Article Text

Download PDFPDF

Standard and Guidelines: Intracranial Dural Arteriovenous Shunts
  1. Seon-Kyu Lee1,
  2. Steven W Hetts2,
  3. Van Halbach2,
  4. Karel terBrugge3,
  5. Sameer A Ansari4,
  6. Barb Albani5,
  7. Todd Abruzzo6,
  8. Adam Arthur7,
  9. Michael J Alexander7,
  10. Felipe C Albuquerque8,
  11. Blaise Baxter9,
  12. Ketan R Bulsara10,
  13. Michael Chen11,
  14. Josser E Delgado Almandoz12,
  15. Justin F Fraser13,
  16. Don Frei14,
  17. Chirag D Gandhi15,
  18. Don Heck16,
  19. Muhammad Shazam Hussain17,
  20. Michael Kelly18,
  21. Richard Klucznik19,
  22. Thabele Leslie-Mazwi20,
  23. Ryan A McTaggart21,
  24. Philip M Meyers22,
  25. Athos Patsalides23,
  26. Charles Prestigiacomo24,
  27. G Lee Pride25,
  28. Robert Starke26,
  29. Peter Sunenshine27,
  30. Peter Rasmussen28,
  31. Mahesh V Jayaraman29,
  32. on behalf of the Standard and Guidelines Committee for the Society of Neurointerventional Surgery
  1. 1Department of Radiology, the University of Chicago, USA
  2. 2Department of Radiology, University of California at San Francisco, USA
  3. 3Department of Medical Imaging, University of Toronto, Canada
  4. 4Department of Radiology, Northwestern University, USA
  5. 5Department of Radiology, Christiana Healthcare System, USA
  6. 6Department of Neurosurgery, University of Cincinnati, USA
  7. 7Neurosurgery, Cedars-Sinai Medical Center, Los Angeles, California, USA
  8. 8Division of Neurological Surgery, Barrow Neurological Institute, Phoenix, Arizona, USA
  9. 9Radiology, Erlanger Medical Center, Chattanooga, Tennessee, USA
  10. 10Neurosurgery, Yale University School of Medicine, New Haven, Connecticut, USA
  11. 11Neurological Sciences, Rush University Medical Center, Chicago, Illinois, USA
  12. 12Interventional Neuroradiology, Abbott Northwestern Hospital, Minneapolis, Minnesota, USA
  13. 13Neurological Surgery, University of Kentucky, Lexington, Kentucky, USA
  14. 14Interventional Neuroradiology, Radiology Imaging Associates, Englewood, Colorado, USA
  15. 15Neurological Institute of New Jersey, New Jersey Medical School, Newark, New Jersey, USA
  16. 16Radiology, Forsyth Medical Center, Forsyth Radiological Associates, Winston Salem, North Carolina, USA
  17. 17Cleveland Clinic Stroke Program, Cleveland Clinic, Cleveland Heights, Ohio, USA
  18. 18Neurosurgery, Royal University Hospital, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
  19. 19Neuroradiology, Methodist Hospital, Houston, Texas, USA
  20. 20Neurointerventional Service, Massachusetts General Hospital, Boston, Massachusetts, USA
  21. 21Neurosurgery, Cleveland Clinic Florida, Weston, Florida, USA
  22. 22Radiology and Neurological Surgery, Columbia University, New York, New York, USA
  23. 23Department of Neurological Surgery, New York Presbyterian Hospital, Weill Cornell Medical College, New York, New York, USA
  24. 24Neurological Surgery, New Jersey Medical School, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, USA
  25. 25Neuroradiology, UT Southwestern, Dallas, Texas, USA
  26. 26Neurological Surgery, University of Virginia, Charlottesville, Virginia, USA
  27. 27Department of Radiology, Banner Univeristy Medical Center, Phoenix, Arizona, USA
  28. 28Neurosurgery Department, Cleveland Clinic, Cleveland, Ohio, USA
  29. 29Warren Alpert School of Medical at Brown University, Providence, Rhode Island, USA
  1. Correspondence to Dr Seon-Kyu Lee, Department of Radiology, Univeristy of Chicago, 5801 S Ellis Ave, Chicago, IL 60637, USA; slee{at}

