Thomas H. Milhorat, M.D.
Paolo A. Bolognese, M.D.
Department of Neurosurgery, The Chiari Institute, North Shore – Long Island
Jewish Health System, Manhasset, New York
AUTHOR INFORMATION
Reprint Requests:
Thomas H. Milhorat, M.D.
Professor and Chairman
Department of Neurosurgery
Director, The Chiari Institute
North Shore - LIJ Health System
300 Community Drive
Manhasset, New York 11030
Telephone: (516) 562-3020
Facsimilie: (516) 562-3030
E-Mail Milhorat@nshs.edu
ABSTRACT
OBJECTIVE: We describe an operative technique for Chiari I malformation that employs color Doppler ultrasonography (CDU) as a guide for performing patient-specific posterior fossa decompressions. The technique has been used since 1999 in more than 300 operations.
METHODS: Based on real time anatomical and physiological measurements, the following goals of surgery were monitored: (1) adequate decompression of the cervicomedullary junction; (2) creation of retrocerebellar space of 8-10 cc volume; and (3) establishment of optimal CSF flow between the cranial and spinal compartments.
RESULTS: The size of the craniectomy was tailored to conform to the area of cerebellar impaction as demarcated by compressed subarachnoid spaces. A laminectomy was not performed unless the cerebellar tonsils were herniated below C1. Prior to opening dura, CDU imaging was invaluable in planning operative strategies. A simple duraplasty without additional steps was found to be appropriate treatment in occasional patients with minimal tonsillar herniation (5-8 mm). In all other cases, it was necessary to perform an internal decompression that included lysis of the arachnoid and shrinkage of the cerebellar tonsils to achieve the goals of surgery. Optimal CSF flow through the foramen magnum in anesthetized prone patients was found to have the following characteristics: a peak velocity
of 3-5 cm per second, bidirectional movement, and a wave form exhibiting vascular and respiratory variations. The attainment of surgical goals was confirmed in most patients by postoperative neuroimaging.
CONCLUSION: CDU imaging is an important technological advance that permits the
neurosurgeon to tailor the steps of Chiari surgery according to patient-specific variables.
The success of this technique depends upon the mastery of a new and sophisticated monitor-
ing modality.
KEY WORDS: Chiari I malformation; Posterior fossa decompression; Intraoperative ultrasound imaging; Color Doppler imaging; Cerebrospinal fluid flow.
INTRODUCTION
The surgical treatment of Chiari I malformation (CMI) has not been standardized. Of the available operative techniques, the most widely performed is a posterior fossa decom-presssion consisting of a suboccipital craniectomy, C1 laminectomy, and duraplasty (1, 4, 9, 15). Controversy exists concerning the extent of the bony decompression and the need for additional steps such as shrinkage or resection of the cerebellar tonsils. While experienced neurosurgeons tend to use a fixed operative technique, this approach has limitations given the great variability of patient-specific findings.
Since 1999, the authors have used intraoperative color Doppler ultrasonography (CDU) as a guide to posterior fossa decompression in patients with CMI. The technology can be adapted to measure cerebrospinal fluid (CSF) flow and provides real time information about neural displacements, vascular anatomy, and CSF circulation at the cervicomedullary junction. Distinct advantages of CDU imaging include complication avoidance and the ability to tailor operative steps according to patient-specific variables. The methodology of CDU and the details of the tailored operative technique are presented.
METHODOLOGY OF CDU IMAGING
Historical Background
Modern techniques of intraoperative ultrasonography are based on B-scan real time imaging. The basic methodology provides real time bidimensional anatomical images and has been used in neurosurgery since the early 1980’s for the intraoperative localization of a wide variety of intracerebral and intramedullary lesions (3). First generation ultrasound machines were easy to use and required limited training for competence in data acquisition. With advances in machine design, including the introduction of electronic probes, it was possible to complement anatomical imaging with pulse wave Doppler and color Doppler coding. Specific advantages of triplex imaging (CDU) include the high definition of vascular structures and the ability to measure arterial and venous blood flow patterns.
