Spinal Cord Herniation and Spontaneous Cerebrospinal Fluid Leak

Summary of Key Points

•

Spinal dura is thinner than cranial dura due to its single histologic layer and as such it is more prone to tear.
•

Idiopathic spinal cord herniation (ISCH) and spontaneous cerebrospinal fluid (CSF) leak are both potential sequelae of dural defects and most commonly affect middle-aged females.
•

There is no consensus on the etiology of either disease process.
•

ISCH is typically found in the midthoracic spine and may cause Brown-Séquard syndrome or spastic paraparesis.
•

Once diagnosed, ISCH requires surgical intervention with reduction of the cord herniation and either patching or widening of the dural defect to prevent recurrent trapping of the spinal cord.
•

Spontaneous CSF leaks will lead to spontaneous intracranial hypotension, resulting in persistent, often positional headaches and potentially subdural fluid collections.
•

Treatment of spontaneous CSF leaks may include observation, medical management, and epidural blood patch.

Anatomy and Histology of Spinal Dura

The meninges consist of three layers throughout the cranial and spinal components: the dura, arachnoid, and pia. The dura is also known as the pachymeninx, and the arachnoid and pia are collectively termed the leptomeninges. Intracranially, both layers of the dura—meningeal layer and periosteal layer—are adhered, except for at the venous sinuses, where the two dural layers separate from each other.

However, the anatomic architecture of spinal dura is different. Rather than the two layers of dura adhering tightly to each other, as dura reaches the foramen magnum, the two layers separate. The outer layer is continuous as the periosteum of the spinal canal, whereas the inner layer forms the dural sac, enclosing the spinal cord and nerves. The space between the two layers is known as the epidural space, even though by definition it is inside the dura. The epidural space contains loose connective tissue, fat, and the venous plexus. If one were to make a corollary, the intracranial venous sinuses would be analogous to the epidural space—they are both spaces where the two layers of dura have diverged to house anatomic components, in one case venous blood and in the other ligamentum flavum, fat, and venous plexus.

As the spinal nerve roots exit from the spinal canal through the intervertebral foramen, the two layers of spinal dura join. At the inferior aspect of the dura mater, the dural sac encloses the cauda equina and terminates at the S2 level. It continues below this level as the filum of the dura mater, which is anchored to the sacral periosteum by the fibrous coccygeal ligament. The spinal cord itself is attached to the dural sac by the dentate ligaments, between successive nerve roots. The denticulate ligament also divides the anterior and posterior nerve roots.

Delving deeper into the dura, Vandenabeele and associates studied human dura under electron microscopy (EM) and found three distinct layers: a fibroelastic outer layer, a fibrous middle layer, and a cellular inner layer. The authors further concluded through EM that dural fibers ran in different directions, contrary to the common teaching that dural fibers run longitudinally. A similar study using EM by Dittmann and colleagues concluded that in the outermost epidural layer of the dura the collagen fibers are bunched together in bands that run in all directions: longitudinal, horizontal, and transverse. However, on the innermost arachnoid side, thin collagen fibers are fused into layers in such a way that the innermost layer resting on the arachnoid has a smooth, shiny appearance comparable to that of serosa. Despite what EM studies have revealed, the greatest strength lies in the longitudinal direction. One study confirmed the higher ultimate stress of the spinal dura when stretched longitudinally rather than transversely or circumferentially.

One interesting controversy surrounds a possible connection between the cervical musculature and the dura mater. Several cadaver studies cite a connection between the nuchal ligament and the spinal dura mater, where a connection between the rectus capitis posterior minor and the dura mater at the foramen magnum is seen. The result of this “myodural bridge” is thickening of the posterior cervical spinal dura. Clinically, authors have implicated its relation to headache and neck pain. Furthermore, it is possible that if a muscle tear or nuchal ligament rupture occurs, this may result in a cerebrospinal fluid (CSF) leak if the myodural bridge is torn and pulls the dura mater with it. As such, in a clinical scenario such as a fall or motor vehicle accident where the posterior neck musculature is stretched, this may potentially result in a leak of CSF.

Idiopathic Spinal Cord Herniation

Epidemiology

Idiopathic spinal cord herniation (ISCH) is an exceedingly rare condition with an estimated incidence of 0.08%. Relatively few case series are found in the literature, though certain centers have reported a greater number than others. The first known case dates from 1974 as described by Wortzman and coworkers, and since then roughly 180 cases have been reported.

This condition is typically found among the middle-aged or elderly, with an age range of 21 to 78 years and a mean of 49 years reported. It occurs more commonly in the female population, and in fact in one series 6 out of 7 patients were female.

Pathophysiology

ISCH occurs when the spinal cord herniates through a ventral dural defect and is subsequently incarcerated, leading to neurologic injury secondary to ischemia, scarring, and torsion of the spinal cord. It is most typically found in the midthoracic spine around the apex of thoracic kyphosis, where the spinal cord is most ventrally located within the canal. Various authors have proposed multiple theories of pathogenesis, but there is still no consensus. The theories range from congenital to acquired etiologies.

With regard to congenital causes, some have speculated a duplication of the ventral dura with subsequent spinal cord herniation through the inner layer of the dura, as well as the possibility of a ventral meningocele or extradural arachnoid cyst into which the cord herniates. However, this has not been conclusively supported radiographically or pathologically. In addition, the later onset, de novo nature of clinical symptoms argues for an acquired rather than congenital etiology.

Acquired causes may be traumatic or activity related, degenerative, or inflammatory. Though some patients have reported a remote history of trauma, most do not. Abnormal thoracic flexion-extension activity or exaggerated thoracic kyphotic curvature has been theorized to increase the risk of ISCH. Some have speculated that disc herniations may erode through the dura, creating a defect through which the spinal cord subsequently herniates. However, this is less likely given that the most common level of thoracic disc herniation is T4-5, whereas the most common level for ISCH is midthoracic. In addition, cases of spinal cord herniation may occur at the level of the vertebral body and not just the disc space.

