Download these documents to share with your physician(s), family and friends:

>> Spinal CSF Leak Diagnosis

>> Classification of Spontaneous Spinal CSF Leaks

The most important clues to the diagnosis of intracranial hypotension lie in the patient history.

The hallmark symptom is a headache that is worse upright and improved with horizontal positioning. It is important to note that not all headaches related to spinal CSF leaks are positional and not every patient with a positional headache has a spinal CSF leak. Common and uncommon associated symptoms are listed on our Symptoms page.

History of medical procedures or trauma
Lumbar punctures, epidural injections and spinal surgery occurring prior to onset of symptoms may be responsible for a spinal CSF leak. Significant trauma may also result in a spinal CSF leak. See our page on Causes.

History of Heritable Disorders of Connective Tissue (HDCT)
Spinal CSF leaks that develop spontaneously or with minimal precipitant may be related to an underlying weakness of the spinal dura. Marfan Syndrome, Ehlers-Danlos Syndrome, other HDCTs as well as Polycystic Kidney disease have been associated with spontaneous leaks. In many cases, the symptoms related to intracranial hypotension are the first recognized evidence of a HDCT.

Known spinal problems
Spontaneous leaks are also associated with calcified discs or bone spurs that may or may not be known prior to the onset of symptoms.

Previous diagnosis of POTS (Postural Orthostatic Tachycardia Syndrome)
The poor tolerance of upright posture associated with POTS can mimic intracranial hypotension, but the two may also coexist. Patients with the hypermobile type of Ehlers-Danlos Syndrome are known to have a higher prevalence of POTS as well as a greater risk of spontaneous intracranial hypotension.

Previous diagnosis of Chiari
Low-lying cerebellar tonsils on cranial imaging can be the result of intracranial hypotension (and fully reversible) and may not be recognized as such. To complicate matters, congenital Chiari is also more common in patients with some types of HDCT.


Evidence of HDCT
There are a number of physical signs of HDCT that may be noted, such as joint hypermobility, high arched palate, dental crowding, tall stature, scoliosis, easy bruising, atrophic scars and a number of others.

Neurologic Findings
There are many findings including involvement of cranial nerves, cognitive changes, gait disturbance, sensory and motor changes, but the exam may be entirely normal.

Clinical mimics
There are a number of disorders that intracranial hypotension may mimic due to the wide range of neurologic presentations. Patients may present with dementia, ataxic gait, Parkinsonism and other movement disorders.

Serious neurologic presentations
Emergency Medicine physicians may see patients presenting clinically with quadriplegia, stupor, coma or other serious neurologic presentations. Death is a reported complication, albeit rare.


Diagnostic testing includes lumbar puncture (not required), cranial (brain) imaging, spinal imaging and occasionally a nuclear medicine study called a radioisotope cisternogram.

Lumbar Puncture (LP) may be done to measure the opening pressure and to collect CSF for analysis but this is not required to make the diagnosis. An opening pressure is often measured at the time of LP for myelography. Normal opening pressure has a reference range of 65 to 195 mm of water. The opening pressure is often <60, and can be unmeasurable or subatmospheric. Normal pressure is not uncommon, however, and does not rule out CSF leaking. High pressures also occur.
Cranial imaging includes MRI and CT

Cranial MRI
MRI of the brain should be done in ALL cases where a spinal CSF leak is suspected. This should be done without and with gadolinium enhancement.
There are 5 findings, remembered by the mnemonic SEEPS, however the absence of these findings does not rule out CSF leaking. The lower prevalence of dural enhancement associated with longer duration of symptoms can contribute to underdiagnosis.
Subdural fluid collections
Enhancement of pachymeninges
Engorgement of venous structures
Pituitary hyperemia
Sagging of the brain

The first cranial MRI image below demonstrates brain sag before treatment, with resolution 48 hours post-treatment.
Reproduced with permission from Wouter I. Schievink, MD and Cedars-Sinai, Los Angeles, CA.

MRI sag1 MRI sag2

Cranial CT
CT imaging of the brain is not infrequently ordered when a patient presents to the emergency room with acute onset of severe headache. Subdural fluid collections or subdural hematomas may be seen. MRI is, however, the preferred imaging of the brain for diagnosis.

These two cranial CT images show subdural hematomas pre-treatment and resolution 1 week post-treatment.
Reproduced with permission from Wouter I. Schievink, MD and Cedars-Sinai, Los Angeles, CA.

f3 f4

Spinal Imaging is performed to localize leaks. It should be emphasized that spinal imaging may not be necessary since many patients respond well to lumbar epidural blood patching. MRI without intrathecal contrast and CT myelography are used most commonly. Digital subtraction myelography is used increasingly. It should be noted that extradural CSF at C1-2 and the cervicothoracic junction are common false localizing signs. Spinal imaging continues to be refined but remains insufficiently sensitive to image about half of suspected spinal CSF leaks. More than one type of spinal imaging and/or repeat imaging is often necessary.