Beginning in 1983, one of the authors (PAB), working in conjunction with Victor A. Fasano at the University of Torino, investigated the uses of intraoperative ultrasonography in more than 1200 neurosurgical procedures. Included in this experience was the in vivo testing of numerous probes, machines, and calibration techniques. The results of these investigations led to the introduction of CDU in aneurysm surgery (2). In recent years, we have investigated the feasibility of using CDU for imaging CSF flow during Chiari surgery.
Adaptations for Measuring CSF Flow
The imaging of CSF flow using color Doppler technology poses technical challenges. In contrast to blood, CSF has the following characteristics: (1) a very low content of cells and proteins which limits the reflection of ultrasound waves; (2) a low velocity circulation with nonhomogeneous flow through a network of irregular spaces rather than vessels; and (3) a significant component of to and fro movement. To address these technical issues, it was necessary to utilize high performance equipment that was capable of maximizing Doppler sensitivity. The results in the current report were obtained using the Acuson Sequoia 512 system (Acuson Corporation, Mountain View, CA) equipped with standard high definition probes.
Because CDU machines are designed to measure blood flow, the steps of data acquisition must be adapted according to the unique features of CSF flow. This can be accomplished by manually adjusting the controls for B-scan real time processing, pulse wave Doppler, and color Doppler. Adjustments of machine default values include recalibration of the following commands: B-scan gain; pulse wave and color Doppler scales; pulse wave Doppler gain; pulse wave and color Doppler filters; time gain compensation; and focusing of B-scan real time. Each of these commands must be fine tuned to achieve optimal images and Doppler information. Experience is required to recognize artifacts
and the aliasing of color Doppler images. The Acuson Corporation has reviewed the measurements of CSF flow in this study and confirm their validity.
CSF Volume Measurements
The Acuson Sequoia 512 system has an inner-built software package for calculating the volume of an area of interest which is obtained by measuring the area in three planes, using two scan orientations perpendicular to each other. Cisterna magna volumes, for example, were calculated by measuring the superior-inferior (length) and anterior-posterior (depth) dimensions in the sagittal plane. The width of the space was measured in the axial plane. Volume calculations used the following formula:
Volume = D1· D2· D3 
6
where D1, D2,D3 are the three measurements of the cisterna magna.
Surgical Data Base
Between November 1999 and September 2002, CDU was used in 315 operations for CMI. All operations were performed by the senior author. There were 152 primary operations and 163 reoperations for failed Chiari surgery. Prior to surgery, patients underwent whole neuraxis magnetic resonance imaging (MRI), CINE-MRI, and 3-D computed tomography (CT) of the head and neck. Optional studies included flexion/extension x-rays of the cervical spine, flexion/extension MRI, 3-D CT angiography, and metrizamide myelography. MRI scans of the cervical spine and CINE-MRI were obtained 3-12 months after surgery.
TAILORED OPERATIVE TECHNIQUE
Surgical Indications
Radiographic criteria. All of the patients in this series had demonstrable impaction of the foramen magnum with MRI evidence of tonsillar herniation, obliteration of the cisterna magna, and reduced or absent CSF flow at the cervicomedullary junction (CINE-MRI). The vast majority of patients had some osseus features of Chiari I malformation (12) such as hypoplasia of the posterior fossa, reduced length of the supraocciput, increased slope of the tentorium, and reduced length of the clivus. Other findings that were present in some patients included anterior displacement of the cerebellum, kinking of the medulla, retroflexion of the odontoid, basilar invagination, scoliosis, Klippel Feil anomaly, and hydrocephalus. Syringomyelia or a pre-syrinx state was present in 182 patients. In 163 patients undergoing reoperation for failed Chiari surgery, there was radiographic evidence of one or more of the following findings: underdecompression of the posterior fossa; overdecompression of the posterior fossa with cerebellar ptosis; pseudomeningocele formation; surgical meningocele; cranial settling; basilar invagination; and hydrocephalus.
