Child's Nervous System

, Volume 20, Issue 5, pp 323–328

History, anatomic forms, and pathogenesis of Chiari I malformations


    • Section of NeurosurgeryHospital C. Durand
    • Division of NeurosurgeryUniversity of Buenos Aires School of Medicine
Focus Session

DOI: 10.1007/s00381-003-0878-y

Cite this article as:
Schijman, E. Childs Nerv Syst (2004) 20: 323. doi:10.1007/s00381-003-0878-y



Chiari I malformations constitute a group of entities of congenital or acquired etiology that have in common descent of the cerebellar tonsils into the cervical spinal canal. In recent years, since the advent of magnetic resonance imaging, an increasing number of asymptomatic, doubtfully symptomatic, and minimally symptomatic patients with Chiari I malformations have been diagnosed. This has resulted in controversy about the multiple therapeutic strategies indicated for these problems.


With the intention of updating the knowledge that we currently have on the Chiari I malformation and the related syringomyelia we review the literature and discuss the historical background, pathogenesis, anatomic forms, clinical presentation, and diagnostic procedures of these abnormalities.


Chiari I malformationForamen magnum decompressionHydromyeliaPosterior fossa decompressionSyringomyelia


Chiari malformations constitute a group of entities with very different etiology, pathophysiology, and clinic features that have in common the presence of anatomical deformities of the brainstem and cerebellum. They are part of the group of cranio-spinal malformations. Among them, the osseous malformations were the first to be recognized: the atlas assimilation was described by Colombo (mentioned in [26]) in 1572 and the basilar impression by Ackermann in 1790 [4].

In 1883, Cleland was the first to describe what was later called the Chiari II or Arnold-Chiari malformation (mentioned in [62]), with his report of a child with spina bifida, hydrocephalus, and anatomical alterations of the cerebellum and brainstem. The term “Chiari malformations” has been used in recognition of the work of Hans Chiari, a Viennese pathologist, who practiced medicine in Vienna, Prague, and Strasbourg. In his first publication in 1891, Chiari described the case of a 17-year-old woman with “elongation of the tonsils and medial divisions of the inferior lobules of the cerebellum into cone shaped projections which accompany the medulla oblongata into the spinal canal” [20, 22]. Chiari also mentioned that the structure of the tonsils may be normal, soft or fibrotic, and the tonsils may extend down to the level of the atlas or to the axis. Soon after the original description by Chiari, in 1892, the first case of syringomyelia treated by surgery was reported [1, 17, 60]. In 1894, Arnold published a case of a child with spina bifida and elongation and descent of the inferior part of the cerebellum in the spinal canal [8]. In his second presentation on this subject, Chiari reported 14 cases with posterior fossa alterations, describing a descent of the inferior vermis, pons and medulla, and fourth ventricle elongation into the spinal canal. He pointed out that Cleland and Arnold had previously described these alterations [21]. Unlike his previous description in which he considered these alterations to be secondary to hydrocephalus, in this presentation Chiari stated that “other factors must play a role in this condition,” namely insufficient skull growth resulting in increased intracranial pressure (ICP). He also remarked that the degree of tonsillar descent could be directly related to the age of onset, length, and severity of the hydrocephalus, and considered that perhaps hydrocephalus was more precocious and the tonsils and hindbrain descent more important in the forms today known as type II than in that called type I malformation. In 1901, Homén reported the neurological manifestations that result from the osseous malformations in the cranio-cervical region, mainly after basilar impression [40]. In 1907, Schwalbe and Gredig, pupils of Arnold, described four cases with meningomyelocele and alterations in the brainstem and cerebellum, and gave the name “Arnold-Chiari” to the association of these malformations [73]. In 1932, Van Houweninge Graftdijk was the first to report the surgical treatment of Chiari malformations (mentioned in [69]). He wrote a monograph entitled Over Hydrocephalus, presented as a thesis for his doctorate in medicine at the University of Leyden, in which he described the resection of the occipital bone, the incision of the dura mater in cases of Chiari malformation and also the resection of the tonsils when they are redundant, with the intention of relieving the CSF flow obstruction at the foramen magnum. All his patients died as a consequence of the surgery or after postoperative complications.

