortical subarachnoid space and too much venous blood in the substance of the brain and in the dural venous sinuses.

   
   
3. 
   

   (3) There is a pressure gradient across the foramen magnum due to low pressure in the spinal subarachnoid space. This can be due to spontaneous intracranial hypotension such as from a leaking perineural cyst or from overdrainage of a non-valved lumboperitoneal shunt.

   
   
4. 
   

   (4) Finally, there is traction on the distal end of the spinal cord from severe tethering due to lipoma, thickened filum terminale, or myelomeningocele that had been previously repaired.

Hydrocephalus and Chiari I: which came first?

Based on the new consensus classification of hydrocephalus, hindbrain herniation can cause hydrocephalus in two ways. It can either block the outlet foramina of the fourth ventricle if both the foramen of Magendie and the foramina of Luschka are occluded. This would lead to severe dilatation of the fourth ventricle and would be clearly shown to be an obstructive cause of hydrocephalus using the classic classification of Dandy that a dye placed into the lateral ventricle would not be able to be recovered from the spinal subarachnoid space. The second possibility would be that the CSF that did exit the foramina into the spinal subarachnoid space would be prevented from flowing from the spinal subarachnoid space into the cortical subarachnoid space, which Dandy would classify as communicating even though there is a very clear point of obstruction.

Both of these processes are quite rare, particularly in adults. Because there is a failure of communication between the lateral ventricles and the cortical subarachnoid spaces, both of these mechanisms should respond to endoscopic third ventriculostomy (ETV).

By its very nature, hydrocephalus adds volume to an already closed intracranial skull volume and therefore may be the cause of the hydrocephalus itself. In infants it is often very difficult to discern which of the conditions is the cause and which is the result. In cases in which the question is unanswered or unanswerable performing a Chiari decompression could lead to a refractory pseudomeningocele. If one plans to do a Chiari I decompression and the patient has hydrocephalus that seems well compensated, we strongly recommend postoperative intracranial pressure monitoring for early diagnosis of decompensation of the compensated hydrocephalus. If the patient exhibits signs of increased intracranial pressure it is wise to treat the hydrocephalus first either with a shunt or with an ETV.

Hydrocephalus and the Chiari II malformation

The Chiari II malformation is a complex developmental disorder associated exclusively with open neural tube defects, most commonly spina bifida aperta but also found with encephaloceles. As demonstrated by the work of McLone [5–7], it is due to the constant leaking of CSF from the intracranial compartment via the open defect causing distortion of brain development and, at least transiently, a state of microcephaly during intrauterine development. Figure 25.1b is an artist’s representation of the major findings in Chiari II malformation including a small posterior fossa, vertical and incompetent tentorium, tectal beaking, and herniation of the brainstem and cerebellar vermis into the cervical canal. This deranged anatomy causes hydrocephalus in four different ways and the patient may have any one of the specific points of obstruction or may have two, three, or all four. In the mildest forms of the condition, the manifestations of the Chiari II malformation may be so mild as to be difficult to detect. The clinical presentation and to some extent the treatment options differ depending on the point of obstruction. The specific types of hydrocephalus that depend on a specific point of obstruction are described below. They include:

1. 
   

   (1) Closure of the aqueduct of Sylvius secondary to upward herniation of the top of the cerebellar vermis. These patients are generally found in utero by ultrasound to have severe ventriculomegaly that progresses fairly rapidly in the last trimester. Delivery here when recognized is always by cesarean section due to cephalopelvic disproportion. These babies should be delivered before the thickness of the brain at the foramen of Monro (cerebral mantle) falls below one centimeter, which may be a point of no-return implying that the brain cannot reconstitute after shunting and the babies would have severe developmental delays that could have been avoided by early intervention. At the time of shunt failure these children always will be symptomatic and show increase in ventricular size if they have ever gotten smaller.

   
   
2. 
   

   (2) Occlusion of the outlet foramina of the fourth ventricle due to the herniation of the entire fourth ventricle into the cervical spine. In this case the entire brainstem is distorted, but because the anatomy of the fourth ventricle is determined by the narrow space in which it is confined it may not be appear to be distended. These patients will frequently present with overt brainstem dysfunction at the time of shunt failure. They will be found to have a fourth ventricle that is elongated into the cervical spine, which it may not be possible to visualize on MRI when the shunt is working.

   
   
3. 
   

   (3) The “stopper in the bottle” phenomenon. If the CSF can exit through the outlet foramina of the fourth ventricle, that fluid can mix with the remainder of the CSF in the cortical subarachnoid space but cannot get back into the intracranial compartment to access the dural venous sinuses in order to be terminally absorbed. This type of hydrocephalus typically is present early in postnatal life and progresses slowly if at all. It may be observed without shunting for a few months to make certain that the hydrocephalus has become compensated and the patient does not need a shunt. That this compensation can occur can usually be ascertained within the first half year of life [8].

