Endovascular treatment of a sacral dural arteriovenous fistula

  1. Mokshal H Porwal 1,
  2. Sean Tutton 2,
  3. Grant P Sinson 1 and
  4. Raphael Sacho 1
  1. 1 Neurosurgery, Medical College of Wisconsin, Wauwatosa, Wisconsin, USA
  2. 2 Radiology, Medical College of Wisconsin, Wauwatosa, Wisconsin, USA
  1. Correspondence to Dr Raphael Sacho; rsacho@mcw.edu

Publication history

Accepted:27 Jan 2021
First published:05 Feb 2021
Online issue publication:05 Feb 2021

Case reports

Case reports are not necessarily evidence-based in the same way that the other content on BMJ Best Practice is. They should not be relied on to guide clinical practice. Please check the date of publication.

Abstract

Spinal dural arteriovenous fistula (SDAVF) is a rare pathological communication between arterial and venous vessels within the spinal dural sheath. Clinical presentation includes progressive spinal cord symptoms including gait difficulty, sensory disturbances, changes in bowel or bladder function, and sexual dysfunction. These fistulas are most often present in the thoracolumbar region. Diagnoses of SDVAFs are commonly missed, possibly due to the low index of suspicion, non-specific symptoms and challenging imaging. In this case report, we describe a rare presentation of a sacral SDAVF which was detected by collective efforts between endovascular neurosurgery and interventional radiology. We outline the diagnostic and imaging challenges we faced to discover the fistula. In particular, mechanical pump injection instead of hand injection during angiography was required to reveal the fistula. Following identification, the fistula was successfully treated endovascularly by using onyx (ethylene vinyl alcohol glue), a less invasive alternative to surgical intervention.

Background

Spinal dural arteriovenous fistulas (SDAVFs) are rare pathological conditions presenting with 5–10 cases per year per 1 million individuals in the population.1 This inappropriate arteriovenous communication causes venous hypertension and congestion in the spinal cord, leading to sensory disturbances and further progressing to myelopathy.2 The combination of rarity, non-specific symptoms and difficult imaging often leads to missed diagnoses.3 SDAVFs most often present in the thoracolumbar region. This report describes a rare pelvic presentation of an SDAVF which was further complicated by imaging challenges. Awareness of less common fistula locations and difficulties in the diagnostic process is important to provide timely treatment to reduce the degree of irreversible deterioration by early intervention.

Case presentation

An adult male (age 60–70) with history of diabetes and chronic kidney disease presented to establish care with a primary care provider. He had problem of progressive lower limb weakness and myelopathy. The patient described that he initially had sensory disturbances, which progressed to motor weakness in the lower limbs over the preceding 3 months. He was able to ambulate independently but had difficulty running. Diabetic polyneuropathy was suspected given the patient’s presentation and history; however, referral to neurology was requested for further workup.

During the neurology visit, the patient reported several recent falls and heaviness in both lower limbs. Additionally, he experienced burning sensations from the waist down including the rectum and perineal region. MRI scan of the brain, cervical, thoracic and lumbar spine demonstrated increased T2 signal intensity within the caudal thoracic cord from the level of T9 to the conus which enhanced with contrast. These findings raised suspicion for a dural arteriovenous malformation; therefore, consultation to neurosurgery was initiated.

The patient presented to the neurosurgery clinic reporting a particularly sharp decline in functioning over the prior months. Examination was compatible with upper motor neuron distribution of bilateral lower limb weakness with 4/5 power. Given the patients age, clinical presentation and referral history, additional MRI imaging was requested since a spinal dural fistula was suspected.

The new spinal cord MRI was suspicious for a spinal dural fistula, revealing oedema of the lower thoracic spinal cord and conus with degenerative canal stenosis and serpiginous intradural extramedullary veins (figure 1). Although the level of the fistula was not clear at this point, magnetic resonance angiography (MRA) of the spine demonstrated a prominent tortuous vessel along the left T8–T9 level (figure 2). Given this evidence, we decided to perform a digital spinal angiogram to locate and potentially treat the fistula.

Figure 1

Lateral thoracic MRI (A) and lateral lumbar MRI (B). Thoracic MRI shows dilated pial/subdural vessels and spinal cord oedema within the lower thoracic and conus region, noted by the green arrows (A). Lumbar MRI was consistent with no further notable characteristics (B). This is a classical appearance for a dural arteriovenous fistula.

Figure 2

Sagittal thoracic MRA. Thoracic MRA shows a prominent tortuous vessel along the left T8–T9 level suspicious for a spinal dural arteriovenous fistula (red arrows).

A spinal angiogram was performed under general anaesthesia, using predominantly hand injection technique for the intercostal and pelvic vasculature. An obvious spinal fistula was not identified initially, but on further inspection of the angiographic images, there was a suspicion of an abnormal pelvic vessel along the midline. Additionally, we had not been able to identify the right-sided L5 intercostal artery. We therefore decided to repeat the spinal angiogram, focusing on the pelvis and with a view to perform a cranial angiogram, should this be unremarkable. In conjunction with interventional radiology, repeat angiography was performed of the lower aorta and left-sided internal iliac artery using a power injection technique on this occasion. Glucagon was used to reduce peristaltic bowel shadows which had been noticed during the first angiogram. Resultingly, a pelvic dural arterial venous fistula arising from the left sacral artery, leading to a venous pouch and then radicular vein, was identified. Figure 3 presents the results of the two angiograms, showing the radicular draining vein on the second angiogram.

Figure 3

Anteroposterior projection with hand injection (A) and power injector (B) of left internal iliac artery angiogram. Note that the radicular draining vein is now visualised following power injection angiogram and glucagon was used to optimise image quality by reducing peristaltic bowel shadows in (B).

