Prevalence of cerebral vein thrombosis among patients with spontaneous intracranial hypotension

Abstract

Background and purpose

Cerebral venous sinus thrombosis (CVST) is a rare condition but an important complication of spontaneous intracranial hypotension (SIH). We reviewed our experience of patients with SIH and assessed for the presence of CVST.

Methods

We reviewed the medical records and imaging studies from our tertiary referral institution, assessing patients with clinically suspected SIH as well as imaging findings of intracranial hypotension. All relevant brain MRIs were reviewed for the presence of brain sag, pachymeningeal enhancement, and CVST.

Results

Among 563 patients with a clinical diagnosis of SIH, 431 (76%) demonstrated brain sag. In the overall patient cohort, a total of 5 patients had CVST (0.9%) and all 5 of these patients demonstrated findings of brain sag. Of the patients with CVST, 3 had significant complications, including dural arteriovenous fistulas (2 patients) and lobar hemorrhage with seizure (1 patient).

Conclusion

SIH is a risk factor for the development of CVST. In our review of 563 patients with clinical and/or imaging findings of SIH, 0.9% of patients were diagnosed with CVST and 3 of these patients (60%) had additional severe complications.

Introduction

Spontaneous intracranial hypotension (SIH) classically presents with orthostatic headache which resolves with recumbency and worsens in the upright position. Other symptoms include posterior neck stiffness, nausea, vomiting, hypacusia, blurred vision, tinnitus, and imbalance.^1,2 Patients with SIH may be misdiagnosed leading to delays in treatment and potentially serious consequences. ³

Cerebral venous sinus thrombosis (CVST) is an infrequently encountered condition with an annual incidence of 3–4 cases per 1 million population ⁴ . SIH has been shown to be a risk factor for the development of cerebral venous thrombosis, with one series estimating the prevalence of CVST in SIH patients as high as 2%. ⁵ Additional studies are needed to further evaluate the prevalence of CVST in SIH as this is a potentially life-threatening complication of the disease. We reviewed our experience in patients with SIH in order to further investigate the relationship between these two entities.

Methods

Patient population

We reviewed medical records and imaging studies performed at our institution, a tertiary referral center for SIH. A text search for the key words “intracranial hypotension”, “brain sag”, and “SIH” of brain MRI reports from January 1, 2000 through October 1, 2020 was performed and all examinations were reviewed for the presence of brain sag, CSF hypotension, and pachymeningeal enhancement. From this database, all brain MRI examinations with retrievable images were reviewed and those with positive findings (as detailed in the Image Review section) were included in our study. Additionally, an institutional database from the Department of Neurology identifying patients with the clinical diagnosis of SIH based off the International Classification of Headache Disorders criteria ⁶ was utilized. Only patients with at least one brain MRI were included in our retrospective review.

Image review

All brain MRIs from the patient's included in our study were reviewed for the presence of brain sag and pachymeningeal enhancement. Specifically, sagittal sequences were reviewed for the presence of brain sag and post-gadolinium images were reviewed in assessment for pachymeningeal enhancement. Definitions of brain sag and pachymeningeal enhancement were as specified by Dobrocky et al., and included effacement of the suprasellar cistern (<4.0 mm), effacement of the prepontine cistern (<5.0), or decreased mamillopontine distance (<6.5 mm). ⁷ Some patients had multiple MRI examinations and all of these imaging studies were reviewed. Patients who had a head CTV and/or MRV examination were also reviewed for the presence and sequela dural venous sinus thrombosis. The presence of dural venous sinus thrombosis was the primary outcome of the study. All examinations were initially interpreted and reported by a board-certified neuroradiologist at the time the imaging was acquired. The images were subsequently reviewed by a neuroradiology fellow for purposes of this study.

Statistical analysis

Descriptive statistics are reported as mean (SD) and N (%). All statistical analyzes were performed using JMP 14.0 (www.jmp.com, Cary, NC).

Results

We identified 563 patients with a clinical diagnosis of SIH. Their mean age was 48.2 years, and 64% were female. Of these patients, 431 had MRI findings of brain sag and/or pachymeningeal enhancement (76%, mean age 49.6 years, 67% female).