Najjar and colleagues proposed an inflammatory mechanism based off of their case experience. They have suggested an initial inflammatory event with spinal cord and meningeal involvement which, given the ventral position of the thoracic spinal cord, leads to arachnoid adhesions between them. This scarring causes ventral spinal cord tethering, which subsequently progresses to gradual herniation through the dura secondary to the pulsations of the spinal cord. However, arguing against this theory, a history of infection or intraoperative evidence of inflammation is not always present.

Presentation

Clinical sequelae are related to the thoracic spinal cord incarceration, typically with insidious onset. Some have reported a period of intercostal pain prior to neurologic deterioration. Presenting signs and symptoms related to thoracic myelopathy may include Brown-Séquard syndrome or spastic paraparesis. Brown-Séquard syndrome tends to be the most common clinical presentation, generally 70% to 80% in prior series, though as low as 42% in one. Given the obscure nature of the disease unfortunately patients are frequently misdiagnosed for years prior to appropriate surgical treatment. A mean of roughly 5 years from onset of symptoms to surgical treatment has been reported in various case series.

Imaging

Either magnetic resonance imaging (MRI) or a computed tomography (CT) myelogram may reveal a ventrally displaced spinal cord with a C-shaped deformation. The differential diagnosis includes a dorsal arachnoid cyst, which is the most common misdiagnosis. A CT myelogram may be useful in distinguishing this, as an arachnoid cyst or other extramedullary lesions will not fill with contrast. Alternatively, phase contrast MRI may also be used to illustrate CSF flow. Nerve roots may also be seen traversing the dorsal subarachnoid space in ISCH but not in arachnoid cysts. In addition, certain extramedullary or extradural tumors may result in scalloping or bony remodeling of the vertebral body, which is not typical of spinal cord herniation. Finally, sharp angulation or kinking of the spinal cord is characteristic of ISCH and would not be typically found in the case of extradural compression. MRI may also demonstrate atrophy or spinal cord edema secondary to the entrapment. Additional differential diagnoses with these findings may include intramedullary tumors or transverse myelitis.

Imagama and colleagues proposed a classification scheme based on imaging findings, which delineates the progression of spinal cord herniation and correlates with the severity of preoperative symptoms and prognosis of postoperative recovery. Based on sagittal imaging, the authors described a kink type (type K) with ventral displacement of the spinal cord, a discontinuous type (type D) with disappearance of the cord at the site of herniation, and a protrusion type (type P) in which the ventral subarachnoid space disappears but there is no spinal cord kink. Axial imaging was used to classify the dural defect as either central (type C) or lateral (type L). Better prognosis was associated with type P herniation, which may represent the mildest form of spinal cord herniation. More severe preoperative myelopathy and poorer prognosis were associated with type C herniation, as the central location of the defect may result in greater torsion of the spinal cord and may involve its central portion, as well as make surgical access more difficult. The presence of bony vertebral defect along with the dural defect was also related to more severe preoperative symptoms.

Treatment and Outcome

Once diagnosed, surgical intervention is typically necessary to reduce the spinal cord herniation and prevent further neurologic injury. Although a small number of patients with mild, sensory-only symptoms have been followed conservatively, any progression of symptoms, especially myelopathic, should necessitate surgery.

Although Wortzman and coworkers initially described an anterior transthoracic approach, for most cases a posterior approach is standard and involves laminectomy, dural opening, and lysis of the dentate ligaments to mobilize the spinal cord and identify the defect. Use of intraoperative neuromonitoring including motor-evoked potentials (MEPs) is important given the need for spinal cord manipulation. It is inadequate to simply drain the subarachnoid fluid without reducing the cord herniation, as has been illustrated in cases where ISCH was initially diagnosed as a dorsal arachnoid cyst. The spinal cord must be restored to its normal anatomic position and recurrent herniation prevented. Given the ventral location of the dural defect, direct primary repair is difficult and involves a high risk for spinal cord injury from a posterior surgical approach. As such, typically a dural graft is used instead to cover the defect and prevent reherniation. Different grafts have been utilized including synthetic (Dacron, Teflon), allograft, and fascial autograft. Postoperative MRI is obtained and should demonstrate the spinal cord restored to its anatomic position and surrounded by CSF.

Some authors have advocated for simply widening the dural defect with resection of the dural ring to prevent entrapment and strangulation of the spinal cord. This is felt to require less manipulation of the compromised spinal cord than with repair of the dural defect and was found in some series to have better outcomes (90% as opposed to 70% with dural grafting). Though rare, recurrence has been reported as far as 10 years out from initial surgery despite dural patching.

Postoperative outcome and prognosis is dependent on both the severity and duration of preoperative symptoms, as well as radiographic morphology. In a literature review by Summers and colleagues, out of 159 surgical cases, 74% of patients had improved neurologic outcome, 18% unchanged, and 8% worsened. Favorable outcome is more likely among patients with Brown-Séquard symptoms (90%) as opposed to spastic paraparesis (69%). Motor weakness is more likely to improve than spasticity, and most patients are left with some degree of residual numbness. The rule of thumb for these patients tends to be the earlier the diagnosis and treatment, the better the prognosis. Najjar and associates found that patients presenting with less than 3 years of symptoms had a better chance of good outcome (86%) than those with more than 3 years (77%). Radiographically, as previously discussed, the degree and central location of spinal cord kinking as well as the presence of bony erosion or herniation into the vertebral body denotes a worse prognosis.

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