In 1934, Ebenius reported a case of basilar impression treated with a suboccipital decompression [28]. In 1935 Russell and Donald described ten cases with posterior fossa abnormalities in children with meningomyelocele, naming them Arnold-Chiari malformations [69]. They suggested that hydrocephalus might be secondary to the Arnold-Chiari malformation and postulated that decompressing the spinal cord at the foramen magnum might facilitate the CSF circulation, pointing out that such an operation “has not yet been carried out.” Nevertheless, as an addendum, they mention that while their paper was in press, they became aware of the above-mentioned monograph by Van Houweninge Gaftdijk in 1938 [63]. Also in 1938, McConnell and Parker reported five cases of surgically treated hydrocephalus in adults [49]. All the patients died after surgery. Autopsy showed a tonsils descent through the foramen magnum in three of them. In the same year, Aring reported the case of a 20-year-old boy with signs of increased ICP [6]. On posterior fossa exploration a tonsils descent to the axis level was noticed. The patient died soon after surgery. In 1940, Gustafson and Oldberg described two patients with an intraoperative diagnosis of Chiari malformation. In one of them, the autopsy showed that besides Chiari a basilar impression and syringomyelia were present [38]. Walsh et al. described in 1941 a case of Chiari malformation operated on under the diagnosis of basilar invagination [79]. Also in 1941, List reported two cases as being “Arnold-Chiari malformation” suspected preoperatively by ventriculography in the first and by myelography in the second case [47]. The same year, Adams et al. reported a case with a preoperative diagnosis of spinal tumor with myelographic blockage at the level of C3 [5]. A Chiari type I malformation with the tonsils at C3 was found during surgery. These authors were the first to describe in detail the clinic picture of the Chiari type I syndrome, classifying it into five forms of clinical presentation: increased ICP, compromise of cranial nerves, brainstem compression, spinal cord compression, and cerebellar signs. In 1948 Chorobski and Stephen operated on a patient with basilar impression, observing during surgery anatomical alterations of the cerebellum that the authors “recognized” as those described by Chiari, and indicating that the association between both entities must be suspected [23]. In 1949, Russell was the first to bring some order to the nomenclature when she suggested that the Arnold-Chiari term should only be used in patients with spina bifida [68]. Later on, Baker described 11 cases as Arnold-Chiari malformations with tonsils descent to the C1 level as seen by myelography [11]. In addition, he remarked that the normal level of the cerebellar tonsils is when the tips are above a line between the clivus tip and the posterior rim of the foramen magnum on a lateral X-ray. Three cases were operated on, describing the rest of them as “mild forms of Arnold-Chiari malformation,” which was the reason why they did not require surgery. Adult patients with Chiari malformations were reported sporadically [31, 35]. The confusion regarding the nomenclature persisted until not long ago; even in 1971 it was suggested to reserve the “Chiari malformation” name only for those cases that take place during infancy and associated with spina bifida and hydrocephalus [26]. In 1974, Bloch et al. described the tonsils position using myelography [15]. According to the authors, under normal conditions, they are between 7 mm above and 8 mm below the upper tip of the foramen magnum, and in cases with Chiari malformations between 3 mm above and 25 mm below the foramen. They concluded that “tonsillar herniation can be seen as an incidental finding.” The advent of MRI has been an important factor for the diagnosis of Chiari malformations and related syringomyelia, which have since been recognized with increasing frequency. In 1985, Aboulezz et al. established by MRI that in normal conditions the tonsils tip can extend up to 3 mm below the foramen magnum and in patients with Chiari they can exceed 5 mm [2]. For Barkovich et al. the limit in normal cases is 5 mm below the foramen magnum, describing in Chiari cases the “peglike aspect” of the tonsils and narrowing or complete effacement of the subarachnoid space at the foramen magnum [13]. The evaluation of the CSF flow dynamics at the posterior fossa and foramen magnum by cine-flow MRI [7, 59, 76] has contributed to the understanding of the pathophysiology of the Chiari I malformation as we know it today.

Anatomic forms of Chiari malformations

Four types of Chiari malformations have been classically described.

Chiari I

There is variable descent of the tonsils into the upper cervical canal. Hydrocephalus is infrequent and of constrictive type. Usually, it is a disorder of mesodermal origin but neuroectodermal and acquired forms have been reported.

Chiari II

There is descent of the inferior vermis and cerebellar hemispheres through the foramen magnum with a displacement of the brainstem (medulla, fourth ventricle, and lower portion of the pons) inside the spinal canal and aqueduct and fourth ventricle elongation. This is the most frequent form of Chiari malformation, also named Arnold-Chiari malformation. It is associated with spina bifida and other cerebral, spinal and meningeal abnormalities. Hydrocephalus is present in more than 70% of the cases and is of constrictive or obstructive (aqueduct stenosis) type [66]. Its origin is a neuroectodermal disturbance.