Statistics from


Intracranial dural arteriovenous shunts (DAVS), also known as dural arteriovenous fistulas or dural arteriovenous malformations, are abnormal connections between dural (and occasionally pial) arteries and the veno vasora within the dura mater, comprising the walls of the dural sinuses, the leptomeningeal (bridging) veins, or the transosseous emissary veins within or adjacent to the dura mater. DAVS are rare, accounting for about 5–15% of intracranial vascular malformations.1 ,2 Although incompletely understood, DAVS are thought to be acquired lesions3 resulting from dural sinus or cortical venous thrombosis, possibly precipitated by hormonal changes, hypercoagulability states, trauma, or a combination of these factors.4–6 The clinical implications of intracranial DAVS are directly associated with its venous drainage pattern. For example, intracranial DAVS can cause either intracranial hemorrhages or non-hemorrhagic neurologic events such as regional or global venous congestive encephalopathy.

The aims of this document are (1) to review existing knowledge about the natural history, diagnostic methodology, and treatment modalities/techniques for DAVS; and (2) to provide recommendations on management strategies for intracranial DAVS using evidence-based medicine approaches when possible but, of necessity, relying frequently on expert opinion concerning this rare disease. Recommendations follow the American College of Cardiology/American Heart Association (ACC/AHA) Classification of Recommendation/Level of Evidence (COR/LOE) and the definition of classes and levels of evidence used in the AHA/American Stroke Association (AHA/ASA) recommendations (tables 1 and 2).

Table 1

ACC/AHA Classification of Recommendations and Level of Evidence (COR/LOE)

Table 2

Definition of classes and levels of evidence used in AHA/ASA recommendations

Natural history

Given the rarity of DAVS and the challenge of diagnosing them with non-invasive tools, the natural history of intracranial DAVS is not completely understood. There are few data on the progressive enlarge of DAVS over time with respect to recruitment or enlargement of arterial feeders or appearance of de novo fistulas over time. Nonetheless, it appears that the location and pattern of venous drainage are the major determinants of the natural history of DAVS.7–12 The location of venous drainage is likely to be related to presenting symptoms.7 ,13 ,14 The pattern of venous drainage—especially the presence of angiographic cortical venous reflux (CVR) or drainage:arterial flow of DAVS shunts into cortical vein(s) directly and/or through adjacent dural sinus—correlates with an aggressive clinical course such as intracranial hemorrhage, seizures, or venous congestive encephalopathy including venous ischemia and infarct.15 In addition to CVR, male gender, older age at presentation, posterior fossa location, and focal neurologic deficits are all independently associated with intracranial hemorrhage.16

The pattern of venous drainage is determined by local availability of either cortical veins or dural sinuses, and this may be altered by local venous thrombosis.

Cortical Venous Reflux/Drainage

There are limited studies available evaluating the natural history of intracranial DAVS without CVR; however, existing evidence and clinical experience indicates that the clinical course of intracranial DAVS without CVR is usually non-aggressive. Satomi et al9 reported that 1% of 68 patients who did not have CVR developed intracranial hemorrhage and no patient developed non-hemorrhagic neurologic deficits, with a mean follow-up period of 27.9 months.

Intracranial DAVS with CVR or drainage appear to have an aggressive natural history. Duffau et al12 reported 20 patients with cortical venous drainage who presented with intracerebral hemorrhage; 35% of these patients experienced rebleeding within 2 weeks of the original hemorrhage. van Dijk et al8 reported a 15% annual rate of intracranial events (8.1% annual risk of intracranial hemorrhage, 6.9% risk of non-hemorrhagic neurologic deficits) and 10.4% mortality rate in patients with DAVS with CVR. This study was based on 20 patients with a mean follow-up time of 4.3 years. A more recent study of untreated DAVS with CVR found an overall annual incidence of hemorrhage of 13% following diagnosis, whereas reflux with venous ectasia had an annual incidence of hemorrhage of 27% and reflux without ectasia had an annual incidence of hemorrhage of 3.5%.17

Recently, some authors have suggested that angiographically demonstrated CVR itself may not be the only risk factor for future intracranial hemorrhage, emphasizing the importance of presentation with intracranial hemorrhage.