Clinical Criteria. The definition of what constitutes a symptomatic Chiari I malformation has been addressed elsewhere (12). In the current series, the indications for operation and were limited to one or more of the following criteria: (1) evidence of progressive clinical deterioration; (2) presence of intractable or disabling symptoms with a Karnofsky Disability Score of 70 or less; and (3) MRI evidence of syrinx propagation. The mean Karnofsky Disability Score in this series was 60 (unable to perform normal activities; requires occasional assistance). The indications for reoperation on patients with failed Chiari surgery were the same as those for patients undergoing primary operations.
Surgical Decision Making
The presence of clinically significant anterior compression of the cervicomedullary junction by a retroflexed or invaginated odontoid was regarded as a contraindication for primary decompression of the posterior fossa. Anterior compressions were divided into two groups by cervical traction. Irreducible compressions were managed by transoral odontoidectomy followed by a craniovertebral fusion at the time of posterior decompression. Reducible compressions were managed by a one-stage posterior decompression and fusion.
Each of the operative steps of posterior fossa decompression was guided by CDU using real time anatomical and physiologic measurements. The goals of surgery were as follows:
(1) adequate decompression of the cervicomedullary junction;
(2) creation of a retrocerebellar space of 8-10cc volume; and
(3) establishment of optimal CSF flow between the cranial and spinal compartments.
External Decompression
Posterior fossa decompressions were performed under somatosensory evoked potential monitoring with the patient in the prone position and the head flexed in a Mayfield headholder. The suboccipital area was exposed through a midline incision that extended from the inion to the second cervical spine. When
a pericranial graft was harvested for duraplasty, the incision was extended 2 fingerbreadths above the inion.
The size of the bony decompression was guided by CDU. After making a small sub-occipital opening, the atlantooccipital membrane was excised to facilitate imaging. Thereafter, the craniectomy was enlarged in a stepwise manner to expose dura overlying the area of cere- bellar impaction, as demarcated by compressed or obliterated subarachnoid spaces. The superior limit of the craniectomy was never above the prepyramidal fissure. Laterally, the craniectomy was widened to create a nearly circular bony opening. The final dimensions of the craniectomy were generally in the range of 3.5 x 3.5 cm.
The size of the bony decompression was guided by CDU. After making a small sub-occipital opening, the atlantooccipital membrane was excised to facilitate imaging. Thereafter, the craniectomy was enlarged in a stepwise manner to expose dura overlying the area of cere- bellar impaction, as demarcated by compressed or obliterated subarachnoid spaces.
The superior limit of the craniectomy was never above the prepyramidal fissure. Laterally,
the craniectomy was widened to create a nearly circular bony opening. The final dimensions
of the craniectomy were generally in the range of 3.5 x 3.5 cm.
Following completion of the craniectomy, a decision was made whether or not to perform a aminectomy. The determining factors were the extent of tonsillar herniation and the required length of the dural incision. CDU was used to establish the true position of the tonsillar tips which was typically 3-6 mm lower than that predicted by MRI. A laminectomy was not performed for herniations above C1. In patients with intermediate herniations (12-15 mm), the superior aspect of the C1 arch was sometimes underbitten (Fig. 1). Herniations of greater than 15 mm generally required a standard C1 laminectomy. Similar guidelines were applied to herniations at lower levels. The relationship of the cerebellar tonsils to bony land-marks was altered by anomalies such as assimilation of the atlas.
|
|
  |
FIGURE 1. CDU images of cervicomedullary junction in 42 year old female with CMI and syringomyelia. A, sagittal image showing 13 mm tonsillar herniation that extends to within 0.38 cm (distance between asterisks) of C1. Trapezoid outlines color Doppler sample volume. B, sagittal image after underbiting superior arch of C1 which increases space between C1 and tonsillar tips to 0.97 cm (distance between asterisks). D, dura; P, posterior inferior cerebellar artery; VA, vertebral artery; T, cerebellar tonsil; SC, spinal cord; tbv, tonsillar blood vessel. Dotted line demarcates shadow artifact of bone.