Chiari III

It is characterized by a caudal displacement of the medulla and herniation of part of the cerebellum in an occipital, cervical or occipital-cervical meningocele. Sometimes part of the hindbrain is also herniated. Raimondi extended the Chiari III concept to the vertex encephaloceles with the osseous defect at the sagittal suture [65]. Hydrocephalus is present in 50% of the cases and is of obstructive type due to aqueductal stenosis or to an associated Dandy-Walker malformation [66]. It is a neuroectodermal malformation.

Chiari IV

It is the least frequent form of Chiari malformation, characterized by a hypoplasia or aplasia of the cerebellar hemispheres and alterations of the pons with a “pigeon breast deformity” of the brainstem. These anatomic alterations provoke a marked dilatation of the fourth ventricle, cisterna magna, and basal cisterns, even though hydrocephalus is infrequent [66]. It is a disorder of neuroectodermal origin.

Other forms

Recently, anatomic forms different from the classical Chiari malformation have been described with a clinical picture similar to the Chiari I. The Chiari 0 malformation is an alteration characterized by some degree of posterior tilt of the pons and medulla with caudal displacement of the medulla oblongata, low tip of the obex and normal position of the tonsils [41, 42, 77, 80]. Occasionally, a caudal migration of the brainstem and cerebellar tonsils ectopia have been observed, as in the Chiari II malformation, but with the absence of spina bifida. It has been called Chiari 1.5 malformation [41, 80, 81]. These Chiari malformations show, as in the Chiari type I and type II malformations, “crowding” of the foramen magnum.

Pathogenesis of Chiari I malformation

The Chiari I malformation constitutes an heterogeneous and multifactorial entity, in which congenital forms of isolated presentation or with a genetic background and forms of acquired etiology exist. All of them have in common a “crowded” foramen magnum and some present, in addition, a “crowded” posterior cranial fossa, which can be of normal or small size.

While the Chiari type II, III, and IV malformations are considered as primary neural anomalies (neuro-ectodermal anomalies), most frequent types of Chiari I seem to be the result of a mesodermal defect [51]. It has been pointed out that a possible para-axial mesoderm insufficiency after the closure of the neural folds could lead to underdevelopment of the basichondrocranium resulting a posterior fossa that is too small and shallow [48, 56]. The consequence would be overcrowding and a cerebellum progressive deformation leading to a tonsillar herniation below the foramen magnum [10, 55, 64, 70, 74].

A small posterior fossa and a short clivus have been observed in two-thirds of Chiari I cases and the smaller the posterior fossa, the more important the tonsillar descent [57, 72, 74, 78]. In a morphologic study of the posterior fossa done by MRI in patients with Chiari I, a reduced height of the supraocciput, clivus hypoplasia, an increased slope of the tentorium, and a median volume diminution of the CSF in the posterior cranial fossa of 40% were observed [55]. Syringomyelia is found in 30–85% of patients with Chiari I malformation [14, 52, 55]. For a long time, its presence was explained by Gardner’s hydrodynamic theory, according to which there would exist an incomplete embryonic opening of the outlets of the fourth ventricle and a forceful CSF diversion from the fourth ventricle into the central canal of the cord [34]. These forms of “communicating” syringomyelia, in which the syrinx communicates with the fourth ventricle, are often seen in hydrocephalus cases but not in patients with Chiari malformation [54]. On the contrary, in these cases the syrinx usually forms at a considerable distance from the fourth ventricle, from which it is separated by a large segment of normal cord [33]. The cine-mode MRI studies done in these patients do not usually show CSF flow from the obex into the central canal of the cord [7, 46, 59].

