LOW CEREBROSPINAL FLUID HEADACHE

Published on 10/04/2015 by admin

Filed under Neurology

Last modified 22/04/2025

Print this page

This article have been viewed 2495 times

CHAPTER 62 LOW CEREBROSPINAL FLUID HEADACHE

Bahram Mokri

BACKGROUND AND TERMINOLOGY

In 1938, Schaltenbrand used the term aliquorrhea to describe the spontaneous occurrence of an entity manifested by very low cerebrospinal fluid (CSF) opening pressures and orthostatic headaches, among other features.^1,² Sometimes referred to as Schaltenbrand’s headaches, this later came to be known as spontaneous intracranial hypotension.³ It is now realized that practically all cases of spontaneous intracranial hypotension result from spontaneous CSF leaks,⁴ often at the level of the spine (particularly the thoracic spine⁵) and only rarely at the skull base. CSF leak leads to CSF volume depletion. Terms such as spontaneous CSF leak, CSF hypovolemia, and CSF volume depletion have been used interchangeably with spontaneous intracranial hypotension, because some patients with this disorder have consistently normal CSF opening pressures.^5,⁶ True hypovolemic state (reduced total body water), CSF shunt overdrainage, dural holes or tears as the result of lumbar puncture, epidural catheterization, surgery, and trauma all may lead to loss of CSF volume. In this chapter, we focus on spontaneous CSF leaks.

ETIOLOGY OF SPONTANEOUS CEREBROSPINAL FLUID LEAKS

The cause of spontaneous CSF leaks often remains undetermined, but two etiologic factors should be considered: trivial trauma and weakness of the dural sac.

Some patients report a history of trivial trauma, such as coughing, pushing, trivial falls, lifting, and sports activities. Evidence for weakness of the dural sac is accumulating. Meningeal diverticula are seen more frequently in patients with spontaneous CSF leaks. Increased frequency of meningeal diverticula and also CSF leaks have been reported in Marfan’s syndrome.⁷^–⁹ Some patients with spontaneous CSF leaks show stigmata of the disorders of a connective tissue matrix (e.g., marfanoid features, hyperflexible joints, hyperextensible skin).^5,¹⁰

In uncommon instances, a dural tear from a spondylotic spur or disc herniation ^11,¹² may cause dural defect and CSF leak.

CLINICAL MANIFESTATIONS

Headache is the most common manifestation. It is “classically” orthostatic (present in upright position, relieved by recumbency).^3,⁵ It is often not throbbing, but it may be throbbing; it is often bilateral, but it may be unilateral; and it is often aggravated or sometimes even triggered by Valsalva-type maneuvers. The headache may be frontal, fronto-occipital, holocephalic, or occipital. Not all headaches in CSF leaks are orthostatic, and variability is substantial (Table 62-1).^5,¹³^–¹⁷ Furthermore, not all orthostatic headaches result from CSF leaks. For example, they can be the dominant clinical manifestation in some patients with postural tachycardia syndrome.¹⁸

TABLE 62-1 Headache Variations in Cerebrospinal Fluid (CSF) Leaks

Orthostatic headaches may be preceded by neck pain or interscapular pain or both or by lingering nonorthostatic headaches.

With chronicity, sometimes the orthostatic features may dampen and the headache may be transformed into chronic lingering headaches that are present even in recumbency.⁵

Sometimes the headaches may have an acute thunderclap-like onset mimicking a subarachnoid hemorrhage ¹³ before orthostatic features are recognized.

In rare cases, a paradoxical postural headache (headache present in recumbency, relieved in upright position) may occur.¹⁴

Sometimes a second-half-of-the-day headache may occur. Typically, patients are headache free in the morning, but by later morning or early afternoon, they develop a headache that would increase in severity if the patients continued to be up and about; the headaches decrease clearly or vaguely in recumbency.¹⁵

Exertional headaches occur in isolation or are accompanied by some of the other clinical manifestations of CSF leaks and without any orthostatic features.¹⁶

With intermittent CSF leaks, the headaches may appear and disappear for variable periods of time.