In the full patient cohort, 5 of 563 patients (0.9%) had CVST. Among the patients with brain sag, 5/431 had CVST (1.2%). Of the patients with CVST, mean age was 47 years and 3 patients (60%) were male. One patient had a remote diagnosis of lower extremity DVT, however no additional risk factors were identified following a thrombophilia consult. None of the patients had prior use of contraceptives with estrogen or other risk factors for venous thrombosis. Three patients with CVST had major complications, including dural arteriovenous fistulas (2 patients) and lobar hemorrhage with seizure (1 patient), and all three had brain sag on MRI.

Representative case examples

Patient 1

A 43-year-old male presented with a 10-day history of nausea and orthostatic headaches. Initial neurologic exam and head CT were normal. MRI demonstrated diffuse dural thickening and enhancement (Figure 1(a)) with effacement of the prepontine cistern. One week later he acutely developed right-sided weakness, and CTV demonstrated acute CVST (Figure 1(b) and (C)). The patient was anticoagulated with unfractionated heparin and admitted to the intensive care unit. On the first night, the patient had a generalized tonic-clonic seizure, and repeat head CT demonstrated a left parietal lobe hemorrhage (Figure 1(d)).

Figure 1.

(a) Sagittal CT venogram demonstrates occlusive thrombus involving the superior sagittal sinus along the vertex (black arrowheads). (b) Non-contrast head CT head shows a large intraparenchymal hematoma within the left parietal lobe (white arrow). A hyperdense thrombosed cortical vein (black arrowhead) is visualized along the superior aspect of the left parietal lobe. (c) Sagittal T2WI with fat saturation demonstrates a large ventral epidural fluid collection (white arrowheads). (d) Hyperdynamic CT myelogram demonstrates a CSF leak at the level of T1-T2 through a small dural defect (black arrows) and contrast was noted to fill in the large ventral epidural collection.

Cervical spine MRI performed 17 days after presentation demonstrated a large anterior extradural fluid collection spanning from C2 through T4 (Figure 1(e)) indicative of an underlying CSF leak. A hyperdynamic prone CT myelogram demonstrated a fast CSF leak due to a spiculated osteophyte at the T1-T2 interspace (Figure 1(f)). A focal blood patch and fibrin glue injection was performed at T1-2.

The patient's symptoms and imaging findings improved but then worsened again over several months. Surgery was performed 4 months after presentation, during which a ventral dural defect at T1-T2 was visualized and directly repaired. His symptoms and imaging findings resolved.

Patient 2

A 49-year-old male with two years of fluctuating headaches presented with intermittent pulsatile and non-pulsatile tinnitus. Initial brain MRI demonstrated numerous findings of SIH including severe brain sag, diffuse pachymeningeal enhancement, pituitary engorgement, and abnormal signal in the distal right transverse and right sigmoid sinuses suggestive of CVST (Figure 2(a)–(c)). MRA/MRV confirmed thrombus within the right sigmoid sinus with numerous enlarged and tortuous right-sided external carotid and trans-calvarial collateral arteries consistent with a dural arteriovenous fistula (Figure 2(d)). Cerebral angiogram confirmed a Cognard Type I/Borden Type I dural AVF involving the right sigmoid sinus, which was supplied by the posterior and petrous divisions of the right middle meningeal artery and trans-osseous branches of the right occipital artery (Figure 2(e)).

Figure 2.

(a) Axial gadolinium bolus MRV demonstrates a linear filling defect within the distal right transverse sinus (white arrow). (b) Sagittal T1WI reveals severe brain sag including effacement of the suprasellar and prepontine cisterns (white arrows), flattening of the ventral pons, and decreased mamillopontine distance. Additionally, there is engorgement of the pituitary gland with increased T1 signal suggestive of pituitary apoplexy (white arrowhead). (c) Cerebral angiogram from a right external carotid artery injection better demonstrates the dural arteriovenous fistula involving the right sigmoid sinus. The fistula is supplied by the posterior and petrous divisions of the right middle meningeal artery (black arrowheads) and trans-osseous branches of the right occipital artery (black arrow). (d) On 6-month follow-up, sagittal T1WI demonstrates marked decrease in size of the pituitary gland and resolution of the findings of intracranial hypotension.

