Resolution of spontaneous intracranial hypotension due to a spinal CSF-venous fistula after a dynamic myelogram: illustrative case

Abstract

BACKGROUND

CSF-venous fistulas (CVFs) are abnormal connections between the spinal subarachnoid space and the adjacent venous plexus. They are typically treated by surgical disconnection, transvenous embolization, or percutaneous therapies such as epidural fibrin glue injection or blood patches that are known to be less effective for full resolution.

OBSERVATIONS

Here, the authors report a case of intracranial hypotension due to a spinal CVF that spontaneously resolved after a diagnostic spinal myelogram in the right lateral decubitus position. One hypothesis for this unique observation is the high viscosity of the contrast agent used.

LESSONS

CVFs can potentially resolve spontaneously after high-volume CSF drainage during a diagnostic spinal myelogram due to the high viscosity of contrast material in addition to a decrease in transfistula gradient.

https://thejns.org/doi/10.3171/CASE25194

CSF-venous fistulas (CVFs) are irregular connections between the spinal subarachnoid space and the adjacent venous plexus contributing to CSF hypovolemia, first proposed in 2014 as a treatable source of spontaneous intracranial hypotension (SIH).¹ Diagnosing CVFs is challenging as there is a lack of extradural contrast pooling, rendering them radiographically occult to conventional myelography techniques. One can use contrasted imaging to calculate the Bern score, which predicts the likelihood of having a spinal CSF leak.²

In patients with a high clinical suspicion of CVF, studies such as dynamic myelogram or dynamic CT myelogram with the patient in the lateral decubitus position can reveal the fistula.³ Currently accepted treatment options for CVFs include surgical disconnection, transvenous embolization,⁴ and percutaneous therapies such as epidural fibrin glue injection, which are known to be less effective than surgical intervetions.⁵ However, to the best of our knowledge, spontaneous resolution of CVF has not yet been reported in the literature.

Here, we report the spontaneous resolution of intracranial hypotension caused by a spinal CVFs due to the high viscosity of contrast agent used after a diagnostic spinal myelogram.

Illustrative Case

A 56-year-old female with a history of stage 3 triple-negative breast cancer in remission postlumpectomy and chemoradiation therapy and a history of pulmonary embolus on chronic apixaban therapy presented for progressively worsening headache, neck pain, and light-headedness over several months. Symptom onset appeared to be spontaneous, without associated trauma or procedures. Her headaches demonstrated an orthostatic nature, intensifying within 15 minutes of standing and nearly resolving when supine. This strongly suggested a diagnosis of intracranial hypotension. Her neurological examination was nonfocal. Her MRI brain study demonstrated signs that corroborated the diagnosis of SIH,⁶ including diffuse smooth dural enhancement (Fig. 1A), a diminished ponto-mamillary distance of 3.7 mm (abnormal distance < 5.5 mm) (Fig. 1B), and venous distension of the transverse sinus (Fig. 1C).

FIG. 1.

A: Postcontrast axial T1-weighted MR image demonstrating homogeneous dural enhancement. B: Postcontrast sagittal T1-weighted MR image with diminished ponto-mammillary distance (3.7 mm). C: Postcontrast sagittal T1-weighted MR image with biconvex transverse sinus, a sign of venous distension (arrow).

Both radiological and clinical presentation were highly suspicious of intracranial hypotension in this patient given a Bern score of 6 (2 points for venous engorgement, 2 points for pachymeningeal enhancement, 1 point for ponto-mamillary distance, and 1 point for size of prepontine cistern). However, her oncological history and meningeal enhancement seen on imaging also led to a high-volume lumbar puncture for CSF cytology as well as a dural biopsy to exclude leptomeningeal metastasis. These studies were negative for malignancy, further supporting a diagnosis of CSF leak. However, conventional CT myelography performed at an outside institution as well as dedicated thin-cut fat-saturated T2-weighted MRI of the spine failed to localize an epidural collection, suggesting a type 3 CSF leak.⁷ To confirm the possibility of a CVF, a dynamic myelogram with cone beam CT in the right lateral decubitus position was obtained, revealing an opening pressure of 13 cm H₂O. Approximately 20 ml of CSF was withdrawn and sent for cytology. The dynamic myelogram using 22 ml of contrast agent (Omnipaque, 300 mgI/ml, GE Healthcare) was performed and confirmed a low-flow right-sided CVF at T10 (Fig. 2A and B). The patient tolerated the procedure well and noted resolution of headaches immediately after the dynamic myelogram study.