Spinal MRI – MR imaging of the spine includes sequencing that makes CSF appear obvious without the use of intrathecal contrast. Because this is non-invasive, this is often the first type of spinal imaging used.

MR myelogram is MR imaging of spine with the use of specific sequencing/weighting (no intrathecal contrast).

Intrathecal gadolinium-enhanced spinal MRI – Spinal MRI may also be done following injection of gadolinium into the intrathecal space via lumbar puncture, although this is an off-label use of gadolinium.

The series of images below compares a normal MR myelogram on the left with 5 abnormal MR myelograms.
Reproduced with permission from Wouter I. Schievink, MD and Cedars-Sinai, Los Angeles, CA.

CTM normal CTM abN1 CTM abN2 CTM abN3 CTM abN4 CTM abN5

CT myelogram involves an injection of radiographic contrast into the intrathecal space via lumbar puncture, followed by a CT of the spine. An opening pressure may be measured and CSF may be collected for analysis. Several variables can be manipulated, depending on the circumstances, such as timing of the imaging after the injection of contrast (dynamic, early, delayed), the intrathecal injection of normal saline to raise the intrathecal pressure, or the use of digital subtraction imaging (see below).

Dynamic CT myelogram involves the performance of the CT as the contrast is being injected. This is most commonly used to localize rapid or high-flow spinal CSF leaks.

Digital Subtraction Myelogram is a myelogram (injection of intrathecal contrast + imaging) done under fluoroscopy with the additional ability to digitally subtract a pre-contrast image to enhance the visualization of the contrast. This is a dynamic form of imaging. This can be done under general anesthesia to stop breathing for a short time, thereby preventing movement artifact on the imaging. A CT can be done soon after or with a delay. This is used to image rapid leaks, ventral (front of the spinal cord) leaks or leaks not associated with an obvious extrathecal CSF collection, such as a CSF-venous fistula.

Digital Subtraction Myelography (DSM) image showing a ventral (anterior) leak.
Ventral spinal CSF leaks are best imaged with DSM.
Reproduced with permission from Wouter I. Schievink, MD and Cedars-Sinai, Los Angeles, CA.

Digital Subtraction Myelography (DSM) image showing a CSF-venous fistula.
This type of spinal CSF leak is also best imaged with DSM.
Reproduced with permission from Wouter I. Schievink, MD and Cedars-Sinai, Los Angeles, CA.

Radioisotope cisternogram is a nuclear medicine study that is occasionally used. This can help to confirm the presence of CSF leaking, but is insensitive in localizing leaks.

Most radiology departments can perform these procedures, however, in cases where a leak cannot be localized, neuroradiology departments that see larger volumes of spinal CSF leak patients may offer the additional expertise with interpretation as well as with optimization of imaging sensitivity. The understanding of anatomic types of leaks has resulted in refined imaging strategies to characterize a leak in a specific patient to guide treatment.

Classification of Anatomic Types of Spontaneous Spinal CSF Leaks
[Neurology. 2016 Aug 16; 87(7) 67379]

Type 1 – dural tear [26.6%] – extradural CSF collection in almost all
1a – ventral CSF leaks (96%) [associated with spinal pathology]
1b – posterolateral CSF leaks (4%)

Type 2 – meningeal diverticula [42.3%] – extradural CSF collection in 22.1%
2a – simple diverticula (90.8%)
2b – complex meningeal diverticula / dural ectasia (9.2%)

Type 3 – CSF-venous fistula [2.5%] – extradural CSF collections absent

Type 4 – indeterminate [28.7%] – extradural CSF collections in 51.5%

Presence or absence of extradural CSF collection
– extradural CSF collections were present in 50.5% of the patients
– of those patients,
52.3% Type 1
29.4% Type 4
18.5% Type 2
zero Type 3
This understanding of anatomic leak types has contributed to the evolution of imaging strategies
tailored to the suspected leak type. Treatment approach is then planned on the basis of specific leak type and location.

Diagnostic Criteria for Spontaneous Spinal CSF Leaks
Proposed Diagnostic Criteria for Spontaneous Spinal CSF Leaks (2011)
A: orthostatic headache;
B: the presence of at least one of the following:
-low opening pressure (≤60 mm H2O),
-sustained improvement of symptoms after epidural blood patching,
-demonstration of an active spinal cerebrospinal fluid leak,
-cranial magnetic resonance imaging changes of intracranial hypotension
(e.g., brain sagging or pachymeningeal enhancement);
C: no recent history of dural puncture; and
D: not attributable to another disorder.
(Headache 2011;51:1442-1444)