Total spine CT myelogram was unrevealing except for numerous nerve root sleeve diverticula, bilaterally. A multifocal epidural blood patch was performed using a total of 8 transforaminal and 3 dorsal interlaminar needles and 70 mL blood. At 6-month follow-up, the patient's symptoms and constellation of imaging findings of intracranial hypotension had significantly improved. The patient elected to observe the dural AVF with non-invasive imaging.

Patient 3

A 39-year-old female with history of Chiari 1 malformation and Klippel-Trenaunay-Weber Syndrome with multiple calvarial and mandibular vascular malformations endorsed a history of transient orthostatic headaches that would come and go for weeks at a time over the previous several years. Extensive work-up and imaging at outside institutions were not available for review but per report the patient demonstrated evidence of intermittent CSF hypotension on four separate MRI examinations including varying degrees of diffuse pachymeningeal enhancement and bilateral subdural fluid collections. At time of presentation to our institution, she reported more severe symptoms, including orthostatic headaches that had persisted for 8 months. Headaches were bifrontal in location and throbbing in character. She had no other symptoms.

Remote imaging prior to the patient's course of illness also demonstrated variant anatomy of the deep venous system with a persistent falcine vein and underdevelopment of the vein of Galen and straight sinus. On imaging at presentation to our facility, MRI demonstrated findings of brain sag and MRV demonstrated occlusion of the persistent falcine vein. She was treated with warfarin and, over serial examinations, the persistent falcine vein recanalized (Figure 3(c)). No definite CSF leak was identified, and no epidural blood patch was performed. Following resolution of the thrombosis, the patient was headache free for 4 years. She then had recurrence of postural headache, though less severe, with continued varying degrees of positive imaging findings including brain sag and diffuse pachymeningeal enhancement.

Figure 3.

(a) Sagittal T1 weighted MRI demonstrates brain sag, including effacement of the suprasellar and prepontine cisterns. Inferior descent of the cerebellar tonsils was similar to prior examinations and compatible with the patient's Chiari 1 malformation. (b) MRV 3D MIP reconstruction images demonstrate variant anatomy with underdevelopment of the vein of Galen and straight sinus (arrowhead). The previously visualized persistent falcine vein is occluded (arrow). Additionally, numerous calvarial malformations are noted in keeping with the patient's diagnosis of Klippel-Trenaunay-Weber Syndrome. (c) Subsequent sagittal post-gadolinium MRV demonstrates the recanalized persistent falcine vein (arrow).

Patient 4

A 68-year-old female initially presented with headache that worsened while upright and while bending over. Headache, nausea, and vomiting could all be relieved by lying flat. The initial imaging, performed elsewhere and not available for our review, reportedly showed CVST and prompted treatment with warfarin. The patient had no history of prior venous thromboembolism. Approximately 3 months later, her CVST had resolved and warfarin was discontinued. Her headaches never fully resolved, and approximately 8 months after initial presentation her headaches worsened, and repeat imaging showed recurrence of CVST. Warfarin was restarted and a single-level lumbar epidural blood patch was performed. MRI 9 months after initial presentation demonstrated chronic thrombosis of the superior sagittal and right transverse sinuses and severe brain sag (Figure 4).

Figure 4.

(a) Reconstruction images gadolinium bolus brain MRV demonstrates chronic thrombosis of the superior sagittal sinus (white arrowheads). Additionally, there was chronic occlusion of the right transverse sinus and a small linear non-occlusive filling defect within the left transverse sinus (not shown). (b) Sagittal T1WI MRI shows marked sag of the brain including effacement of the suprasellar (white arrow) and pre-pontine (black arrow) cisterns. (c) T1WI post gadolinium MRI demonstrates diffuse dural enhancement and thickening.

