CHAPTER 93 Syringomyelia

David M. Benglis, Jr., MD, Andrew Jea, MD, Steve Vanni, DO, Barth A. Green, MD

Syringomyelia, or cavitation within the substance of the spinal cord without an ependymal lining, has been recognized for more than 300 years as a pathologic entity. Etienne is credited for the first pathologic description in 1564 in La Dissection du Corps Humain; he described a cystic lesion in the spinal cord that contained a “fluid, reddish, like the fluidity of that of the ventricles.”¹

Portal, in 1804, first appreciated and connected the clinical syndrome of an intramedullary cyst with the pathologic changes of the spinal cord.² Ollivier then coined the term “syringomyelia,” combining the Greek words for “tube or pipe” and “marrow.” He documented a connection between the fourth ventricle and this cystic structure, which he believed to be a congenital anomaly.^1,^3,⁴

The classic clinical description of syringomyelia is described as a dissociated, suspended segmental sensory loss, consisting of loss of sensation to pain and temperature and preservation of sensation to proprioception and light touch. Slowly progressive distal motor dysfunction may also occur. A syrinx may extend to the medulla and cause lower brainstem dysfunction; in this case, the pathologic lesion is termed syringobulbia.

Those ependymal-lined cavities that appeared to be pathologic dilatations of the central canal were termed hydromyelia. Some authors viewed hydromyelia, in which the central canal was dilated but preserved, and syringomyelia, with or without a connection to the central canal, as stages of a common process. Hence unification of the terms resulted in the concept of syringohydromyelia or hydrosyringomyelia.¹^–³ ⁵

In 1973 Barnett published the first English-language monograph on syringomyelia.⁶ He proposed a classification based on a variety of clinical and experimental observations and studies. The classification scheme consisted of two broad categories: (1) communicating syringomyelia (e.g., Chiari I malformation, Chiari II malformation, basilar arachnoiditis) and (2) noncommunicating syringomyelia (e.g., occurring with spinal dysraphism, spinal cord trauma, spinal cord tumor, spinal arachnoiditis) (Box 93–1).

Recent experimental and clinical work including that of Oldfield and Milhorat and their colleagues ⁷^–¹³ has helped to clarify the pathophysiology and treatment of this complex syndrome.

Etiology, Pathology, Pathophysiology, Prominent Theories

Historical Perspective—Early Theories of Syringomyelia

Although the pathogenesis of syringomyelia has not yet been completely defined, the association between syringomyelia and congenital abnormalities was appreciated long ago. Tamaki and Lubin ¹⁴ credited Baulmer of establishing this relationship in 1887. Poser noted that Schlesinger’s 1895 monograph stated that there was an associated congenital abnormality in one third of the cases of syringomyelia that he had reviewed.¹⁵

Ollivier d’ Angers formulated and Leyden further refined the developmental theory of syringomyelia formation.^15,¹⁶ These authors stated that syringomyelia must be considered a congenital disorder associated with embryonic maldevelopment, specifically incomplete occlusion of the primitive fold. This improper fusion of the two folds of the primitive medullary groove allowed the abnormal lining of germinal cells to persist, resulting in simple hydromyelia.

Another theory from the congenital viewpoint was that there was a problem inherent in the environment to which the fetus was exposed. Kahler and Pick ¹⁶ in 1879 theorized that chronic intrauterine inflammation resulted in gliosis and aberrant development of the spinal cord that subsequently led to syrinx formation. In 1910 Haener,¹⁶ on the other hand, proposed that events during the act of birth (e.g., trauma) may arouse neural activity in abnormally enclosed tissue, with resultant syrinx formation.

W.J. Gardner—Hydrodynamic Theory “Water Hammer”

In a series of landmark papers, W.J. Gardner expounded his hydrodynamic theory of the pathogenesis of syringomyelia.¹⁷^–²³ Gardner’s theory was the first among three prominent theories to survive up to modern clinical practice. He based his theory on three observations: (1) dye injected into the ventricular system was recovered from the syrinx at operation, (2) fluid withdrawn from the syrinx at operation strongly resembled cerebrospinal fluid (CSF) found in the ventricular system, and (3) experimental hydrocephalus produced by obstruction of the normal outflow of CSF from the fourth ventricle resulted in the formation of syringomyelia that was in communication with the ventricular system.²⁴

Syringomyelia could be explained by failure of the embryonic rhombic roof to fenestrate during a critical period of development. The inability of CSF in the fourth ventricle to gain the usual access to the subarachnoid space during the sixth to eighth weeks of embryogenesis forced the hindbrain to herniate through the foramen magnum. A Chiari malformation was thereby created, and the failure of the CSF to expand the subarachnoid space resulted in communicating hydrocephalus. Gardner believed that the effect of the hindbrain malformation was to increase the obstruction to outflow at the foramen of Magendie and deflect the pulse wave of CSF into the opening of the central canal at the obex.