   
   
4. 
   

   (4) The final point of obstruction relates to the position of the torcula within the cervical spine seen in a significant number of patients with spina bifida. This is due to the small volume of the posterior fossa and the vertical position of the tentorium. Hydrocephalus develops somewhat slowly in these cases and it is due to the fact that the head is still distensible and the absorption of CSF requires a pressure differential of at least 5–7 mmHg in order for CSF to be absorbed. Once a shunt is inserted the ventricles decrease in size and the skull becomes a fixed volume. When the shunt fails in such circumstances the skull cannot distend and therefore the ventricles do not enlarge. The patient now exhibits a severe form of the “slit ventricle syndrome” (SVS). This is seen in about 25% of patients with spina bifida shunted in infancy. Management here is quite difficult. One of the authors has treated such a patient in whom intracranial pressure (ICP) was measured to be over 80 mmHg with no evidence of ventricular distension [9].

Figure 25.1 (a) Stopper in the bottle analogy for hindbrain herniation. (b) Artist’s rendition of a sagittal view of Chiari II malformation showing the overall anatomy of the posterior fossa and upper cervical spine. The four potential points of obstruction causing hydrocephalus are demonstrated as the secondary closure of the aqueduct of Sylvius, closure of the outlet foramina of the fourth ventricle, impaction of the brainstem causing poor flow from the spinal subarachnoid space and the cortical subarachnoid space, and compression of the torcula causing high venous pressure in the dural venous sinuses.

Shunt dependency in this population is a severe burden. The published mortality for shunt-dependent spina bifida patients has been reported to be 1%/year into adulthood. Shunt infection in this population if it occurs within the first 6 months of life tends to result in a 25 point drop in predicted IQ.

The recent publication of the results of the “Management of Myelomeningocle Study” has led to new forms of treatment and exciting new potential for more normal lives for the affected individuals [10]. The data are quite convincing that if the neural tube defect is repaired in the 24th to 26th week of gestational life and the baby can be carried to term or near term, the radiographic appearance of the Chiari II malformation is much improved and the incidence of hydrocephalus that needs to be shunted falls to a minority of patients.

The Chiari I malformation and benign intracranial hypertension or pseudotumor cerebri

About 10% of patients seen at the Chiari Institute will be found to have pseudotumor cerebri. Sometimes this diagnosis is made prior to intervention when papilledema is found on fundoscopic examination. Usually, however, it is found when the patient develops recurrent pseudomeningocele postoperatively. Lumbar puncture or formal ICP monitoring shows them to have raised ICP.

The most likely situation in which Chiari I malformation and pseudotumor will coexist occurs in the context of morbid obesity. Morbid obesity results in increased intracranial pressure, especially in women due to the effect on the pressure in the right atrium. Acetazolamide in this situation acts as a diuretic and lowers intracranial pressure. High right atrial pressures also lead to high sagittal sinus pressures, an increase in the venous volume of the brain and therefore the total volume of the brain and in some cases may actually be the cause of the Chiari I malformation. Operating on the craniovertebral junction in morbidly obese patients is both difficult and dangerous. Bariatric surgery carries the highest likelihood of a good outcome in patients with Chiari who are morbidly obese and has been shown to be effective treatment of the increased ICP [11,12]. We strongly recommend bariatric surgery for all patients with pseudotumor, with or without Chiari I malformation, who have a body mass index of 35 or greater.

In non-obese patients pseudotumor is often overdiagnosed and while shunting may help to treat a pesky pseudomeningocele every effort should be made to get rid of the shunt as soon as possible. Often after the pseudomeningocele is healed it is possible to remove the shunt and place an external ventricular drain and record the ICP. This frequently shows that the CSF absorptive deficit either no longer exists or was not there in the first place.

In thin patients who have documented pseudotumor and a Chiari I malformation the management requires some thought. Treatment with ventriculoperitoneal shunts often leads to recurrent failure of the ventricular catheter. The absorptive deficit is in the cortical subarachnoid space and it is difficult for the CSF to flow retrograde from the cortical subarachnoid space to the ventricle to get out of the intracranial compartment via the shunt. The ventricle will frequently collapse around the catheter. There are three potential strategies that have been shown to be effective in this situation:

1. 
   

   (1) Decompression surgery showing a reconstituted cisterna magna will allow the management of this condition with a lumboperitoneal shunt with good success.

   
   
2. 
   

   (2) If a ventricular shunt is felt to be needed it is likely that a second catheter in the cervical subarachnoid will be needed to balance the pressure on the inside and outside of the brain and make certain that the ventricle does not collapse around the ventricular catheter.

   
   
3. 
   

   (3) Finally, all pseudotumor is thought to be due to increased pressure in the superior sagittal sinus. If a point of restriction to flow is found, recent studies have shown the usefulness of venous stenting in the treatment of pseudotumor without shunting [11,13,14].