Treatment

Using a coaxial access system, a microcatheter was advanced into the lateral sacral artery, and selective angiography performed (figure 4). Embolisation of the fistula was successfully performed using onyx (ethylene vinyl alcohol) glue to close the fistula (figure 5).

Figure 4

Note the venous pouch (red arrow) that exists within the fistulous epidural network prior to the radicular vein exiting. Microcatheter injection of feeding artery arising from left lateral sacral artery demonstrating a radiculomedullary vein running along nerve root as it exits the fistula. The radicular vein drains superiorly and then enters the congested medullary venous plexus (yellow arrow).

Figure 5

Final result of endovascular obliteration using onyx glue, demonstrating the cast within the venous pouch and draining radicular vein.

Outcome and follow-up

The patient was seen 3 months after the embolisation of his fistula with significant recovery in his lower extremity weakness. Additionally, MRI of the spine showed significant improvement of oedema. Occupational therapy notes that the patient met all goals following treatment and is being discharged with continued home exercise programmes. Physical therapy notes significant improvements in ambulation; however, there are still some functional deficits. They are satisfied with the progress and anticipate he will be discharged within the next few weeks. Delayed catheter angiography by neurosurgery is planned for 6 months following his treatment for monitoring.

Discussion

While SDAVFs are the most common spinal vascular malformation, they tend to be underdiagnosed partly due to the non-specific presentation and imaging challenges.3 This can lead to progressive spinal cord symptoms including gait difficulty, sensory disturbances, changes in bowel or bladder function, and sexual dysfunction.2 We present an extremely rare case of an SDAVF within the pelvic region, branching from the superior gluteal artery. To our knowledge, fewer than 50 cases of SDAVFs have been reported in the sacral region to date.4 This further complicated the process of making an accurate diagnosis which asserted the importance of interspecialty collaboration, greater awareness of atypical fistula locations and thorough knowledge of the imaging challenges that may occur.

The patient presented with classical progressive myelopathy symptoms of a spinal arteriovenous fistula in conjunction with dilated veins on his MRI. A previous literature review of sacral dural arteriovenous fistulas revealed that most patients present with progressive lower limbs weakness, hyperintensity within the spinal cord, prominent perimedullary veins and intramedullary enhancement.4 Tortuous intradural veins are hypothesised to be seen in MRIs of patients with sacral SDAVF because the filum terminale vein is the only longitudinal collecting vessel seen below L2 level.5 Our patient’s presentation and MRI imaging was consistent with this literature which raised suspicion for a sacral SDAVF. A spinal angiography is warranted if presentation and MRI imaging raises any suspicion for an SDAVF.

Spinal angiography revealed a sacral SDAVF supplied by the superior gluteal artery. Branches of the internal iliac artery supply the sacral dural sheath. This case contributes to the current literature and provides some important lessons about the diagnostic process. During the first angiogram, we were unable to locate the fistula using a hand injection method. SDAVFs typically present in the thoracic and lumbar region; however, we were aware of the possibility of a pelvic fistula. Although hand injection is the simpler and preferred method for spinal investigation, we hypothesised that it may not have generated adequate force against the flow of blood in our patient. Many of the vessels injected during spinal angiography are smaller, radicular intercostal vessels, where hand injection may be the preferred method for performing angiographic runs. The novelty of our case emphasises that performing hand runs of the pelvic vasculature may not be adequate to exclude a pelvic source for spinal vascular malformations, and that more robust mechanical pump runs may help diagnose and localise a subset of spinal dural AV fistulas that can be challenging to detect. Collaboration with interventional radiology allowed us to use a mechanical injection method and glucagon administration to reduce bowel shadows during the second angiography. These collective efforts revealed the rare sacral fistula in our patient. Mechanical injection has been shown to greatly reduce radiation dose in cerebral angiographies, without any statistical differences in image quality and vertebral reflux.6 Based on our imaging challenges, we hypothesise that it may be beneficial to approach atypical SDAVFs with mechanical injection methods for accuracy and safety. However, we did not find any literature exploring the diagnostic sensitivity of SDAVFs using hand versus mechanical injection, indicating that further research about this topic is warranted.

We opted to embolise the fistula using ethylene vinyl alcohol glue instead of a surgical approach. Although open surgery is a definitive treatment, embolisation in selected cases has high success rates and is a safe, less invasive method.7 Additionally, a multidisciplinary study discovered that initial endovascular management of SDAVFs resulted in outcomes equal to surgery in 25% of patients.8 The safety and efficacy of endovascular embolisation suggests that it may be beneficial to offer endovascular treatment as first line, while using surgical approaches for refractory cases. Our patient presented with a rare fistula which was challenging to locate, requiring repeated imaging. We opted for endovascular embolisation for initial treatment because we could close the fistula in a minimally invasive manner during the imaging procedure itself. This could potentially eliminate the need for additional hospital visits and invasive surgery.

Learning points

  • Spinal dural arteriovenous fistulas are challenging to diagnose due to the often slowly progressive, non-specific symptoms.

  • Awareness of rare fistula locations, such as those arising from pelvic regions, are imperative to ensure accurate diagnosis and treatment.

  • Mechanical pump injection can be a useful adjunct to hand injection to detect a spinal dural arteriovenous fistulas arising from the pelvis.

Footnotes

  • Contributors MHP was responsible for compiling the patient information and writing the majority of the case report. RS was involved in conducting the care for the patient and provided technical assistance and interpretation of data. ST was involved in conducting the care for the patient and provided interpretation of the case along with technical assistance. GPS provided guidance and oversight for the project along with analysis of the case and interpretation. RS, ST and GPS together conceived the idea for the case report.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent for publication Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

References

Use of this content is subject to our disclaimer