In a cohort of 83 patients with DAVS with CVR, Söderman et al10 showed a 7.4% annual risk of rebleeding in patients who presented with hemorrhage but only a 1.5% annual risk of bleeding in patients who had incidentally discovered asymptomatic lesions. Based on 28 patients, Strom and colleagues11 also reported a 7.6% annual hemorrhage risk for patients with DAVS who presented with intracranial hemorrhage or neurologic deficits (presumably due to cortical venous hypertension) compared with a 1.4% risk in patients who presented without intracranial hemorrhage or neurologic deficit. They also showed that patients who presented with cortical venous hypertension (congestive encephalopathy) had an 11.4% annual risk of neurologic deficit compared with 0% in those who presented without CVR.

Conversion of Venous Drainage Pattern

Patients who do not show CVR on initial angiography can develop CVR over time.9 ,18 Satomi et al9 reported that 4% of patients who did not have angiographic CVR showed angiographic conversion to CVR on follow-up angiography over a mean follow-up period of 27.6 months. In this study, complete or partial resolution of an associated subjective bruit indicated conversion of a benign venous drainage pattern to an aggressive pattern (CVR) in two-thirds of cases. Kim et al18 also showed a 4% conversion rate based on the same database, but with more patients (112 patients) and a longer follow-up period (40.4 months). Shah et al19 reported a 1% annual risk of conversion from a fistula without CVR to a fistula with CVR, although in their group of 23 patients neither of the two patients who developed CVR hemorrhaged over a mean follow-up period of 5.6 years. Spontaneous closure of DAVS has also been reported.18 ,20–22 For example, Kim et al18 reported a 12.5% rate of spontaneous closure over periods ranging from 1 months to 10 years, primarily for cavernous sinus and transverse sinus DAVS. Based on reports of changing venous drainage over time, many authors have suggested that it is important to follow patients with intracranial DAVS regularly by clinical examination and that diagnostic catheter angiography should be considered when a patient's clinical symptom(s) have changed in order to assess the possibility of conversion of venous drainage pattern or spontaneous thrombosis of the DAVS, both of which can be difficult to confirm by non-invasive testing.

Clinical presentations

The development of non-invasive imaging studies such as CT angiography (CTA), MRI, and MR angiography (MRA) may raise the diagnostic possibility of clinically asymptomatic intracranial DAVS. Most patients with intracranial DAVS present with clinical symptoms. The clinical presentation of intracranial DAVS is dependent on the anatomic location and the pattern of venous drainage.13 ,23 For example, if an intracranial DAVS drains into the cavernous sinus, potential clinical presentations include conjunctival injection, chemosis, proptosis, ophthalmoplegia, decreased visual acuity, and retro-orbital pain. If an intracranial DAVS drains into the sigmoid or transverse sinuses, it may present with pulsatile tinnitus or hearing loss.7 ,13 Anterior cranial fossa DAVS are peculiar as the local venous drainage pattern is unique since there is no adjacent dural sinus. Thus, anterior cranial fossa DAVS can drain through transosseous emissary veins, thereby presenting with epistaxis, subdural hemorrhage, or intracerebral hemorrhage. DAVS involving the skull base can cause cranial neuropathies.24 Venous drainage of DAVS into the deep venous system may induce personality changes and increased intradural pressures which may result in hydrovenous disorders (hydrocephalus), memory impairment, and even dementia. The most devastating clinical presentations include intracerebral hemorrhage, seizures, and progressive focal neurologic deficits.