The patient subsequently presented to our institution 11 months after initial headache onset. She remained symptomatic with a persistent posterior and left temporal headache that was relieved by lying flat. She had also developed pulsatile tinnitus. She was evaluated for a spinal CSF leak including a total spine MRI, a total spine CT myelogram, and a lateral decubitus digital subtraction myelogram, all of which failed to show a focal leak or CSF-venous fistula despite contrast in the renal collecting systems on early imaging which can be suggestive for a CSF-venous fistula (21). ⁸ A non-targeted dorsal interlaminar epidural blood patch was performed at T11-T12 and L2-L3. The patient had approximately 1 month of improvement of her headache before her symptoms returned. In the interim, the patient also developed progressive bilateral vision loss. A cerebral angiogram was performed after congested vessels were suggested on MRI and showed a complex Cognard Type IIa + b/Borden Type II dural arteriovenous fistula (dAVF) involving the bilateral transverse-sigmoid sinus regions (Figure 5). This was presumed to have developed secondary to the CVST, which was presumed to have developed secondary to the spinal CSF leak. Right and left dAVFs were treated sequentially with embolization and gamma knife radiosurgery, with improvement of symptoms.

Figure 5.

Cerebral angiogram including right external carotid artery (a) and left external carotid artery (b) injections. On the right, there is flow to the dural arteriovenous fistula from branches of the right middle meningeal artery, right occipital artery and right ascending pharyngeal artery. On the left, there is flow to the dural arteriovenous fistula from branches of an enlarged left occipital artery and the posterior branch of the left middle meningeal artery. Following both injections, prompt reflux of contrast is visualized into the distal superior sagittal sinus (white arrows) and additional components of the deep venous system.

Patient 5

A 40-year-old male initially presented to an outside institution with a dull bifrontal headache. He later developed blurred vision affecting both eyes and was found to have bilateral papilledema. Lumbar puncture opening pressure was elevated to 410 mm H2O. He was started on acetazolamide and underwent left optic nerve sheath fenestration. His headaches improved, but only for a few weeks. He was again found to have papilledema, and, per report, the MRV showed acute CVST of the bilateral sigmoid sinuses. He was started on warfarin and headaches resolved.

Approximately 2.5 years after his initial episode, the patient presented to our institution with a new orthostatic headache that had been present for 7 months. Headaches also worsened with coughing, sneezing, and straining. Brain MRI demonstrated severe brain sag and diffuse pachymeningeal enhancement. MRV showed diffuse irregularity of the dural venous sinuses with irregular filling defects of the superior sagittal sinus, both transverse sinuses, as well as the sigmoid sinuses. Although the more remote MRV images were unable to be reviewed, prior reports detailed much less CVST on the more remote imaging. Total spine MRI and CT myelogram failed to localize a CSF leak. Over the course of the next year, the patient underwent 3 non-targeted L2-3 dorsal interlaminar epidural blood patches (15–20 mL blood per patch). The patient had a dramatic improvement in symptoms, initially, but a milder orthostatic headache returned and persisted despite repeat epidural blood patches. A year after his presentation to our institution, MRI still showed similar findings of brain sag, diffuse pachymeningeal enhancement, and chronic post-thrombotic changes. His symptoms continued to wax and wane but were overall more tolerable than earlier in his course (Figure 6).

Figure 6.

(a) Sagittal T1WI demonstrates brain sag, including effacement of the suprasellar and prepontine cisterns with flattening of the ventral pons (white arrows) and inferior descent of the cerebellar tonsils (white arrowhead). (b) MRV 3D MIP reconstruction shows irregular filling of the superior sagittal sinus from the vertex and extending posteriorly (white arrowheads). Numerous prominent venous collateral pathways are noted. (c) Gadolinium bolus axial MRV images demonstrate non-occlusive linear filling defect within the right transverse and sigmoid sinuses compatible with thrombus (white arrow).

Discussion

Our study examining the prevalence of CVST in a cohort of SIH patients demonstrated several interesting findings. First, the overall rate of CVST was 0.9%, higher than the previously reported annual incidence of 3–4 cases per 1 million population (∼0.0004%). ⁴ Second, we found that complications related to CVST, including dural fistula formation, were fairly common. These findings are important as they suggest that SIH can be complicated by CVST and highlight the fact that timely treatment of this disease is critical.