|
CDU Imaging Prior to Opening Dura
Prior to opening dura, CDU was used for anatomical orientation and to establish baseline measurements
of CSF flow (Fig. 2). Structures that were routinely imaged included the cerebellar tonsils, uvula,
 
|
|
FIGURE 2. CDU images of cervicomedullary junction after external decompression and prior to opening dura. A, sagittal image in 9 year old male with CMI showing 11 mm tonsillar herniation, intact C1, and regional vascular anatomy. B, pulse wave Doppler tracing with sagittal target image (above) in 38 year old male with CMI and syringomyelia showing no cisterna magna and minimal CSF flow caudal to cerebellar tonsils (cursor). D, dura; P, posterior inferior cerebellar artery; VA vertebral artery; T, cerebellar tonsil; SC, spinal cord; av, arachnoid vessel; aa, arachnoid adhesion. Dotted line demarcates shadow artifact of bone.
|
medulla, upper cervical spinal cord, both vertebral arteries, both posterior inferior cerebellar arteries (PICA’s), the marginal sinuses, parenchymal arteries and veins, and bridging vessels suspended by the
arachnoid. The identification of aberrant vascular anatomy, asymmetric herniations, and neural displacements helped to reduce the risk of surgical error (Fig. 3A). In patients undergoing reoperation for failed Chiari surgery, information concerning the location and extent of meningocerebellar scarring was invaluable in planning dissection strategies (Fig. 3B).
|
|
FIGURE 3. A, axial B scan real time image of cervicomedullary junction in 47 year old female with CMI and syringomyelia showing 18 mm asymmetric tonsillar herniation with spinal cord displacement. B, sagittal CDU image of cervicomedullary junction in 3 year old female undergoing reoperation for failed Chiari surgery showing meningocerebellar cicatrix with dura adherent to posterior inferior cerebellar artery and tonsillar branches. D, dura; P, posterior inferior cerebellar artery; T, cerebellar tonsil; tbv, tonsillar blood vessel; SC, spinal cord.
|
CSF circulation at the cervicomedullary junction was assessed immediately prior to opening dura. The following measurements were made and stored: (1) the size and volume of the cisterna magna; (2) the size and volume of the dorsal cervical theca between C1 and the tonsillar tips; (3) CSF velocity/flow in the cisterna magna; (4) CSF velocity/flow in the dorsal cervical theca; and (5) CSF velocity/flow in the premedullary cisterns. Cisterna magna volumes of less than 0.5 cc and CSF flow velocities in the range of 0-0.8 cm per second anterior and posterior to the cervicomedullary junction were typical baseline findings.
Dural Opening
The dura was opened with a “Y” incision across the marginal sinuses unless CDU imaging suggested a better line of entry. The arachnoid was left intact and imaging was repeated in appropriate cases to determine whether duraplasty alone might be sufficient treatment. Opening the dura invariably resulted in some reexpansion of the cisterna magna, but a significant improvement in CSF flow was rarely observed. A simple duraplasty without additional steps was performed in occasional patients who met the following criteria: cisterna -magna volume of at least 4 cc; CSF velocity/flow of at least 2 cm per second; and CSF tracings demonstrating bidirectional movement with vascular and respiratory variations.
Internal Decompression
The arachnoid was opened and CSF was allowed to drain spontaneously. Under mag-nification vision, the arachnoid was resected widely and adhesions to the cerebellar tonsils, PICA’s, and spinal cord were coagulated and divided. The tonsils were mobilized and the PICA’s were protected with moist cotton patties. On the basis of direct inspection, a decision was made whether or not to shrink the cerebellar tonsils. The tonsils were not shrunk if the obex area was open and a pulsatile flow of CSF could be seen to exit from the fourth ventricle into the dorsal cervical theca. In most cases, the tonsils were reduced with low voltage, bi-polar coagulation until the tonsillar tips were positioned at or slightly above the putative level of the foramen magnum.