The pulse wave effect of the diverted CSF acted as a water hammer, gradually dilating the central canal or dissecting the substance of the spinal cord around the canal and creating a syrinx. From the perspective of Gardner, a congenital hindbrain defect that obstructed the CSF flow from the fourth ventricle to the subarachnoid space was the sine qua non of syringomyelia. Ball and Dayan ²⁵ and West and Williams²⁶ questioned Gardner’s theory and the necessity of a direct communication to the fourth ventricle for production of a syrinx. To support this statement, Milhorat and colleagues^7,²⁷ demonstrated in large autopsy studies that the majority of syrinxes did not communicate with the fourth ventricle and that the central canal was not patent in most normal adult patients.

B. Williams—Craniospinal Pressure Dissociation Theory

Williams proposed an alternative theory to explain syrinx formation and speculated that a partial block of the spinal subarachnoid space produced a pressure differential between the ventricular system and the spinal subdural space during Valsalva-type maneuvers. He explained that venous distention associated with these maneuvers produced an increased intracranial pressure that was not evenly distributed to the lumbar subarachnoid space because of a more proximal subarachnoid block. This pressure difference was labeled craniospinal pressure dissociation, and the lower pressure in the lumbar theca caused fluid to be drawn into the syrinx. This phenomenon was termed “suck.”

Williams also believed that the cavity enlarged after its initial formation as the result of compression of the lower end of the cavity with the rapid filling of the epidural venous plexus during a cough or sneeze. The fluid in the syrinx was then propelled rostrally, dissecting the central canal or pericentral parenchyma of the spinal cord. Williams applied the term “slosh” to this part of his theory to explain syrinx extension.²⁸

E. Oldfield—Abnormal Pulse Wave Theory

Oldfield and colleagues ¹³ used magnetic resonance imaging (MRI) with and without cardiac gating, intraoperative ultrasonography, and direct intraoperative observation of the exposed hindbrain and documented the downward movement of the cerebellar tonsils during systole. This group interpreted the data as obviating the necessity of a direct communication with the fourth ventricle, as advocated by Gardner. Moreover, they observed that the syringomyelic cord did not enlarge with Valsalva maneuver and that venous pressure had little to do with syrinx elongation, disputing the Williams “suck and slosh” theory. The authors proposed that the abnormal pulse wave in the spinal subarachnoid space, caused by the partial obstruction by the hindbrain, placed pressure on the spinal cord and dissected the central canal, causing the cyst to enlarge.

The ingress of CSF within the spinal cord parenchyma has recently been suggested to enter through dilated Virchow-Robbins spaces. The blockage of flow creates eddy-like currents analogous to a “boulder in a rapidly flowing river.” These forceful currents enter the cord parenchyma and first create microscopic changes (i.e., myelomalacia). They later develop into more confluent macrocystic cavities by dilation of the central canal and/or peripheral areas of the spinal cord.

Milhorat and colleagues ¹² proposed that normal CSF flow was from the spinal subarachnoid space through the parenchyma of the spinal cord into the central canal. The CSF then flowed into the fourth ventricle outlet at the obex. Their theory was supported with a rodent model of syringomyelia. They injected kaolin into the central canal of rats, causing stenosis of the proximal central canal through an inflammatory reaction. A resultant syrinx was formed. The authors suggested that syrinx formation was due to disruption of normal CSF flow by the inflammatory stenosis (Fig. 93–1).

FIGURE 93–1 Post-traumatic cord cyst. A low-power (×1.25) microscopic section image of a post-traumatic syringomyelia. There is a large, centrally located syrinx in the parenchyma of the spinal cord, which is surrounded by a thick wall of reactive astrocytes (arrows). The pia is thickened, and there are tissue changes that involve the spinal roots (arrowheads). The dura is also thickened (curved arrows).

(From Madsen PW, Falcone S, Bowen B, Green BA: Post-traumatic syringomyelia. In Levine A, Garfin S, Eismont F, Zigler J (eds): Spine Trauma. Philadelphia, WB Saunders, 1998.)

Communicating Syringomyelia

In 1896 Chiari published an addendum to an earlier work in which he described anomalies associated with hydrocephalus. In this latter publication, there were descriptions of patients with hydromyelia. Gardner and Goodall found that a majority of patients undergoing surgical decompression for symptomatic Chiari I malformation had a concurrent syringomyelia (Fig. 93–2).²⁹ Gardner and colleagues³⁰ demonstrated, at operation, communication between the syrinx of the upper cervical cord and the ventricles in patients undergoing suboccipital craniectomy and cervical laminectomy for decompression. Indigo-carmine was injected into the patient’s lateral ventricle, and colored CSF was recovered by direct puncture of the cervical syrinx.