FIG. 2.

A: Sagittal dynamic myelogram with cone beam CT obtained in the right lateral decubitus position, demonstrating CSF drainage into epidural vein (arrows). B: Oblique axial reconstruction of cone beam CT myelogram demonstrating fistula (arrows).

Her remission proved durable through both her remaining hospitalization and subsequent follow-ups at 6 weeks and 1 year. Surveillance MRI of her brain at both the 6-week and 1-year follow-up appointments demonstrated resolution of dural enhancement, normalized ponto-mamillary distance (5.7 mm),⁸ and a concave transverse sinus, suggesting normalization of her intracranial pressures (Fig. 3A–C).

FIG. 3.

A: Postcontrast axial T1-weighted MR image showing resolution of dural enhancement. B: Postcontrast sagittal T1-weighted MR image demonstrating normalized ponto-mammillary distance (5.7 mm). C: Postcontrast sagittal T1-weighted MR image demonstrating concave transverse sinus (arrow).

Informed Consent

The necessary informed consent was obtained in this study.

Discussion

Observations

Spinal CVF is a recently described source of SIH due to spinal CSF leakage, requiring a high degree of suspicion to diagnose. Current definitive treatment options for CVF include transvenous embolization and surgical nerve root ligation;⁴ however, spontaneous resolution in the setting of viscous contrast material has not yet been reported in the literature.

This case highlights the radiographically occult nature of CVF, as standard myelography, postmyelogram CT, and high-resolution MRI were all negative. Furthermore, even normal opening pressure on lumbar puncture, as observed in this case, did not exclude the presence of a CVF, consistent with recent reports.⁹ In our patient, the CVF identified at the T10 level spontaneously resolved after a high-volume lumbar puncture and a dynamic spinal myelogram obtained in the right lateral decubitus position. Retrospectively, it is possible that withdrawing a large volume of CSF prior to obtaining the myelogram reduced the sensitivity of the study by decreasing the pressure gradient across the fistula. This is supported by the observation of only a faint contrast opacification of the outflow vein.

In contrast, some patients with a high degree of suspicion for CVF but negative initial dynamic myelograms may benefit from maneuvers designed to enhance flow across the fistula. Such maneuvers include artificially increasing CSF pressure via intrathecal saline injection or inducing resisted aspiration.^10,11 Based on these studies, we have modified our practice to avoid high-volume CSF taps prior to dynamic myelogram studies in order to preserve diagnostic sensitivity.

Although contrast agents are essential for myelography and are used routinely, there have been no previous reports of spontaneous CVF closure solely attributable to the contrast agent, making this explanation alone unlikely. Similarly, patients with CVF live in a persistent state of CSF hypovolemia without spontaneous fistula resolution. We hypothesize that in this case, a combination of a transient reduction in CSF pressure caused by a high-volume diagnostic lumbar puncture performed at the time of the myelogram and the increased viscosity of the contrast agent given may have facilitated spontaneous fistula closure by promoting thrombosis within the outflow vein.

One weakness of this study is the absence of imaging to confirm fistula closure. As such, the CVF may reemerge at some point, causing rebound intracranial hypotension. However, the length of follow-up along with both clinical and radiographic resolution of intracranial hypotension is reassuring. Although repeat dynamic myelography would confirm closure of the fistula, the patient has remained asymptomatic and further invasive testing could not be justified.

Lessons

This is the first report of intracranial hypotension caused by a CVF presumptively resolved without direct surgical or percutaneous intervention. The mechanism of the purported resolution of CVF due to a combination of high viscosity of iodine contrast along with a decrease in transfistula gradient warrants further study.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: all authors. Acquisition of data: Peto, Hayman. Analysis and interpretation of data: Bales, Vakharia. Drafting the article: Bales. Critically revising the article: all authors. Reviewed submitted version of manuscript: Bales, Peto. Approved the final version of the manuscript on behalf of all authors: Bales. Administrative/technical/material support: Bales. Study supervision: Peto.

Correspondence

Natalie J. Bales: University of South Florida, Tampa, FL. nbales@usf.edu.