Sometimes patients with documented CSF leaks and low CSF pressures and typical MRI abnormalities may have no headache at all—the acephalgic form.¹⁸

MRI, magnetic resonance imaging.

CLINICAL MANIFESTATIONS OTHER THAN HEADACHES

Many of the patients with spontaneous CSF leaks have one or, often, more symptoms in addition to the headaches (Table 62-2).^3,^5,¹⁹^–³²

TABLE 62-2 Clinical Features Other Than Headaches

Interscapular or neck pain (sometimes orthostatic), low back pain

Nausea, sometimes emesis, often orthostatic

Horizontal diplopia (unilateral or bilateral sixth cranial nerve palsy)¹⁹

Cochleovestibular manifestations (dizziness, change in hearing, tinnitus)^3,^5,²⁰

Photophobia, visual blurring

Upper limb numbness or pain ⁵

Rare manifestations: facial numbness or weakness, diplopia due to third or fourth cranial nerve palsy,²⁰^–²³ galactorrhea,²⁴ Menière’s disease–like manifestations,²⁵ upper limb radiculopathy,²⁶ encephalopathy,²⁷ stupor,²⁸ coma,²⁹ parkinsonism,³⁰ ataxia, incontinence, gait unsteadiness,³¹ frontotemporal dementia³²

Proposed mechanisms of clinical manifestations in CSF leaks and intracranial hypotension are listed in Table 62-3.^3,^5,¹⁹^–³⁴

TABLE 62-3 Mechanisms of Clinical Manifestations or Cerebrospinal Fluid Volume Depletion

Clinical Manifestation	Proposed Mechanism
Headache	Descent of the brain, stretch and distortion of pain-sensitive suspending structures of the brain^3,^5,^33,^34,^34a
Cranial nerve palsies	Stretching or compression of related cranial nerves¹⁹^–²³
Dizziness, change in hearing	Stretching of eighth cranial nerve or pressure changes in perilymphatic fluid of the inner ear^5,²⁵
Galactorrhea and increased prolactin	Distortion of pituitary stalk²⁴
Radicular upper limb symptoms	Stretching of cervical nerve roots or irritation by dilated epidural venous plexus^5,²⁶
Encephalopathy, stupor, coma	Diencephalic compression²⁷^–²⁹
Cerebellar ataxia, parkinsonism	Compression of posterior fossa and deep midline structures³⁰
Frontotemporal dementia	Compression of frontotemporal lobes³²
Gait disorder	Spinal cord venous congestion³¹

DIAGNOSIS

Cerebrospinal Fluid Examination

Considerable variability, not only between different patients but also in the same patient if multiple taps are performed at different times, is to be expected.

The opening pressure is typically low, sometimes unmeasurable, and sometimes it is “low normal” or even consistently within normal limits.

Analysis reveals the following characteristics:

Color: often clear, occasionally xanthochromic.

Protein concentration: normal or elevated (protein concentrations up to 100 mg/dL are common, and a concentration up to 1000 mg/dL has been reported).¹⁵

Glucose concentration: never low.

Leukocyte count: normal or elevated (pleocytosis with counts up to 50 cells/mm³ is common, and a count up to 222 cells/mm³ has been reported).⁵

Erythrocyte count: normal or elevated (difficult and traumatic taps are common in this disorder, and epidural venous plexus may be dilated).

Cytological and microbiological profiles: always negative.

Radioisotope Cisternography

Indium 111 is the radioisotope of choice. This is introduced intrathecally, typically through a lumbar puncture, and its movement is monitored by sequential scanning at various intervals, up to 24 or even 48 hours. Normally, by 24 hours (but often earlier), substantial radioactivity can be detected over the cerebral convexities. When a spinal CSF leak exists, the activity typically does not extend much beyond the basal cisterns. Therefore, images at 24 or even 48 hours reveal either absence or paucity of activity over the cerebral convexities.³⁵^–³⁷ This finding is the most common cisternographic abnormality in CSF leaks. Detection of parathecal activity that may point to the level or approximate site of the leak, although more desirable, is noted much less commonly (Fig. 62-1). Furthermore, meningeal diverticula, if large enough, may appear as foci of parathecal activity. Another cisternographic observation in CSF leaks is the early appearance of radioactivity in the kidneys and urinary bladder (<4 hours versus 6 to 24 hours), indicative of early entrance of extravasated isotope into the venous system and its early renal clearance and early appearance in the urinary bladder.