FIGURE 93–2 Communicating syringomyelia (syringohydromyelia) with associated Chiari I malformation. (A) T1-weighted sagittal images show ectopia of the cerebellar tonsils and an intramedullary cystic cavity extending from C2 to T2. The T1-weighted axial images demonstrate the central location of the cyst (arrow) at C5-C7 (B–D).

Appleby and colleagues ³¹ established that a “communicating” type of syringomyelia could also be acquired from chronic arachnoiditis involving the basal cisterns and obstructing the outflow of CSF from the fourth ventricle.

Noncommunicating Syringomyelia

Syringomyelia Associated with Spinal Arachnoiditis

The association between spinal arachnoiditis and syringomyelia was first reported by Vulpian in 1861 and by Charcot and Joffroy in 1869. Some authors believed occlusion of blood vessels supplying the cord from profound arachnoid scarring was the underlying pathophysiologic process for intramedullary cavitation.³²^–³⁴

As previously described, Williams implicated craniospinal pressure dissociation, secondary to obstruction of the subarachnoid space, as the factor responsible for cyst formation and extension in this disease entity.³⁵ In 2004 Chang and colleagues³⁶ explained that the blockage of the spinal subarachnoid CSF pathway produces a relative pressure gradient inside the spinal cord distal to the blockage point that induces CSF leakage into the spinal parenchyma and the formation of syringomyelia. Barnett³⁷ and Milhorat⁷ considered this type of syringomyelia to be of the “noncommunicating” type because no connection between the cyst and fourth ventricle could be demonstrated.^, Koyanagi and colleagues³⁸ reviewed a series of 15 patients who underwent various shunting procedures for syrinx treatment caused by spinal arachnoiditis. Although neurologic improvement was found in a decent percentage of patients (60%), many required additional shunting procedures over time due to catheter blockage or failure.

Syringomyelia Associated with Spinal Cord Tumors

The association between syringomyelia and spinal cord tumors has been well established. Simon in 1875 was the first to report the simultaneous occurrence of syringomyelia and spinal cord tumors.³⁹ Proposed mechanisms for syrinx formation in this environment include (1) edema, (2) blockage of the perivascular spaces with resultant tissue fluid stasis, (3) cavitation secondary to disturbance of blood supply to the spinal cord, and (4) spontaneous hemorrhage or autolysis of the mass.^33,^40,⁴¹ Other authors contend that syringomyelia associated with a spinal cord tumor is due to a direct effect of the neoplasm.^15,^39,⁴²^–⁴⁴ Some authors believed that this disordered gliosis observed with tumor presence was also the underlying pathophysiologic cause even in cases not associated with an intramedullary neoplasm; this led some physicians to recommend radiation therapy as a rational but extreme form of primary therapy.^1,⁴⁵^–⁴⁸ Today, the use of radiation therapy should be restricted only for primary therapy of a known neoplasm or as an adjunct to surgical resection of tumor (Fig. 93–3).

FIGURE 93–3 Syringomyelia associated with invasive ependymoma of the cord. T1-weighted sagittal image obtained before (A) and after (B) gadolinium administration demonstrate the isointense, enhancing intramedullary mass at T11. Immediately inferior to the mass is an associated cyst or cysts (arrows). The cord is enlarged. C, A T2-weighted sagittal image shows the high signal intensity of fluid within the cyst(s) (arrows) and the relatively low signal intensity of the tumor. D, A T1-weighted axial image shows the eccentrically located cystic cavity (arrows), containing low-signal-intensity fluid, within the enlarged cord.

Spinal CSF Dynamics in the Presence of a Neoplasm

Total or subtotal obstruction of CSF flow by an intramedullary or sometimes extramedullary/intradural tumors may be a significant factor for the development of syringomyelia. The subarachnoid and the extracellular space of the central nervous system should be considered as a single-fluid compartment with no barrier to fluid movement between them. Interference with this normal CSF flow influences extracellular fluid flow out of the spinal cord. The high ratio of syrinx cavities associated with intramedullary tumors may be attributable to their simultaneous influence on the subarachnoid and extracellular spaces. Fluid transudation from tumor vessels, breakdown products of tumor cells, and in some cases active secretion also raise the protein content and thus the viscosity of the extracellular fluid contributing to further aberrances on normal flow dynamics.⁴⁹ Less commonly, intramedullary cysts can present in association with an extramedullary neurofibroma or meningioma. The removal of the extramedullary mass lesion is usually followed by a spontaneous collapse of the associated syrinx.