Imaging studies

Computerized Tomography (CT)

Non-contrast enhanced CT (NCCT) is usually the first-line imaging modality for patients who have a suspected intracranial lesion. NCCT, however, has limited sensitivity and specificity for making the diagnosis of intracranial DAVS. The spectrum of NCCT findings of intracranial DAVS includes a normal NCCT scan, engorged vascular structures which most likely are draining veins, brain parenchymal edema, subarachnoid hemorrhage, intracerebral hemorrhages, and bony resorptions. Enlarged vascular foramina crossing the skull base or calvarium can suggest a high-flow intracranial vascular lesion and may prompt further evaluation with a dedicated vascular imaging study.

Compared with NCCT, CTA can provide more specific and potentially diagnostic information. The diagnostic sensitivity of CTA for DAVS has been improved from 15.4%25 to 93% with technical developments such as the introduction of 64-detector array CT scanners.26 The CTA can provide the approximate location of the intracranial DAVS by demonstrating an enlarged draining venous structure. Advanced CT imaging techniques such as 4D CTA may also demonstrate CVR and the principal venous drainage pattern of an intracranial DAVS.26 ,27 However, CTA is still not considered a definitive diagnostic test for DAVS since its capacity to demonstrate arteriovenous shunting and characterize the venous drainage pattern is limited.

Magnetic Resonance Imaging (MRI)

MRI can demonstrate multiple imaging findings suggesting the diagnosis of DAVS: prominent and asymmetrically increased vascularity via increased signal voids, hypointensity (blooming) of draining cortical veins/sinuses on susceptibility weighted imaging due to increased deoxyhemoglobin content, parenchymal edema due to venous congestive encephalopathy, and parenchymal, subdural, or subarachnoid hemorrhage.

Time-of-flight (TOF) MRA usually shows abnormal flow-related enhancement of affected dural sinuses and associated draining cortical veins; however, it is not always easy to differentiate feeding arteries from draining veins. The diagnostic sensitivity of TOF MRA is reported to be about 50%.25 The diagnostic sensitivity of the MRA can be increased up to 93% on contrast-enhanced time-resolved MRA (tr-MRA) at 3 T. Thus, tr-MRA holds promise as a screening and follow-up imaging modality for DAVS, particularly given its non-invasive nature and lack of ionizing radiation.28 ,29 Advanced MRI techniques such as arterial spin labeling may also improve the non-invasive detection and monitoring of DAVS.30 ,31

Digital Subtraction Angiography (DSA)

Digital subtraction angiography (DSA) provides higher temporal resolution than tr-MRA, submillimeter spatial resolution, and excellent contrast resolution that cannot be matched by other current imaging modalities for the evaluation of cerebral vasculature. Thus, DSA remains the gold standard imaging modality for the diagnosis, classification, and follow-up of DAVS. In particular, in patients with DAVS with CVR, surveillance DSA might need to be performed to confirm durable occlusion of the DAVS. Some DAVS that are not apparent on non-invasive CT or MRI are only evident on DSA. Selective DSA can further clarify patho-anatomic details to elucidate clinic-angiographic correlations and establish platforms for endovascular and/or surgical treatment planning. Diagnostic cerebral angiography for patients with DAVS should be a complete selective six-vessel angiography since the clinical presentation is dependent on venous drainage of the DAVS, so the location of the shunts could be remote from the symptom location. For example, intracranial DAVS can drain into the spinal veins and produce myelopathy, or spinal DAVS can drain superiorly and cause brainstem edema and dysfunction.

Classification of DAVS

Several classification systems to describe or predict the natural history of intracranial DAVS with regard to risk of intracranial hemorrhage have been proposed.32 Borden et al33 proposed a classification system based on the angiographic characteristics of venous drainage, while Cognard et al23 presented a classification based on the clinical presentation and angiographic features. Recently, Zipfel et al34 incorporated clinical symptoms into both the Borden and Cognard classification systems and proposed a new modified classification system. A novel anatomic classification system proposed by Geibprasert et al35 emphasized the relations ship between the location of the DAVS and the local bridging transdural veins at the site of lesion formation which form the basis of clinical symptoms. The lateral epidural space was associated with male predominance and hemorrhagic presentation while the ventral epidural space was associated with female predominance and benign symptoms. Baltsavias et al36 suggested a leptomeningeal venous drainage pattern-based classification which they called ‘DES’ (Directness, Exclusivity and venous Strain). The DES system can result in eight subgroups; the directness represents the “exact location of the shunt”, the exclusivity expresses the “anatomical and functional status of the involved sinus(es)”, and the strain is the expression of the “venous decompensation”.