Expansile Duraplasty
The dura was closed with a graft of autogenous pericranium or reconstituted cadaveric dura. Cadaveric grafts, which most closely resemble living dura, were discontinued for general use in 2001 because of limited supply and concerns related to transmissible prions. A graft of approximately 5 cm in length and 2.5 cm in width was usually sufficient to produce a competent retrocerebellar space. The graft was anchored to the poles of the incision and sewn in place with continuous, locking sutures of 5-0 Goretex. Prior to tying down the last suture, the retrocerebellar space was inflated with 30-40 cc of sterile saline to expand the graft and eliminate intradural air bubbles that can degrade CDU images. Valsalva maneuvers were performed to assure a watertight closure. CSF leaks were corrected by oversewing the suture line.
CDU Imaging After Closing Dura
Following dural closure, CDU imaging was repeated to assess the goals of surgery. Optimal CSF flow was found to have the following characteristics: a peak velocity of 3-5 cm per second, bidirectional movement, and a wave form exhibiting arterial, venous, and respiratory variations (Fig. 4). Figures 5 and 6 show typical findings before and after lysis of the arachnoid, shrinkage of the tonsils, and duraplasty. Postoperative neuroimaging demonstrated a normal-sized cisterna magna and unrestricted CSF flow anteriorly and posteriorly through the foramen magnum in most patients.
 |
 |
FIGURE 4. Optimal CSF flow characteristics. Pulse wave Doppler tracings in 38 year old female with 19 mm tonsillar herniation after lysis of arachnoid, shrinkage of tonsils, and duraplasty show a peak CSF velocity of 4 cm per second, bi-directional movement, arterial pulsations (A), and respiratory and venous variations (B).
|
 |
 |
FIGURE 5. CSF flow in 39 year old female with 10 mm tonsillar herniation (same case as Fig. 1) before and after lysis of arachnoid, shrinkage of tonsils, and duraplasty. A, pulse wave Doppler tracing with sagittal target image (above) prior to opening dura showing no significant CSF flow caudal to cerebellar tonsils (cursor). B, pulse wave Doppler tracing with sagittal target image (above) after internal decompression showing CSF velocity of 3 cm per second. Red signal on target image (cursor) represents CSF flow; multicolor blots represent aliasing from vascular structures. C (below left), sagittal CDU image after internal decompression showing CSF flow (red signal) through cisterna magna and dorsal cervical theca. Multicolor blots represent aliasing from vascular structures. D, dura; DP, duraplasty; P, posterior inferior cerebellar artery; T, cerebellar tonsil; SC spinal cord.
|
 |
FIGURE 6. CSF flow in 40 year old male undergoing reoperation for failed Chiari surgery after lysis of arachnoid, shrinkage of tonsils, and duraplasty. Sagittal CDU image shows jet stream flow of CSF (blue signal) out of fourth ventricle (arrow) with peak velocity of 4 cm per second as compared to no flow prior to internal decompression. Variations in color-coding are a function of direction of flow. D, dura; T, cerebellar tonsil; SC, spinal cord.
|
Overly generous duraplasties and iatrogenic meningoceles were associated with sub-optimal CSF flow velocities (<1 cm per second) (Fig 7A). The problem could usually be corrected by tightening the graft with reinforcing sutures or adding a restrictive graft (Fig. 7B). Excessive CSF flow velocities (>8 cm per second) were most often encountered during reoperations in which the dura was densely scarred and thickened. Such observations are consistent with principles of fluid mechanics governing rates of flow through spaces of varying compliance and cross sectional area.
 |
 |
FIGURE 7. CDU images of cervicomedullary junction in 49 year old female undergoing reoperation for failed Chiari surgery and meningocele formation. A, pulse wave Doppler tracing with sagittal target image (above) showing no CSF flow in cisterna magna (cursor) prior to repairing meningocele . Multicolor blots represent aliasing from vascular structures. B, pulse wave Doppler tracing with sagittal target image (above) showing CSF velocity of 3 cm per second in cisterna magna (cursor) after repairing meningocele with restrictive duraplasty. M, meningocele; T, cerebellar tonsil; DP, old duraplasty; RD, restrictive duraplasty.
|
Wound Closure
The suboccipital craniectomy was covered with a titanium mesh/acrylic cranioplasty to protect the duraplasty and limit the extent of extradural scarring. The paraspinal muscles were sewn to the inferior border of the plate. In patients at risk for a CSF leak, a lumbar drain was inserted prior to leaving the operating room.