Syringomyelia and Spinal Cord Tumors—Clinical Studies

In a surgical series of 100 intramedullary tumors, 45% presented with associated syringes.⁴⁹ A syrinx was more likely to be found above than below the tumor level. Ependymomas and hemangioblastomas were the most common tumor types to be associated with syringes. Astrocytomas, on the other hand, tended to demonstrate syringes less often. The higher the spinal level, the more likely a syrinx was encountered.

Similarly, the mechanism for the extension of the syringomyelia remains a matter of controversy leading to several mechanisms being proposed. Some view the rostral and caudal extension of the syrinx, which produces late neurologic symptoms, being a result of a one-way valvelike trapping effect of the subarachnoid space into the cavity.^11,⁶²^–⁶⁶ Less frequently observed presentations of post-traumatic cysts include patients with a history of a herniated cervical or thoracic disc with resultant ventral compression from a bony or soft gibbus and disturbance of normal CSF flow. Spinal puncture with injection of an irritative dye (e.g., methylene blue) can cause an ascending arachnoiditis that may develop associated subarachnoid and/or intramedullary cysts. Patients who have subarachnoid hemorrhage may also develop spinal cord tethering with associated intramedullary or subarachnoid cysts.

One author has suggested that post-traumatic syringomyelia may be classified into two types. These authors contend that successful reestablishment of normal CSF flow dynamics by untethering the spinal cord stops the progression of clinical decline and in some cases may return valuable lost function. A preoperative distinction could be made on the basis of the presence (high-pressure type) or absence (low-pressure type) of the flow-void sign on a T2-weighted magnetic resonance image. With a midline myelotomy, fluid within a high-pressure syrinx would pour out, resulting in sustained neurologic improvement. In the low-pressure type, drainage of the syrinx would not collapse the expanded spinal cord and the surgical outcome would be modest at best.⁶⁷

Clinical Features

The onset of syringomyelia is commonly between the ages of 25 and 40. Males are somewhat more affected than females. Syringomyelia most often affects the cervical or thoracic spinal cord yet sometimes extends rostrally into the medulla.

Hydrocephalus may be found in 10% to 33% of patients but is more likely related to an associated Chiari malformation.²⁶ This frequent association supports the concept of a hydrodynamic mechanism for the formation of syringomyelia.

Syringomyelia—Physical Examination

Clinical features are variable and dependent on the anatomic structures involved in a cross-sectional area, as well as longitudinally. Anterior horn involvement results in weakness and wasting, especially in the upper extremities, and fasciculations. Posterior horn and decussating spinothalamic fiber involvement result in loss of pain and temperature sensation, usually in a suspended, segmental distribution, involving the arms and trunk and sparing the legs. Patients may occasionally feel pain that is characterized as boring or lancinating. The large fibers of the dorsal columns are usually unaffected, and therefore proprioception and light touch are preserved.

The autonomic pathways of the interomediolateral column may be affected, resulting in Horner syndrome, trophic changes of the skin, a neurogenic bladder, and dyshidrosis. Corticospinal tract involvement may give spastic paraparesis. Patients experience a loss of deep tendon reflexes in the upper extremities. A skeletal survey may reveal congenital anomalies including basilar impression and invagination, Klippel-Feil deformity, and spina bifida occurring primarily at C1.⁶⁸ Developmental scoliosis may occur as a result of cord cavitation and is probably not an unassociated congenital lesion.⁶⁹ Although discussed extensively in the literature, painless joint destruction (Charcot joints) occur in less than 5% of patients with syringomyelia.⁷⁰

Cranial nerve involvement is seen with syringobulbia. Most often signs and symptoms are unilateral. CN XII involvement results in tongue weakness and atrophy. CN XI dysfunction can result in weakness and wasting of the sternocleidomastoid and trapezius muscles, whereas CN IX and X involvement produces dysphagia and dysarthria. CN VII involvement results in a facial palsy, and involvement of the spinal tract of V produces sensory loss of pain and temperature of the face in an “onionskin” distribution, as well as reduced corneal sensation and reflex.

The clinical course of syringomyelia is usually insidious, worsening over several years to decades. Alternatively, the patient may worsen in an abrupt stepwise fashion punctuated by intervals of clinical stability.⁷¹ The prognosis of untreated syringomyelia is compatible with modified but productive survival for decades in 50% of surviving patients.⁷² The remainder become incapacitated and die as a direct result of the pathologic process. Seki and Fehlings⁷³ designed a rodent model to further elucidate the role of post-traumatic syringomyelia in spinal cord injury.

< div class='tao-gold-member'>

Only gold members can continue reading. Log In or Register to continue