DAVS can simply be classified into ‘non-aggressive (type 1)’ and ‘aggressive (type 2)’ DAVS based on the presence of angiographic CVR which reflects the natural history of DAVS. In addition, each of these types can be further subdivided into ‘asymptomatic’ and ‘symptomatic’ lesions which direct the urgency of active intervention (table 3).

Table 3

Dural arteriovenous shunt classifications


The management of DAVS should be based on the predicted natural history, degree of symptomatology, patient preferences, and treatment-related risks. Considering the complexity of the management decision-making process and the potential risks related to active interventions, the management of DAVS might be best performed at a center capable of offering all treatment modalities including microsurgery, endovascular surgery, and radiosurgery. Ideally, patient management should be discussed in a multidisciplinary planning conference.

Complete and high quality selective DSA should be performed when a DAVS is suspected based on clinical presentation and/or non-invasive imaging findings to help risk stratify patients according to the expected natural history and guide treatment planning.

After obtaining a DSA, if a patient does not demonstrate angiographic CVR, conservative management can be considered since the risks of intracranial hemorrhage or of developing non-hemorrhagic neurologic deficits are likely to be low. However, if a patient does have angiographic CVR, active intervention should be strongly considered in view of the poor natural history. Additional indications for active intervention of a DAVS include—but are not limited to—secondary glaucoma that is not responsive to medical treatment or intolerable orbital pain in DAVS involving the cavernous sinus, and clinically intolerable pulsatile tinnitus in DAVS involving the transverse or sigmoid sinus.

In the presence of CVR, the goal of active intervention should be complete disconnection/obliteration of the CVR since incomplete disconnection/obliteration does not eliminate the risks of intracranial hemorrhage or the development of non-hemorrhagic neurologic deficits. Asymptomatic DAVS without CVR may not need any active intervention. However, DAVS without CVR but with clinical symptoms creating lifestyle modification or disability may warrant active intervention. In these circumstances, the goal of active intervention would be palliation or resolution of clinical symptoms not necessarily requiring complete obliteration of the DAVS. Prospectively, however, this may be difficult to assess, and complete fistula cure is favored in situations in which re-accessing a partially treated fistula could be challenging or in which there is no certainty of serial imaging follow-up to exclude progression of a partially treated lesion to a higher grade lesion with CVR.

Conservative Management

If initial complete and high quality DSA demonstrates that a patient does not have angiographic CVR, conservative management or therapy aimed at palliation of symptoms can be considered. Although there is no consensus on the clinical and/or imaging follow-up interval, 3–6-month clinical and imaging follow-up to establish the stability of the lesion and then annual clinical and non-invasive imaging follow-up have been adopted empirically by some groups.18 ,37 Especially in patients who have DAVS involving the cavernous sinus, fundoscopic examination, visual field examinations, and intraocular pressure measurements should be performed during both the initial investigation and follow-up to monitor any disease progression that may prompt considering active intervention. Any change in clinical symptoms (deterioration or improvement) warrants repeat imaging.