Figure 62-1 Indium 111 cisternography in a patient with spontaneous cerebrospinal fluid (CSF) leak. Note parathecal activity at 2 hours pointing to the level of CSF leak. At 24 hours, there is paucity of activity over the cerebral convexities, although activity can be seen in the spinal canal and, in particular, the basal cisterns.

Head Computed Tomography

Head computed tomography (CT) only infrequently shows subdural fluid collections or increased tentorial enhancement and overall is of very limited diagnostic value in this disorder.

Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) has truly revolutionized the diagnosis of spontaneous CSF leaks and has been instrumental in recognizing its broad clinical spectrum. Head and spine MRI abnormalities are listed in Table 62-4.

TABLE 62-4 MRI Abnormalities in Cerebrospinal Fluid Leaks

Head MRI

Diffuse pachymeningeal enhancement: the most common head MRI abnormality; may be thick or thin but is typically uninterrupted, nonnodular, bilateral, both supratentorial and infratentorial, and without leptomeningeal involvement (see Fig. 62-2A)^5,^41,⁵⁴

Descent (“sagging” or “sinking”) of the brain: manifested by descent of cerebellar tonsils, which may mimic Chiari type I malformation (see Fig. 62-3)⁵⁵; by obliteration of prepontine or perichiasmatic cisterns; by crowding of the posterior fossa; and by flattening of the optic chiasm (see Figs. 62-2A and 62-3)

Enlargement of pituitary gland: sometimes may mimic pituitary adenoma or hyperplasia (see Fig. 62-2A)^56,⁵⁷

Subdural fluid collections: may be unilateral but are often bilateral, are usually hygromas but may be or may become hematomas

Engorged cerebral venous sinuses ⁵⁸

Decrease in size of the ventricles, “ventricular collapse”²

Spine MRI

Extra-arachnoid fluid collection:^59,⁶⁰ often extends across several spinal levels; therefore, usually does not enable determination of the exact site of the leak (see Fig. 62-4)

Extradural extravasation of fluid: typically extends across fewer levels and, when seen, is more likely to point to the site or approximate site of the leak

Meningeal diverticula: may be single or multiple, small or large; may or may not be the site of the CSF leakage (see Fig. 62-4)

Spinal pachymeningeal enhancement ⁶¹

Engorgement of spinal epidural venous plexus ⁶²

MRI, magnetic resonance imaging.

Figure 62-4 Spine magnetic resonance imaging in a patient with spontaneous cerebrospinal fluid leak. A, Extra-arachnoid fluid extending across several spinal levels (arrows). B, Meningeal diverticulum (arrows).

(From Mokri B: Spontaneous cerebrospinal fluid leaks: from intracranial hypotension to cerebrospinal fluid hypovolemia—evolution of a concept. Mayo Clin Proc 1999; 74:1113-1123. Reprinted with permission of the Mayo Foundation for Education and Research.)

Computed Tomographic Myelography

Computed tomographic myelography (CTM) is myelography with water-soluble contrast material, followed by CT, typically at each spinal level, unless the myelogram itself or a previous cisternography or spine MRI reliably pointed to a more limited area of the spine to be investigated. CTM may show extra-arachnoid and extradural extravasation of contrast material (and therefore, the CSF) and so far is the most reliable test for locating the actual site of CSF leakage. It also shows meningeal diverticula and dilated nerve root sleeves and helps identify whether a noted meningeal diverticulum is the actual site of the CSF leak (Figs. 62-2B,C and 62-5).⁵ This test, like cisternography, also provides an opportunity to measure CSF opening pressure.