ICHD-3-beta, the International Criteria of Headache Disorders, 3rd edition, beta version, 2013
These criteria were based on the proposed criteria published in 2011 (above):
7.2.3 Headache attributed to spontaneous intracranial hypotension
Orthostatic headache caused by low cerebrospinal fluid (CSF) pressure of spontaneous origin. It is usually accompanied by neck stiffness and subjective hearing symptoms. It remits after normalization of CSF pressure.
Diagnostic criteria:
A. Any headache fulfilling criterion C
B. Low CSF pressure (C. Headache has developed in temporal relation to the low CSF pressure or CSF leakage, or has led to its discovery
D. Not better accounted for by another ICHD-3 diagnosis.
7.2.3 Headache attributed to spontaneous intracranial hypotension cannot be diagnosed in a patient who has had a dural puncture within the prior month. The headache in patients with spontaneous CSF leaks or spontaneously low CSF pressure may resemble 7.2.1 Post-dural puncture headache, occurring immediately or within seconds of assuming an upright position and resolving quickly (within 1 minute) after lying horizontally. Alternatively it may show delayed response to postural change, worsening after minutes or hours of being upright and improving, but not necessarily resolving, after minutes or hours of being horizontal. Although there is a clear postural component in most cases of 7.2.3 Headache attributed to spontaneous intracranial hypotension, it may not be as dramatic or immediate as in 7.2.1 Post-dural puncture headache. The orthostatic nature of the headache at its onset should be sought when eliciting a history, as this feature may become much less obvious over time.
Although autologous epidural blood patches (EBPs) are frequently effective in sealing CSF leaks, the response to a single EBP may not be permanent, and complete relief of symptoms may not be achieved until two or more EBPs have been performed. However, some degree of sustained improvement, beyond a few days, is generally expected. In some cases, sustained improvement cannot be achieved with EBPs and surgical intervention may be required. In patients with typical orthostatic headache and no apparent cause, after exclusion of postural orthostatic tachycardia syndrome (POTS) it is reasonable in clinical practice to provide autologous lumbar EBP. It is not clear that all patients have an active CSF leak, despite a compelling history or brain imaging signs compatible with CSF leakage. Cisternography is an outdated test, now infrequently used; it is significantly less sensitive than other imaging modalities (MRI, CT or digital subtraction myelography). Dural puncture to measure CSF pressure directly is not necessary in patients with positive MRI signs such as dural enhancement with contrast. The underlying disorder in 7.2.3 Headache attributed to spontaneous intracranial hypotension may be low CSF volume. A history of a trivial increase in intracranial pressure (e.g. on vigorous coughing) is sometimes elicited.
Postural headache has been reported after coitus: such headache should be coded as 7.2.3 Headache attributed to spontaneous intracranial hypotension because it is most probably a result of CSF leakage.

Diagnostic Difficulties

Misdiagnosis and delayed diagnosis of spontaneous spinal CSF leak remain common, largely due to low familiarity by health care professionals, as we see with all uncommonly recognized disorders. There is considerable variability in clinical presentations and many diagnostic challenges. Severity of symptoms and degree of disability are often underappreciated. Not surprisingly, many patients suffer for months or years before a correct diagnosis is made.

Diagnoses often considered
• primary headache disorders such as migraine headache, tension headache
• sinus headache
• cervicogenic headache (arising from neck problems)
• meningitis
• occipital neuralgia, trigeminal neuralgia
• psychogenic disorder (arising from psychological issues), depressive disorder, anxiety disorder

Differential or concurrent diagnoses
• POTS (postural orthostatic tachycardia syndrome) – may be an alternate or concurrent diagnosis
• Chiari malformation (congenital, with abnormal posterior fossa) – low-lying cerebellar tonsils in intracranial hypotension (reversible with treatment) may be mistaken as congenital Chiari. Intracranial hypotension may also worsen cerebellar tonsillar descent in patients with congenital Chiari.
• patients with Ehlers-Danlos Syndrome often have multiple comorbidities
• in some cases, intracranial hypotension may lead to a diagnosis of a Heritable Disorder of Connective Tissue not previously recognized; screening for vascular complications such as thoracic aortic aneurysm or intracranial aneurysm should be considered on a case by cases basis

Clinical presentations that may be unrecognized as possibly secondary to intracranial hypotension
• subdural hematoma
• dementia
• Parkinsonism or other movement disorders
• stupor or coma
• stroke
• reversible cerebral vasoconstriction
• posterior reversible encephalopathy syndrome
• spinal manifestations (radiculopathy, myelopathy, syringomyelia, quadriplegia, bibrachial amyotrophy) – large extradural fluid collections usually evident

Additional diagnostic challenges
• not every patient with intracranial hypotension has headache; there are many clinical presentations
• not every headache related to a spinal CSF leak is positional
• not every patient with a positional headache has a spinal CSF leak
• lumbar puncture is not required to make the diagnosis
• normal or high CSF pressures do occur and do not rule out the diagnosis
• interpretation of imaging requires experience / training
• normal cranial MRI does not rule out the diagnosis
• extradural CSF at C1-2 and the cervicothoracic junction are common false localizing signs
• sensitivity of spinal imaging is inadequate to localize a significant percentage of leaks, even with the most refined techniques, which limits treatment options
• more than one type of spinal imaging and/or repeat imaging is often necessary