In addition to clinical follow-up, manual compression of the external carotid artery or one of its branches supplying the fistula can be applied intermittently in certain situations, such as patients who present with mild or asymptomatic DAVS without CVR of the transverse, sigmoid, or cavernous sinus. Higashida et al38 reported that 30% of DAVS involving the cavernous sinus showed complete closure with the intermittent external manual compression method. Compression is typically accomplished using the contralateral hand and is performed for several minutes several times a day. Potential mechanisms of the manual compression technique include reducing blood flow through external carotid artery branches such as the occipital artery, encouraging thrombosis of small or residual partially embolized fistulas, or increasing jugular vein pressure which then increases cavernous sinus pressure. However, this method has not been evaluated in a rigorous manner, and thus its efficacy has not been determined precisely. It is therefore most often used as an adjunct to other treatment methods and typically not as a sole management option for DAVS. In addition, if a patient with DAVS has CVR, manual compression should not be considered as a viable management option. DSA is not usually required for routine follow-up imaging of patients who do not show CVR on the initial DSA. However, if there is any change of clinical symptoms such as disappearance or sudden worsening of conjunctival injection or pulsatile tinnitus, or any change in non-invasive imaging findings, follow-up DSA should be considered to assess possible interval development of CVR or spontaneous regression of the DAVS.

Endovascular Treatment

Endovascular treatment is considered to be a preferred first-line option in patients with favorable anatomy when an active intervention is indicated for many types of intracranial DAVS. A thorough analysis and understanding of the anatomy, angioarchitecture, and the location of the DAVS, especially with consideration for catheter accessibility, are the main considerations in choosing endovascular treatment planning.

Endovascular treatment techniques include transarterial, transvenous, and direct puncture. The transarterial approach is often chosen as the first endovascular treatment technique for the aggressive type of DAVS associated with CVR. It requires superselective microcatheterization of the distal aspect of the feeding artery or arteries, allowing embolic agents to be passed through the fistula(s) and to reach the venous side. In some cases the transarterial approach alone can be curative; however, it can be used as an adjunct to other treatment methods including the transvenous approach, open surgery, and radiosurgery.

When there is an approachable transvenous route to reach the DAVS and it is technically feasible to catheterize veins draining the fistula, the transvenous approach is safe and effective.39 ,40 For indirect carotid cavernous fistulas, for example, transvenous access to the fistula site via the inferior petrosal sinus or superior ophthalmic vein is a standard first-line approach for endovascular management.41 Coil obliteration of the fistula site as well as the adjacent cavernous sinus via the transvenous route is generally well tolerated. Similar transvenous approaches can be used for many skull base fistulas including those of the condylar vein, marginal sinus, sigmoid sinus, transverse sinus, and superior petrosal sinus. Transvenous embolization requires careful attention to whether the venous segment to be embolized is also being used by the brain for venous outflow. If a venous segment has stasis or retrograde flow (cerebropetal), then such segments usually can be occluded safely in the context of transvenous DAVS treatment. If the brain continues to drain antegrade to a venous segment into which a DAVS drains, then flow in such segments should be preserved in order to prevent potential venous ischemia of the brain.

In some circumstances, transvenous embolization can be combined with a surgical approach with minimal exposure, especially when the diseased dural sinus or cortical vein is isolated or anatomically not possible to catheterize.42–44 For instance, direct cut-down to the superior ophthalmic vein via eyelid incision has been shown to be safe and effective.43 Transtorcular and other direct puncture techniques are essentially a subset of transvenous approaches, as they traditionally have been a means to place coils in the fistula site and venous outflow when these sites could not be accessed with catheters transvenously or transarterially. As catheter technology has improved, direct puncture approaches have become less frequent.

Endovascular Embolic Agents

Non-permanent particulate embolic agents historically including silk suture45 and, more recently, agents such as polyvinyl alcohol (PVA) particles46 can be used for palliative clinical symptom reduction or as an adjunct to other treatment modalities.47 ,48 The role of PVA as a sole embolic agent for the cure of DAVS is very limited, given the high rate of fistula recanalization after the use of particulate embolic agents alone.

Absolute alcohol has been used as a sclerosing agent to treat DAVS.49 However, the lack of controllability under fluoroscopy and its significant inflammatory response from the target organ and adjacent tissue limits its safe usage.