An association between CVST and iatrogenic CSF leak was first described in 1987, ⁹ with the first reported case of spontaneous CSF leak and CVST appearing in 2004. ¹⁰ SIH has been identified as a risk factor for CVST, and prior studies have estimated the prevalence of CVST at 1–2% in patients with SIH.^5,11 The orthostatic headache and other symptoms of SIH are thought to result from the downward sagging of the brain secondary to the loss of CSF buoyancy, leading to traction on pain-sensitive structures and the cranial nerves and producing diplopia, facial numbness or pain, changes in hearing, vertigo and dysgeusia.^1,2

The various abnormalities seen on brain MRI in patients with cerebrospinal volume depletion can be explained by the Monro-Kellie doctrine. This doctrine states that because the cranial vault is a rigid structure with a fixed volume filled with CSF, brain parenchyma and blood, the loss of one component must be compensated by a gain in one or two of the other components. Since brain volume largely remains constant, cerebrospinal fluid depletion will trigger a compensatory intracranial hyperemic state primarily affecting the venous system. This results in diffuse meningeal venous hyperemia, engorgement of the venous sinuses and hyperemia of the pituitary gland. ¹²

Several mechanisms have been postulated to explain how intracranial hypotension can lead to CVST. Applying the Monro-Kellie doctrine of a constant or near-constant intracranial volume, the cerebrospinal fluid–depleted state will have a compensatory increase in cerebral blood volume. In order to accommodate increased intracranial blood volume containing primarily venous blood, the diameter of venous sinuses such as the sagittal or transverse sinuses will increase. ¹³ Since flow velocity is inversely proportional to the cross-sectional diameter of a vessel, venous stasis will occur as the caliber of the venous sinuses increases. ¹⁴ Indeed, following lumbar puncture, a 47% decrease in flow velocity was identified on transcranial doppler ultrasound of the straight sinus, which is more rigid and less susceptible to changes in diameter than other sinuses. ¹⁵ Additionally, depleted CSF volume will result in decreased CSF absorption into the cerebral venous sinuses and thereby contribute to increased blood viscosity. ¹⁶ Finally, sagging of the brain due to loss of CSF buoyancy may result in a negative intracranial pressure gradient that stretches or damages the venous endothelial lining ¹⁷ or results in mechanical distortion of the vessel wall and contributes to a prothrombotic state. ³

In our review, one of the patients (patient 5) initially had clinical findings that were compatible with intracranial hypertension and per report had subsequently developed CVST. Given that the prior imaging was not available for review, the timeline and causation of the thrombosis are not entirely clear. Given that the patient presented at our institution with a several month history of clinical symptoms SIH with progressive head and imaging findings of SIH on brain MRI, the causation of the apparent worsening sequela of CVST was thought to be related to the patient's intracranial hypotension however this relationship cannot be entirely proven. Additionally, this raises the possibility that patients who have intracranial hypertension, possibly from underlying CVST, could be at an increased risk for subsequent self-decompression through a spinal dural defect and develop symptoms of intracranial hypotension. Conversely, through the mechanisms described above, patients with primary SIH may develop CVST and be at risk for developing rebound intracranial hypertension if the dural venous sinus thrombosis remains present at the time of curing the spinal CSF leak. The intricate relationship between SIH and intracranial hypertension and their respective associations with CVST warrants further study.

Our study has several limitations given the nature of a retrospective review performed at a single academic institution without a dedicated control group. While every patient included in our study had a brain MRI for review, many patients did not have dedicated MRV or CTV examinations performed to assess for the presence CVST. There is significant overlap in both the clinical and imaging findings of SIH and other conditions, complicating estimates of the true prevalence of SIH. Also, not all patients in our review had findings of both brain sag and clinical evidence for SIH. Additionally, numerous patients carried the diagnosis of SIH, given their classic clinical presentation, but did not demonstrate abnormalities on imaging studies. The lack of findings on imaging may relate to the transient nature of imaging findings, the severity of the condition or other mechanisms of compensating for CSF volume depletion that are not yet understood. ¹⁸ Further, images were not blindly reviewed, and we did not have a control group of non-SIH patients.

Conclusion

CVST is an extremely rare condition which demonstrates an increased prevalence in patients with SIH at 0.9% in our review. SIH can be a challenging entity to diagnosis, however the clinical implications of a delayed diagnosis can be both debilitating and possibly devastating given the increased prevalence of CVST and its associated complications.

Glossary

Abbreviations

SIH

= spontaneous intracranial hypotension

CVST

= cerebral venous sinus thrombosis