Williams has postulated as pathogenesis of the syrinx a craniospinal pressure dissociation provoked by the foramen magnum obstruction [82]. During systole, there are increases in cerebral volume and cerebral blood volume; under normal conditions this intracranial volume increase is “compensated” by the rapid exit of blood from the cranial cavity to the venous system and by the exit of CSF from the ventricles to the spinal subarachnoid space accompanied by a descent of the posterior fossa structures. During diastole, on the other hand, CSF passes from the spinal canal to the intracranial subarachnoid space. In patients with Chiari I malformation the herniated tonsils produce a CSF flow blockage at the foramen magnum and constrain the subarachnoid spaces of the posterior cranial fossa, especially posterior and lateral to the cerebellum. This makes the CSF passage to the spinal canal difficult during systole and it results in a cranio-spinal pressure dissociation. As a consequence, the intracranial pressure content remains elevated during each cardiac cycle due to a decrease in CSF compliance and the lost of the damping effect that in normal conditions allow compensation, by blood venous volume and CSF pressure changes, of the intracranial pressure oscillations that take place during breathing, cardiac cycle, positions changes, and physiologic activities that involve a Valsalva maneuver [55]. Frequently, some Chiari I symptoms, such as suboccipital headaches and dizziness, increase during physical activity and Valsalva maneuvers. The lack of CSF compliance that increases the intracranial pressure in each systole stresses the impact of the tonsils in the foramen magnum, acting as a piston on the “entrapped” CSF at the spinal subarachnoid space and driving it into the Virchow-Robin perivascular and interstitial spaces and then into the central canal of the cord [12, 55, 58, 59]. The central canal, usually obliterated, would stay “patent” at some level at which the CSF would accumulate [53, 58, 75]. There would initially be a spinal cord edema, like a “pre-syrinx state,” and once this extra-cellular fluid exceeds the resorptive capacity of the tissue it would form the syrinx [32, 43, 45, 58]. According to Aboulker, the posterior rootlets vicinity would be the CSF entry point into the cord [3]. A small posterior fossa is also present in patients with Chiari I malformation associated with certain forms of craniosynostosis. Syndromic—Crouzon and Apert syndrome—and nonsyndromic craniosynostosis, especially those cases with a precocious closure of the lambdoids and posterior part of the sagittal sutures are accompanied, on some occasions, by a Chiari I malformation due to a tight posterior fossa [50]. The latter, associated with a frequently present sigmoid sinus obstruction, leads to a tonsillar descent into the spinal canal. Some congenital forms of Chiari I are genetically transmitted. They have been reported in monozygotic twins and triplets, in cases of familiar aggregation and conditions such as hypophosphate rickets, acromegaly, achondroplasia, autosomal dominant spondyloepiphysial dysphasia tarda, and the Hedju-Cheney and Klippel-Feil syndromes [19, 55, 74]. Chiari I malformation is occasionally associated with neurological disorders such as mental retardation, developmental delay in motor or language function, and epilepsy. The etiology of these forms of Chiari I malformation seems to be different from the classical mesodermal disorder, more akin to the Chiari II, III, and IV with a neurogenic etiology and where the tonsillar herniation represents only one of several cerebral dysgenetic abnormalities [16, 29, 30, 37].

On occasion, Chiari I presents with a “crowded,” normal size posterior fossa [72, 78]. Vascular malformations, mainly large dural or vein of Galen fistulae that provoke an important intracranial venous hypertension, are conducive secondarily to a hydro-venous congestion with a progressive amygdaline descent—what has been described as “pseudo-Chiari” malformation [36]. In three cases of vein of Galen fistulae in Girard et al.’s series, MRI during the neonatal period did not show any amygdaline descent and it was observed in control studies done months later. The tonsil herniation gets better or disappears with early endovascular treatment of the fistula [27, 36]. Some cases of Chiari I with a “crowded” but normal size posterior fossa could be a form of presentation of a benign intracranial hypertension syndrome—pseudotumor cerebri—a condition with CSF absorptive problems secondary, according to Rekate, to increased intracranial venous pressure (Rekate HL, McDougal C et al Pseudotumor cerebri: management should be directed by retrograde venography with pressure recording. Presented at ASPN, January 27-February 1 2002, Nevis, West Indies; [67]). Acquired forms of Chiari malformations are often seen after lumbar shunting procedures such as the lumbo-peritoneal shunt indicated for the treatment of certain forms of communicating hydrocephalus and pseudotumor cerebri, and after external lumbar drainage or serial spinal taps for the treatment of postoperative or post-traumatic CSF leakage [9, 24, 44, 71]. In these cases there is a normal size posterior fossa with a “crowded” foramen magnum. The procedures mentioned would cause a cranio-spinal pressure gradient through the foramen magnum with a downward displacement of the cerebellar tonsils. Payner et al. point out that it would be a similar mechanism to the one proposed by Cameron in order to explain the Chiari II malformation secondary to CSF leakage in patients with myelodysplasia [18, 61]. The clinical picture in these Chiari I cases is variable, from chronic forms without clinical manifestation in which the CSF siphoning effect is conducive to an arrested skull growth with secondary suture stenosis and tonsil herniation when the volume of cerebral growth exceeds the available space, to those cases that present with an acute foramen magnum syndrome with brainstem and upper spinal cord compression [25, 39].

Copyright information

© Springer-Verlag 2004