Figure 62-2 Composite figure pertaining to a patient with spontaneous cerebrospinal fluid (CSF) leak. A, T1-weighted gadolinium-enhanced coronal image while CSF leakage is active. Note diffuse pachymeningeal gadolinium enhancement (upper arrows) and enlargement of the pituitary gland (lower arrow). Also note that the optic chiasm above the pituitary gland is flattened and that the perichiasmatic cistern is partially obliterated. The patient has a large meningeal diverticulum, as seen in myelography (B) and computed tomographic myelography (C), believed to be source of the CSF leakage. After surgical treatment, along with disappearance of symptoms, all of the abnormalities on magnetic resonance imaging also reversed (D).

(From Mokri B: Intracranial hypertension after treatment of spontaneous cerebrospinal fluid leaks. Mayo Clin Proc 2002; 77:1241-1246. Reprinted with permission of the Mayo Foundation for Education and Research.)

Figure 62-5 Water-soluble myelography and computed tomographic myelography in spontaneous cerebrospinal fluid leaks. A, Myelogram shows a meningeal diverticulum (arrows). B, Computed tomographic myelography shows a meningeal diverticulum at T7 (arrow). C, Extra-arachnoid collections of contrast, giving the “dog ears” appearance on axial image. D, Leakage of contrast toward the paraspinal soft tissues (arrow).

(A, B, and D from Mokri B: Spontaneous cerebrospinal fluid leaks: from intracranial hypotension to cerebrospinal fluid hypovolemia—evolution of a concept. Mayo Clin Proc 1999; 74:1113-1123. C from Mokri B, Piepgras DG, Miller GM: Syndrome of orthostatic headaches and diffuse pachymeningeal gadolinium enhancement. Mayo Clin Proc 1997; 72:400-413. All reprinted with permission of the Mayo Foundation for Medical Education and Research.)

The rate of CSF leakage (fast-flow or slow-flow) may create special diagnostic challenges.

Fast-Flow Cerebrospinal Fluid Leaks

In this situation, after the myelogram and by the time of CT, a substantial amount of contrast has leaked and has spread across several spinal levels, making it virtually impossible to locate the exact site of the leakage. Dynamic CTM ³⁸ is often effective in solving this problem. In this technique, CT is performed immediately after intrathecal injection of contrast material, and high-speed multidetector spiral CT (which allows obtaining many cuts in a short period of time) is used. This technique is quite effective in detecting fast-flow CSF leaks as well as leaks from multiple sites.

Slow-Flow Cerebrospinal Fluid Leaks

When the leaks are slow, even after CTM, not enough contrast material has leaked to allow detection. A delayed CT, 3 to 4 hours after the initial one, sometimes reveals the site of the leak. Magnetic resonance myelography (spine MRI after intrathecal introduction of gadolinium), a test that awaits more confirmation and experience, may also prove to be helpful, inasmuch as images delayed as long as 72 or even 96 hours may be obtained.³⁹

MECHANISMS OF MAGNETIC RESONANCE IMAGING ABNORMALITIES

CSF volume depletion leads to (1) ventricular collapse and decrease in the size of the ventricles and (2) sinking of the brain (and, therefore, descent of the cerebellar tonsils [Fig. 62-3], decrease in the sizes of prepontine and perichiasmatic cisterns, flattening of the optic chiasm, and crowding of the posterior fossa). Furthermore, the loss of CSF volume, according to the Monro-Kellie doctrine,⁴⁰ mandates replacement of the depleted CSF volume, which is accomplished mainly by an intracranial venous hypervolemia. This results in (1) meningeal venous hyperemia, leading to diffuse pachymeningeal enhancement with gadolinium (leptomeninges have blood-brain barriers, but pachymeninges do not and, therefore, it is only the pachymeninges that enhance)⁴¹; (2) enlargement of the pituitary gland as the result of pituitary hyperemia; and (3) engorgement of cerebral venous sinuses. Subdural fluid collections are also volume-compensatory phenomena. At the spine level, CSF volume depletion and relative collapse of the spinal dura lead to engorgement of the epidural venous plexus.¹⁵

Figure 62-3 T1-weighted midline sagittal view in a patient with spontaneous cerebrospinal fluid leak shows descent of the cerebellar tonsils and crowding of posterior fossa.