DISCUSSION
The surgical treatment of CMI requires familiarity with a great number of technical options, some of which are controversial. Among the unsettled questions are the following: the size of the craniectomy (8, 10, 16); the extent of the laminectomy (10, 15); the need to preserve (13) or open (1) the arachnoid; the need for additional steps such as lysis of the arachnoid (9, 15), shrinkage or resection of the cerebellar tonsils (1,7), stenting of the fourth ventricle (5), and plugging the obex (14); the need to close dura (15) or leave it open (19); the size of the duraplasty (10, 16); the optimal material for duraplasty (1, 9, 10); and the need for a cranioplasty (17). There is no consensus on these matters and little in the way of scientifically controlled comparative data.
Complicating the treatment of CMI is the wide range of patient-specific variables. These include the level of tonsillar descent, the degree of craniovertebral dysmorphism, the tightness of posterior fossa, the completeness of the CSF block, and the presence or absence of syringomyelia. It is doubtful that any operative technique with fixed steps and inalterable dimensions of decompression can deal effectively with all these variables.
The introduction of CDU as an intraoperative guide for posterior fossa surgery is an important technological advance that permits the neurosurgeon to tailor operative steps ac-cording to patient-specific variables on the basis of real time anatomical and physiological measurements. Of its potential uses, CDU is particularly well suited for Chiari surgery in which technical decisions are based largely on preoperative neuroradiological findings. A potential disadvantage of CDU is its sophisticated methodology. Special training is required for competence in the acquisition and interpretation of data.
The development of strategies for tailoring the steps of Chiari surgery using CDU involved a number of assumptions. The first of these was that the area of cerebellar impaction conforms to the area of compressed or obliterated subarachnoid spaces. Supporting this assumption is neuroradiological evidence that effacement of the subarachnoid space is one of the earliest signs of cerebral displacement and that extraaxial masses tend to produce a congruent area of cortical sulcus compression (6). Another assumption was that the ideal volume of the reconstructed retrocerebellar space should be in the range of 8-10 cc. This was based on two pieces of evidence: (1) anatomical estimates of cisterna magna size (18); and (2) volumetric measurements of the posterior cranial fossa showing that total CSF volume (26.7 ± 7.3 cc) is reduced by a mean of 10.8 ± 7.1 cc in patients with CMI (12).
No assumptions could be made concerning normal rates of CSF flow at the cervico-medullary junction. The measurements reported here were obtained in anesthetized patients in the prone position before and after posterior fossa decompression. Under the conditions of operation, an optimal rate of CSF flow was assumed to have been achieved in patients with adequate decompressions, visible evidence of an unrestricted and pulsatile flow of CSF from the fourth ventricle into the dorsal cervical theca, and a competent duraplasty. Mean peak velocities after decompression ranged from 3-5 cm per second as compared to 0-0.8 cm per second prior to decompression. Since normal CSF exhibits to and fro movements with distinct vascular and respiratory variations (11), the presence of these findings on postdecompression wave form tracings suggests that the flow velocities achieved were in a physiological range. Postoperative CINE-MRI confirmed the presence of unrestricted CSF flow anteriorly and posteriorly through the foramen magnum in most patients.
The ability to adjust surgical strategies on the basis of real time anatomical and physiological measurements is a defining step in the evolution of Chiari surgery. It remains to be determined which strategies are optimal. Ultimately, the value of the current technique will depend upon the results of an ongoing longitudinal study that correlates CDU-guided operative steps with pre- and postoperative MRI, pre- and postoperative CINE- MRI, symptoms and signs, and long-term patient outcome.
Disclosure Statement: The authors have received no financial support in conjunction with submission of this manuscript.