Liquid embolic agents including n-butyl cyanoacrylate (n-BCA) (TruFill; Codman, Raynham, Massachusetts, USA) and ethylene vinyl alcohol copolymer (EVOH) (Onyx; Covidien, Mansfield, Massachusetts, USA) are currently the main tools for transarterial endovascular treatment.

n-BCA has proved its effectiveness and stability as a liquid embolic agent for more than 30 years. When the tip of the microcatheter establishes flow arrest in a feeding artery to a fistula—an ideal situation for n-BCA injection—an interventionist can control glue delivery through the arteriovenous shunt toward the venous side to occlude the DAVS. The endovascular cure rate using n-BCA for DAVS has been reported to be up to 89.5% without permanent neurologic deficit and without mortality.50 However, relatively rapid polymerization time, challenges in controlling the injection speed, and lack of operator familiarity or experience have led n-BCA to be less frequently used than Onyx in the last decade.

Onyx is a non-adhesive liquid embolic material consisting of two subunits (ethylene and vinyl alcohol) suspended in an organic solvent, dimethyl sulfoxide (DMSO).51 Although there have been concerns about potential tissue toxicity of DMSO, increased radiation dosage due to increased procedure time, and undetermined long-term durability of the Onyx,52 it has been used more frequently than n-BCA as a liquid embolic material in the management of DAVS. Its popularity is mainly due to the relatively slower precipitating time which can facilitate a more prolonged injection with improved penetration into the venous side of the arteriovenous shunt. Complete obliteration rates of DAVS with Onyx have been reported as 80–100%, which is comparable to that of n-BCA.52 ,53 Although direct comparisons of efficacy between nBCA and Onyx are limited, one recent study suggested a lower recurrence rate of angiographically cured fistulas following Onyx embolization compared with nBCA embolization.54 However, this has to be balanced by the increased transient and permanent complication rates when Onyx is used for skull base and cavernous sinus DAVS. This may suggest that more care is needed with transarterial Onyx deposition when feeding artery(ies) of DAVS supply cranial nerves or carry a high chance of reflux toward the cranial nerve blood supply.

Coils (detachable or pushable) are the primary embolic agent for the transvenous endovascular approach since their placement is often more precise than that of liquid embolic agents. However, coils are rarely used for transarterial embolization; they are usually used as an adjunct to liquid embolic agents since delivering coils transarterially into the venous side of DAVS is often difficult or impossible unless the fistula is very large.

Open Surgical Treatment

Although most DAVS are treatable with endovascular techniques, open surgical treatment of DAVS remains important. Indications for open surgical treatment include (but are not limited to) patients requiring urgent intracranial hematoma evacuation and those with numerous and technically inaccessible arterial feeders, or feeding arteries which either supply or pass through eloquent structures such as cranial nerves or brain.55 Certain anatomic locations such as anterior cranial fossa DAVS may favor surgical intervention. The goal of open surgical treatment for DAVS is not the removal of fistulas but the disconnection or obliteration of draining vein(s) from the fistula.56 ,57 Clinical outcomes after open surgical treatment in appropriately selected patients are favorable, with some studies reporting a 10% procedure-related morbidity.56 ,57


Stereotactic radiosurgery (SRS) can take up to 2 years to have an effect on fistula closure and is thus typically not considered as the first-line treatment option for a patient who has angiographic CVR, given the high risk of interval hemorrhage.58 SRS in combination with endovascular treatment may be considered in selected cases of DAVS since embolization can provide early symptomatic relief whereas SRS offers the potential for delayed complete fistula closure.37 ,58 SRS can also be a treatment option for patients in whom endovascular or open surgical treatments are deemed to carry high procedure-related risks and in patients with residual fistulas following prior treatments. SRS has shown reasonably good obliteration rates; the complete obliteration rates of DAVS after SRS was about 68% with SRS alone and 83% in SRS combined with endovascular treatment.59 ,60 Complication rates were very low with post-SRS hemorrhage, neurologic deficit, and mortality rates of 1.2%, 1.3%, and 0.3%, respectively.60 SRS could therefore be considered as a viable option for patients with DAVS with relatively lower risks of future hemorrhage, such as patients with CVR who do not present with hemorrhage or stroke.