CONSIDERABLE VARIABILITY

In spontaneous CSF leaks, clinical manifestations (including headaches), imaging findings, and CSF abnormalities (including opening pressures, which may range from unmeasurable to entirely normal) exhibit considerable variability. The core pathogenetic factor in this disorder is loss of CSF volume (CSF hypovolemia), which is the independent variable, whereas CSF pressures, imaging abnormalities, and clinical manifestations are variables dependent on the CSF volume (Fig. 62-6).⁴

Figure 62-6 Cerebrospinal fluid (CSF) hypovolemia as the independent variable in CSF leaks. CSF pressures, magnetic resonance imaging abnormalities, and clinical manifestations substantially vary as variables dependent on CSF volume.

TREATMENT

A variety of treatment modalities have been advocated in the management of patients with intracranial hypotension or spontaneous CSF leaks (Table 62-5). Some of these are based on previous experience with post–lumbar puncture headaches rather than direct experience with spontaneous CSF leaks.

TABLE 62-5 Treatment of Cerebrospinal Fluid Leaks

Fortunately, in some patients, the leak stops or slows spontaneously and the symptoms resolve. Bed rest has been traditionally recommended. Because many such patients have orthostatic symptoms, they tend to remain recumbent much of the time anyway.

Hydration (or, more precisely, overhydration), another traditionally advocated measure, is often tried, with variable results and essentially undetermined efficacy. An occasional patient may report a definite although nondurable improvement with intravenous fluid infusion.

Caffeine and theophylline, according to some studies,⁴² are mildly effective, but this effectiveness is doubtful to be impressive or durable.

Some patients report definite improvement with steroids, but many do not. It is unlikely that a substantial and durable improvement will occur with a course of corticosteroid therapy, and chronic corticosteroid use carries the recognized risks of many side effects.

Sporadic reports of successful results with epidural saline infusion exist,⁴³ but the experience is limited. This procedure can be tried, but with limited expectations, in patients in whom repeated epidural blood patches (EBPs) have failed. Reports, mostly anecdotal,⁴⁴ have mentioned the efficacy of epidural dextran injections. Intradural fluid infusions have been effective but typically only for as long as the CSF volume loss is thus compensated. However, in rare cases of rapid deterioration with obtundation or coma, intrathecal fluid injections may prove helpful.⁴⁵ Overall, the concern is about complications such as infections in sustained epidural and intrathecal infusions.

EBP has emerged as the treatment of choice for patients in whom initial conservative management has failed.^46,⁴⁷ The effect of EBP may be transient or sustained. EBP has an early effect that is simply a volume replacement resulting from the tamponade effect of the volume of the epidurally introduced blood, as well as a latent effect related to complete or partial sealing of the leak. A durable response is noted in approximately one third of patients with each EBP.⁴⁷ Thus, many patients may require more than one EBP. The efficacy of EBP in spontaneous CSF leaks is much less than what is seen in post–lumbar puncture headaches or even in the leaks that may follow epidural catheterizations. This difference is related mostly to the complex anatomy of the spontaneous CSF leaks (often not a simple rent or hole) but also to the site of the leak (which may not be in the posterior aspect of the dura or may be distant from the site where EBP is delivered).

Sporadic reports on the effectiveness of epidural injection of fibrin glue ⁴⁸ are encouraging. It can be an alternative when EBPs fail.

Surgery in well-selected cases is effective and can be tried when conservative and less invasive approaches, such as EBPs, have failed.⁴⁹ The surgery, however, may not be entirely straightforward, mostly as the result of the complexity of the anatomy of the leak. Besides, occasional patients have leaks at multiple sites.³⁸ It is essential to determine the site of the CSF leak and perform a thorough preoperative workup before surgery is undertaken.