REFERENCES
- Batzdorf U: Treatment of syringomyelia associated with Chiari I malformation, in Tamaki N, Batzdorf U, Nagashima T (eds): Syringomyelia. Current Concepts in Pathogenesis and Management. Tokyo, Springer-Verlag, 2001, pp 121-135.
- Bolognese PA, Fasano VA: Intraoperative ultrasonography in neurosurgery, in Paletto AE (ed): Trattato di Technica Chirugia, volume XIII tomo primo. Torino, UTET, 1990, pp 157-186.
- Dohrmann GJ, Rubin J M: Use of ultrasound in neurosurgical operations: a preliminary report. Surg Neurol 16:362-366, 1981.
- Ellenbogen RG, Zeidman SM: Craniovertebral decompression for Chiari malformation, in Kaye A, Black P (eds): Operative Neurosurgery. London, Churchill Livingston, 2000, pp 1725-1741.
- Griebel ML, Oakes J, Worley G: The Chiari malformation associated with myelo meningoceles, in Rekate HL (ed): Comprehensive Management of Spina Bifida. Boca Raton, CRC Press, 1991, pp 83-89.
- Grossman RI, Yosem DM: Neuroradiology. The Requisites. St. Louis, Mosby, 1994, pp 149-169.
- Guyotat J, Bret P, Jouanneau E, Ricca A, Lapras C: Syringomyelia associated with type I Chiari malformation. A 21-year retrospective study on 75 cases treated by foramen magnum decompression with a special emphasis on the value of tonsils resection. Acta Neurochir 140:745-754, 1998.
- Holly LT, Batzdorf U: Management of cerebellar ptosis following craniovertebral decompression for Chiari I malformation. J Neurosurg 94: 21-26, 2001.
- Klekamp J, Batzdorf U, Samii M, Bothe HW: The surgical treatment of Chiari I malformation. Acta Neurochir 138:788-801, 1996.
- Lazareth JA, Galarza M, Gravori T, Spinks TJ: Tonsillectomy without craniectomy for the management of infantile Chiari I malformation. J Neurosurg 97:1018-1022, 2002.
- Milhorat TH: Hydrocephalus and the Cerebrospinal Fluid. Baltimore, Williams and Wilkins, 1972, pp 7-11.
- Milhorat TH, Chou, MW, Trinidad EM, Kula RW, Mandell M, Wolpert C, Speer MC: Chiari I malformation redefined: clinical and radiographic findings for 364 symptomatic patients. Neurosurgery 44: 1005-1017, 1999.
- Oldfield EH: Editorial. Cerebellar tonsils and syringomyelia. J Neurosurg 97:1009-1010, 2002.
- Peerless SJ, Durward QJ: Management of syringomyelia: a pathophysiological approach. Clin Neurosurgery 30:531-576, 1983.
- Rekate HL: Management of Chiari malformations, in Dickman CA, Spetzler RF, Sonntag VKH (eds): Surgery of the Craniovertebral Junction. New York, Thieme, 1998, pp 507-521.
- Sahuquillo J, Rubio E, Poca MA, Rovira A, Rodriguez-Baeza A, Cervera C: Posterior fossa reconstruction: a surgical technique for the treatment of Chiari I malformation and Chiari I/syringomyelia complex: preliminary results and magnetic resonance imaging quantitative assessment of hindbrain migration. Neurosurgery 35:874-884, 1994.
- Sakamoto H, Nishikawa M, Hakuba A, Yasui T, Kitano S, Nakanishi N, Inoue Y: Expansive suboccipital cranioplasty for the treatment of syringomyelia associated with Chiari malformation. Acta Neurochir 123:949-961, 1999.
- Schaefer GB, Thompson JN, Bodensteiner JB, Gingold M, Wilson M, Wilson D: Age-related changes in the relative growth of the posterior fossa. J Child Neurol 6:15-19, 1991.
Williams B: Surgery for hindbrain related syringomyelia. Adv Tech Stand Neurosurg 20:107-164, 1993.