  1. All patients with suspected intracranial DAVS based on clinical presentation and/or non-invasive imaging findings should receive complete and high quality DSA in order to confirm and risk stratify their disease. (Class I; level of evidence C)

  2. DAVS with high risk features (eg, CVR) should be treated promptly to reduce the potential risk of intracerebral hemorrhage, venous hypertensive encephalopathy, or other neurologic events. Endovascular treatment is considered as the preferred first-line treatment option with favorable anatomy. Open surgical treatment alone or combined endovascular and open surgical treatment should be considered for high-risk fistulas not curable by endovascular means alone. SRS should be reserved as an adjunctive and/or complementary option for aggressive and symptomatic DAVS. (Class I; level of evidence C)

  3. Non-aggressive but symptomatic DAVS can be considered for definitive treatment. Endovascular treatment, open surgery, and SRS can be considered for this type of DAVS, but only if associated with very low treatment-related risk in view of the benign natural history of these lesions. (Class IIb; level of evidence C)

  4. Non-aggressive asymptomatic (ie, incidental) DAVS lesions without CVR do not warrant active intervention and, if treatment is considered, treatment-related risk versus the natural history of the disease should be thoroughly discussed between the practitioner and patient. Nonetheless, these patients should be followed both clinically and with non-invasive imaging studies in regular fashion. An exception to this recommendation would be a patient who has become asymptomatic who was previously symptomatic, as a change in symptoms can portend a venous outflow thrombosis and, hence, potential change in fistula angioarchitecture and venous drainage pattern that would warrant re-evaluation with DSA. (Class I; level of evidence C)

  5. SRS is a reasonably effective and safe treatment option. Thus, it could be considered as a viable option for DAVS that have a small compact shunt zone in patients who are not good candidates for endovascular or open surgical treatment or those who prefer a less invasive approach. (Class I; Level of Evidence C)

  6. As a rare and incompletely understood disease, intracranial DAVS warrants further scientific investigation both with regard to natural history and clinical course following treatment. Standardized reporting of angiographic and clinical features and development of multi-institutional data collection consortia would benefit our understanding and may improve clinical and surgical outcomes in the future.


A critical review of the current literature demonstrates that intracranial DAVS with CVR often follow an aggressive high-risk clinical course and warrant active intervention. On the other hand, DAVS without CVR may follow a relatively non-aggressive clinical course; thus, management can be tailored to clinical symptomatology and patient's tolerance of those symptoms. The dynamic nature of DAVS warrants regular clinical and imaging follow-up. A variety of treatment approaches exist to address DAVS but, due to the rarity of the disease, it remains challenging to generalize which are most efficacious and for what duration. The successful management of patients with DAVS requires precise and accurate angioarchitectural analysis and benefits from a comprehensive multidisciplinary approach.



  • Contributors All authors contributed to this manuscript.

  • Competing interests SWH: royalty agreement: Penumbra (ChemoFilter); consultant: Stryker Neurovascular, Silk Road Medical, Medina Medical; research grants: NIH-NCI, NIH-NIBIB, Siemens Medical. AA: investor: Lazarus Effect, Valor Medical; research support: Siemens, Sequent, Codman; consultant; Codman, Medtronic, MicroVention, Penumbra, Sequent, Silk Road, Stryker. BB: consultant/speakers bureau: Penumbra, Stryker, Medtronic, Pulsar. JEDA: consultant: Medtronic Neurovascular. DF: consultant: MicroVention, Stryker, Medtronic, Penumbra, Siemens. CP: consultant: Stryker Neurovascular, Codman Neurovascular, Edge Threrapeutics; consultant and shareholder: Thermopeutix; board member: International Brain Research Foundation. GLP: consultant: Sequent Medical. PR: investor and scientific advisory board: Perflow Medical, Blockade Medical; scientific advisory board member: Medtronic Neurovascular, Stryker Neurovascular.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles