Abstract

BACKGROUND: High-grade dural arteriovenous fistulas (dAVFs) can present shunts with very different angio-architectural characteristics. Specific hemodynamic factors may affect clinical history and determine very different clinical courses.

OBJECTIVES: To evaluate the relationship between some venous angio-architectural features in high-grade dAVFs and clinical presentation. Specific indicators of moderate or severe venous hypertension were analyzed, such as altered configurations of the dural sinuses (by a single or a dual thrombosis), or overload of cortical vessels (restrictions of outflow, pseudophlebitic cortical vessels, and venous aneurysms).

METHODS: The institutional series was retrospectively reviewed (49 cases), and the pattern of venous drainage was analyzed in relationship with clinical presentation (benign/aggressive/hemorrhage).

RESULTS: Thirty-five of 49 cases displayed cortical reflux (high-grade dAVFs). This subgroup displayed a benign presentation in 31.42% of cases, an aggressive in 31.42%, and hemorrhage in 37.14%.

CONCLUSIONS: Our data confirm that within high-grade dAVFs, 2 distinct subpopulations exist according to severity of clinical presentation. Some indicators we examined showed correlation with aggressive nonhemorrhagic manifestations (outflow restriction and pseudophlebitic cortical vessels), while other showed a correlation with hemorrhage (dual thrombosis and venous aneurysms). Current classifications appear insufficient to identify a wide range of conditions that ultimately determine the organization of the cortical venous drainage. Intermediate degrees of venous congestion correlate better with the clinical risk than the simple definition of cortical reflux. The angiographic aspects of venous drainage presented in this study may prove useful to assess dAVF hemodynamic characteristics and identify conditions at higher clinical risk.

ABBREVIATIONS

    ABBREVIATIONS
  • dAVF

    dural arteriovenous fistula

  • DSA

    digital subtraction angiography

Intracranial dural arteriovenous fistulas (dAVF) consist of a direct arteriovenous shunt, located in the dural layer, near a dural sinus. The blood supply comes from branches or dural meningeal branches of the cerebral arteries, while their drainage is carried out through the venous system of the brain (dural venous sinuses, brain leptomeningeal veins, emissary veins, or their combination).1 dAVFs are considered acquired vascular lesions and constitute about 10% of all intracranial arteriovenous shunt.2,3

Rizzoli, in 1881, was the first to describe an arteriovenous malformation, “involving the dura,” and Sachs in 1931 reported the first angiographic description of a dAVF. These consist of many small connections between branches of intracranial arteries and veins or dural venous sinuses.1 Classically, the clinical manifestations are classified into 3 major categories: (I) benign clinical presentation, (II) aggressive clinical presentation, and (III) hemorrhagic event4 (Table 1).

TABLE 1.

dAVFs Clinical Presentation

Benign presentation TinnitusHeadache Dizziness Conjunctival chemosis Exophthalmos Decreased visual acuity (not associated with glaucoma) 
Aggressive presentation Progressive neurological deficit Transitory ischemic attack Epilepsy Intracranial hypertension syndrome Cranial nerve deficits Facial nerve paralysis Glaucoma with irreversible decrease in visual acuity 
Hemorrhage  
Benign presentation TinnitusHeadache Dizziness Conjunctival chemosis Exophthalmos Decreased visual acuity (not associated with glaucoma) 
Aggressive presentation Progressive neurological deficit Transitory ischemic attack Epilepsy Intracranial hypertension syndrome Cranial nerve deficits Facial nerve paralysis Glaucoma with irreversible decrease in visual acuity 
Hemorrhage  

dAVFs’ annual risk of bleeding has been estimated between 1.5% and 1.8% per year. However, the risk of hemorrhage varies greatly according to vascular characteristics of the dAVF. It essentially depends on 2 factors: (a) the model of the venous drainage and in particular the presence of venous cortical reflux and (b) the presence or absence of aggressive symptoms at clinical presentation. The current classification of Borden and Cognard defines broad categories of venous drainage which are associated with hemorrhagic risk.5

The presence of cortical venous reflux correlates directly with the aggressiveness of the symptoms.1,2,6 dAVfs presenting with cortical venous reflux (Borden grades 2 and 3 and Cognard 2b, 2a + b, 3, and 4) are commonly called “aggressive or high-grade dAVFs,” while those without cortical venous reflux (Borden grade 1 and Cognard grades 1 and 2) are commonly referred to as “benign or low-grade dAVFs.”

In all classifications, the so-called low-grade dAVFs (Cognard grades 1-2a, and Borden grade 1) have an annual risk of bleeding of 0%, the intermediate lesions (Cognard grades 2b and 2a + b and Borden grade 2) have a risk of 6% per year, while the high-grade lesions (Cognard grades 3 and 4 and Borden grade 3) have a risk of 10% per year.2,5,7-11

Soderman et al12 have documented an increased risk of annual bleeding estimated around 7.4% in patients who have already presented a hemorrhage, while the maximum risk of rebleeding (analyzing Duffau's series) is in the 2 weeks after the first bleeding event (35% of rebleeding within the first 2 weeks after hemorrhage), although the risk of delayed rebleeding is widely described.13

The natural history of dAVFs is further complicated by the dynamic nature of the disease, characterized by potential long-term changes of venous drainage conformation.12 There are several reports of low-grade dAVFs (Cognard grades 1-2a and Borden grades 1 and 2) with tardive development of cortical venous reflux during a long period of follow-up, resulting in a change in clinical manifestations or hemorrhage.14 From the series of Satomi et al,15 it seems that the risk of developing cortical venous drainage “de novo” is approximately 1% per year in low-grade dAVFs (Cognard grades 1-2a, and Borden grades 1 and 2).

Aim of the Study

Recent case studies point out that not all high-grade dAVFs present with bleeding or aggressive symptoms, even if they are characterized by cortical venous reflux.11 It has been suggested that the angiographic demonstration of cortical venous drainage does not necessarily indicate a venous leptomeningeal hypertension.16 In addition, asymptomatic or mildly symptomatic patients with cortical venous drainage were reported to have a less aggressive clinical course compared with patients with cortical venous drainage and aggressive presentation at onset.17

These studies have highlighted that subgroups of patients with high-grade dAVFs (with cortical venous reflux) can have very different clinical course. It has been suggested that, as part of this subpopulation presenting cortical venous reflux, specific hemodynamic factors may affect the clinical history of the disease. High-grade dAVFs are a very heterogeneous group, since they can present shunts with very different angio-architectural characteristics.18,19

Hypertension in the cortical veins district is an important pathogenetic mechanism in dAVFs with significant clinical consequences. In fact, differences in venous topography, the hemodynamic effects of arterialized flow, and dynamic pathological mechanisms of dAVFs such as a dural sinus thrombosis, stenosis, or neoangiogenesis, may (a) influence the characteristics of the venous drainage, (b) be related to intracranial venous hypertension, and (c) be predictive of the risk of aggressive clinical presentation or hemorrhage.18,20 The hemodynamic effects of direct arterialization of venous blood can, thus, cause venous ectasia, tortuosity, and congestion of the cerebral veins. The presence of venous thrombosis and its position with respect to the arterialized flow direction can determine the venous drainage pattern. In addition, the interindividual anatomical variability of the venous network may also affect the impact of the flow on the underlying brain parenchyma.20-22 Thus, the simple definition of “cortical reflux,” as indicated in the current classifications, appears to be inadequate to define a wide spectrum of conditions that determine the “quality” of the venous drainage. However, despite the importance of understanding individual features of venous drainage system in high-grade dAVFs, only few papers have extensively investigated venous topography in relation with clinical course.19

The aim of this study is to evaluate the relationship between various dAVFs venous angio-architectural features in high-grade dAVFs and their relation to clinical presentation. The presence of certain indicators of moderate or severe venous hypertension (specifically particular configurations of the dural sinuses or arrangements of cortical vessels with signs of overload) may be useful in highlighting a subset of patients at risk aggressive clinical presentation or bleeding. So far, these indicators were presented in few papers and analyzed individually. Some experiences mostly focused on the leptomeningeal network arrangement,19,20 while other papers underline the significance of dural sinuses conformation.18 In this study, we aim to pull together these observations, and assess if these correlate with an aggressive clinical presentation or with hemorrhagic risk.

METHODS

The institutional series was retrospectively reviewed, and all cases with a diagnosis of dAVF treated at Institute of Neurosurgery of the Catholic University of Rome, Italy – “Agostino Gemelli” University Hospital, from January 1998 to January 2015, were data based. Institutional review board approval for the study was obtained. Patient consent for the use of medical information in an anonymous form was provided at the time of the angiographic study.

General data such as sex, age, clinical presentation, dAVF location, classification, and treatment were recorded. Only patients whose digital subtraction angiography (DSA) and comprehensive data about clinical story were available were included in the study. Cognard grade 5 cases were not included as these cases drain into cervicomedullary veins and thus represent a rather different spectrum of pathology compared with proper intracranial dAVFs.

Forty-nine patients were included in the final series. Thirty-five cases presented cortical venous reflux (Borden grades 2 and 3 or Cognard 2b, 2a + b, 3, and 4). Within this subgroup, the pattern of venous drainage was analyzed in relation to clinical presentation, and in particular has evaluated the possible correlation between indicators of venous hypertension and clinical course.

Classification According to Clinical Presentation

Clinical presentation has been classified, in accordance with the prevailing literature (Table 1), into (a) benign, (b) aggressive, and (c) hemorrhagic presentation. They were categorized as “benign presentation” cases with reversible symptoms without focal neurological deficit such as tinnitus, headache, dizziness, orbital venous congestion (conjunctival chemosis, ocular pulse, exophthalmos, decreased visual acuity not resulting in glaucoma). With “aggressive presentation,” those cases characterized by the presence of progressive/evolutive neurological deficit such as transient ischemic attack, seizures, intracranial hypertension, cranial nerve deficits, facial nerve paralysis, and glaucoma with irreversible decrease in visual acuity. All cases in which the presence of an intracranial bleeding event was demonstrated on CT were included in the group with “hemorrhagic presentation” (Table 1).

Angiographic Analysis of Venous Congestion Indicators

DSA was carefully reviewed in all high-grade dVAFs (with cortical reflux). Each angiographic study has been evaluated together with an expert interventional neuroradiologist, with particular interest to the venous phase of angiography (in particular the delayed parenchymal and venous phase). For each case, the presence of the following venous drainage characteristics was recorded and entered into the database:

  • Dural sinus single thrombosis: defined as the occlusion of a dural sinus downstream with respect to the normal flow of the venous drainage (Figure 1).

  • Dural sinus double thrombosis (so called “isolated sinus”): those conditions which may result in an obstruction in both the proximal and distal ends of a dural sinus. Note that the sites of the dual occlusion may also be relatively distant to the point of fistula (eg, with the fistulous site on transverse sinus and dual thrombosis at the level of the sigmoid sinus and at the level of the torcular with long portion of the transverse sinus being “isolated”; Figure 2).

  • Restricted cortical venous outflow: those conditions in which venous drainage, after DSA parenchymal phase, consists solely of a single leptomeningeal vein and is generally an expression of restrictions of venous outflow by stenosis or thrombosis of a dural sinus (Figures 1, 3, and 4).

  • Pseudophlebitic appearance of cortical vessels: defined as a tortuous and serpiginous aspect of cortical veins, with a typical trend characterized by the repetition, along the course of a cortical vein, focal ectasia (not aneurysmatic), and caliber narrowings (Figures 3 and 5).

  • Venous aneurysms: focal venous dilation greater than 5 mm or 3 times larger than the diameter of the vein (Figure 3).

  • Conditions not indicative of venous failure: cortical veins dilated evenly throughout their course (including very long drainages), without signs of ectasia and of stagnation of contrast (so called “brisk drainage”). This condition, although certainly expressing a certain degree of venous hypertension, is not to be considered as an indicator of venous congestion (Figure 4).

FIGURE 1.

Example of a single sinus thrombosis; Cognard 2b; indirect cavernous fistula (red arrow point of the fistula) with early venous drainage into the superior ophthalmic vein (blue arrow) to the extremely ectatic angular and facial veins. Inferior petrosal sinus thrombosis (yellow arrow). In this case, the thrombosis realizes a condition of limited outflow as the ophthalmic vein is the only drainage of the dAVF.

FIGURE 1.

Example of a single sinus thrombosis; Cognard 2b; indirect cavernous fistula (red arrow point of the fistula) with early venous drainage into the superior ophthalmic vein (blue arrow) to the extremely ectatic angular and facial veins. Inferior petrosal sinus thrombosis (yellow arrow). In this case, the thrombosis realizes a condition of limited outflow as the ophthalmic vein is the only drainage of the dAVF.

FIGURE 2.

Example of double thrombosis determining a condition of “isolated sinus”; Cognard 2a+b; transverse-sigmoid sinus fistula by suboccipital dural arterial branches (red arrow). The stream is compartmentalized in the sinus by a dual thrombosis at the level of the torcula and at the level of the sigmoid sinus at the level of the shunt (yellow arrows). The arterialized flow is drained through the leptomeningal bridging veins (blue arrow). The flow is also diverted through the straight sinus (green arrow) that displayes an inverted flow towards the deep venous system.

FIGURE 2.

Example of double thrombosis determining a condition of “isolated sinus”; Cognard 2a+b; transverse-sigmoid sinus fistula by suboccipital dural arterial branches (red arrow). The stream is compartmentalized in the sinus by a dual thrombosis at the level of the torcula and at the level of the sigmoid sinus at the level of the shunt (yellow arrows). The arterialized flow is drained through the leptomeningal bridging veins (blue arrow). The flow is also diverted through the straight sinus (green arrow) that displayes an inverted flow towards the deep venous system.

FIGURE 3.

Example of direct duro-pial shunt, with thrombosis, venous outflow restriction (drainage determined by a single leptomeningeal vein), phlebitis, venous aneurysm, and pseudophlebitic appearance of distal cortical veins; Cognard 4; injection of the external carotid artery (above), internal (center), and tardive venous phase (below). The dAVF is supplied by tentorial dural branches of the internal carotid artery and middle meningeal artery branches (red arrow). Venous drainage is determined by a single vein (green arrow) with aneurysmal dilatation along the course of the draining vein and pseudophlebitic appearance of distal vessels (blue arrows). Drainage continues toward ectatic ipsilateral basal veins to the Galen region, straight sinus, and internal cerebral veins. Venous reflux at the sinus shows a clear cut (yellow arrow) consistent with thrombosis of the deep venous circulation.

FIGURE 3.

Example of direct duro-pial shunt, with thrombosis, venous outflow restriction (drainage determined by a single leptomeningeal vein), phlebitis, venous aneurysm, and pseudophlebitic appearance of distal cortical veins; Cognard 4; injection of the external carotid artery (above), internal (center), and tardive venous phase (below). The dAVF is supplied by tentorial dural branches of the internal carotid artery and middle meningeal artery branches (red arrow). Venous drainage is determined by a single vein (green arrow) with aneurysmal dilatation along the course of the draining vein and pseudophlebitic appearance of distal vessels (blue arrows). Drainage continues toward ectatic ipsilateral basal veins to the Galen region, straight sinus, and internal cerebral veins. Venous reflux at the sinus shows a clear cut (yellow arrow) consistent with thrombosis of the deep venous circulation.

FIGURE 3.

Example of direct duro-pial shunt, with thrombosis, venous outflow restriction (drainage determined by a single leptomeningeal vein), phlebitis, venous aneurysm, and pseudophlebitic appearance of distal cortical veins; Cognard 4; injection of the external carotid artery (above), internal (center), and tardive venous phase (below). The dAVF is supplied by tentorial dural branches of the internal carotid artery and middle meningeal artery branches (red arrow). Venous drainage is determined by a single vein (green arrow) with aneurysmal dilatation along the course of the draining vein and pseudophlebitic appearance of distal vessels (blue arrows). Drainage continues toward ectatic ipsilateral basal veins to the Galen region, straight sinus, and internal cerebral veins. Venous reflux at the sinus shows a clear cut (yellow arrow) consistent with thrombosis of the deep venous circulation.

FIGURE 3.

Example of direct duro-pial shunt, with thrombosis, venous outflow restriction (drainage determined by a single leptomeningeal vein), phlebitis, venous aneurysm, and pseudophlebitic appearance of distal cortical veins; Cognard 4; injection of the external carotid artery (above), internal (center), and tardive venous phase (below). The dAVF is supplied by tentorial dural branches of the internal carotid artery and middle meningeal artery branches (red arrow). Venous drainage is determined by a single vein (green arrow) with aneurysmal dilatation along the course of the draining vein and pseudophlebitic appearance of distal vessels (blue arrows). Drainage continues toward ectatic ipsilateral basal veins to the Galen region, straight sinus, and internal cerebral veins. Venous reflux at the sinus shows a clear cut (yellow arrow) consistent with thrombosis of the deep venous circulation.

FIGURE 4.

Example of venous drainage without signs of strain, realized by a single cortical vein; Cognard 3; dAVF’s point of fistula is on the greater sphenoid wing (red arrows) with feeders from the middle meningeal artery, and deep temporal and anterior ethmoid terminal branches of the internal maxillary artery. The venous drainage is realized by a single long temporoparietal pial vein uniformly ectatic, with brisk contrast flow toward the superior sagittal sinus (blue arrows).

FIGURE 4.

Example of venous drainage without signs of strain, realized by a single cortical vein; Cognard 3; dAVF’s point of fistula is on the greater sphenoid wing (red arrows) with feeders from the middle meningeal artery, and deep temporal and anterior ethmoid terminal branches of the internal maxillary artery. The venous drainage is realized by a single long temporoparietal pial vein uniformly ectatic, with brisk contrast flow toward the superior sagittal sinus (blue arrows).

FIGURE 5.

Tentorial edge dAVF; Cognard 4; drainage is realized by multiple veins (blue and green arrows); some cortical veins show signs of strain with pseudophlebitic appearance (green arrows). Multiple fistulous points are indicated with red arrows.

FIGURE 5.

Tentorial edge dAVF; Cognard 4; drainage is realized by multiple veins (blue and green arrows); some cortical veins show signs of strain with pseudophlebitic appearance (green arrows). Multiple fistulous points are indicated with red arrows.

The above aspects of venous drainage may be present simultaneously in a single case.

Statistical Analysis

A χ2 test was used to assess statistically significant differences between clinical presentation in relation to each pattern of venous drainage. The definition of risk for each specific pattern, in relation to the manifestation of symptoms defined as aggressive or benign or upon the occurrence of specific symptoms such as bleeding, intracranial hypertension, and epilepsy, was confirmed by a logistic regression model and Pearson's test. Quantitative data are presented as mean (min–max), while the qualitative data are resented in %. Data analysis was performed using the statistical software SPSS 12.0 for Windows (IBM, Armonk, New York). Statistical significance was set at P = .05.

RESULTS

From the institutional series, 49 cases were selected: 28 males and 21 females (mean age 56.75, range 20-84 years). Fourteen cases (28.57%) were cavernous sinus dAVFs, 7 (14.28%) were of the anterior cranial fossa, 4 (8.16%) of the posterior fossa, 1 (2.04%) of foramen magnum, 6 (12.24%) of the superior sagittal sinus, 3 (6.12%) of the tentorium, and 14 (28.57%) of the transverse/sigmoid sinus (Table 2).

TABLE 2.

General Characteristics of the Studied Population

Demographic No of Patients 49  
Characteristics m/f 28/21  
 age 56,75 (20-84)  
Anatomic location CS 14 29% 
 ACF 14.28% 
 PCF 8.16% 
 FM 2.04% 
 SS 12.24% 
 TEN 6.12% 
 TS 14 28.57% 
Clinical presentation Benign presentation 25 51.02% 
 Aggressive presentation 11 22.45% 
 Hemorrhage 13 26.53% 
Classification Cognard 1 12.24% 
 Cognard 2a 16.32% 
 Cognard 2b 18.36% 
 Cognard 2a+b 18.36% 
 Cognard 3 14.28% 
 Cognard 4 10 20.40% 
 Borden 1 14 28.57% 
 Borden 2 14 28.57% 
 Borden 3 21 42.85% 
Demographic No of Patients 49  
Characteristics m/f 28/21  
 age 56,75 (20-84)  
Anatomic location CS 14 29% 
 ACF 14.28% 
 PCF 8.16% 
 FM 2.04% 
 SS 12.24% 
 TEN 6.12% 
 TS 14 28.57% 
Clinical presentation Benign presentation 25 51.02% 
 Aggressive presentation 11 22.45% 
 Hemorrhage 13 26.53% 
Classification Cognard 1 12.24% 
 Cognard 2a 16.32% 
 Cognard 2b 18.36% 
 Cognard 2a+b 18.36% 
 Cognard 3 14.28% 
 Cognard 4 10 20.40% 
 Borden 1 14 28.57% 
 Borden 2 14 28.57% 
 Borden 3 21 42.85% 

CS: cavernous sinus; ACF: anterior cranial fossa; PCF: posterior cranial fossa; FM: foramen magnum; SS: superior sagittal sinus; TEN: tentorium; TS: transverse-sigmoid sinus.

According to Cognard classification, 6 cases were Cognard 1 (12.24%), 8 cases Cognard 2a (16.32%), 9 cases Cognard 2b (18.36%), 9 cases Cognard 2a + b (18.36%), 7 cases Cognard 3 (14.28%), and 10 cases Cognard 4 (20.40%). According to Borden classification, they were classified as follows: 14 cases Borden grade 1 (28.57%), 14 cases Borden grade 2 (28.57%), and 21 cases Borden grade 3 (42.85%; Table 2).

The clinical manifestations were classified as benign presentation in 25 cases (51.02%), aggressive presentation in 11 cases (22.45%), and hemorrhagic presentation in 13 cases (26.53%). Of the 25 cases with benign presentation, 6 patients (24.0%) showed auditory symptoms (tinnitus), 8 (32.0%) showed signs of congestion or eye drop of visual acuity not linked to glaucoma, 7 (28.0%) headache, 4 (16.0%) were incidental findings. Of the 11 cases with aggressive presentation in 2 (18.18%) was documented an intracranial hypertension syndrome, in 6 (54.54%) evolving cranial nerve deficits, and in 4 cases (36.36%) epilepsy (Table 3).

TABLE 3.

Clinical Presentation, and Their Subdivision in Benign, Aggressive, and Hemorrhage

Benign presentation (25 cases, 51.01% of the entire sample)   
 Uditive 24.0% 
 Ocular 32.0% 
 Headache 28.0% 
 Incidental 16.0% 
Aggressive presentation (11 cases, 22.45% of the entire sample)   
 Intracranial hypertension 16.67% 
 Cranial nerve palsies 58.33% 
 Epilepsy 33.34% 
Hemorrhage (13 cases, 26.53% of the entire sample)   
Benign presentation (25 cases, 51.01% of the entire sample)   
 Uditive 24.0% 
 Ocular 32.0% 
 Headache 28.0% 
 Incidental 16.0% 
Aggressive presentation (11 cases, 22.45% of the entire sample)   
 Intracranial hypertension 16.67% 
 Cranial nerve palsies 58.33% 
 Epilepsy 33.34% 
Hemorrhage (13 cases, 26.53% of the entire sample)   

Regarding the clinical presentation with respect to the anatomical site, among the 14 cavernous sinus dAVFs, 8 (57.14%) had a benign clinical presentation, 6 (42.86%) an aggressive, and 0 (0%) a hemorrhagic presentation; among the 7 anterior cranial fossa dAVFs, 4 (57.14%) had a benign clinical presentation, 0 (0%) aggressive, and 3 (42.85%) hemorrhagic; among the 4 posterior fossa dAVFs, 2 (50%) had a benign clinical presentation, 0 (0%) aggressive, 2 (50%) hemorrhagic course; the single case of the foramen magnum presented a clinical manifestations classified as benign; among the 6 cases at the superior sagittal sinus, 3 (50%) had a benign clinical presentation, 2 (33.34%) aggressive, 1 (16.67%) hemorrhagic; among the 3 tentorial dAVFs, 0 (0%) showed a benign course, 2 (66.67%) an aggressive one, 1 (33.34%) a hemorrhagic event; among the 14 dAVFs belonging to the transverse/sigmoid sinus, 7 (50%) had a benign clinical presentation, 2 (14.28%) an aggressive, and 5 (35.71%) a hemorrhagic event (Table 4).

TABLE 4.

Clinical Presentation According to the dAVFs Anatomical Site

 Benign Presentation Aggressive Presentation Hemorrhage 
CS (14) 8 (57.14%) 6 (42.86%) 0 (0%) 
ACF (7) 4 (57.14%) 0 (0%) 3 (42.85%) 
PCF (4) 2 (50%) 0 (0%) 2 (50%) 
FM (1) 1 (100%) 0 (0%) 0 (0%) 
SS(6) 3 (50%) 2 (33.34%) 1 (16.67%) 
TEN (3) 0 (0%) 2 (66.67%) 1 (33.34%) 
TS (14) 7(50%) 2(14.28%) 5 (35.71%) 
 Benign Presentation Aggressive Presentation Hemorrhage 
CS (14) 8 (57.14%) 6 (42.86%) 0 (0%) 
ACF (7) 4 (57.14%) 0 (0%) 3 (42.85%) 
PCF (4) 2 (50%) 0 (0%) 2 (50%) 
FM (1) 1 (100%) 0 (0%) 0 (0%) 
SS(6) 3 (50%) 2 (33.34%) 1 (16.67%) 
TEN (3) 0 (0%) 2 (66.67%) 1 (33.34%) 
TS (14) 7(50%) 2(14.28%) 5 (35.71%) 

CS: cavernous sinus; ACF: anterior cranial fossa; PCF: posterior cranial fossa; FM: foramen magnum; SS: superior sagittal sinus; TEN: tentorium; TS: transverse-sigmoid sinus.

Subgroup With Cortical Venous Reflux

Analyzing the entire sample, 14 out of 49 cases (28.57% of the entire sample) did not show signs of cortical venous reflux (classified as Cognard 1 and 2a or Borden grade 1). Of these, 13 showed a benign presentation, an aggressive presentation in 1, and none had a hemorrhagic presentation.

Thirty-five cases (71.43% of the entire sample) showed, instead, signs of cortical venous reflux (classified as Cognard 2b, 2a + b, 3, 4 or Borden grades 2 and 3). Of these, 11 had a benign onset (31.42%), aggressive in 11 (31.42%), and hemorrhagic in 13 cases (37.14%).

The angio-architectural angiographic parameters were analyzed within the subgroup displaying cortical venous reflux; in particular, the observation of a “single sinus thrombosis” was recorded 8 times: in a case related to benign symptoms, in 4 cases with aggressive symptoms, and in 3 with hemorrhage. A condition of “double sinus thrombosis” (so called “isolated sinus”) was observed 7 times: in a case related to benign symptoms, in a case with aggressive symptoms, and in 5 with hemorrhage. The condition of “restricted cortical venous outflow” was observed 20 times: in 4 cases associated with benign symptoms, in 8 with aggressive symptoms, and in 9 hemorrhagic presentations. A “pseudophlebitic appearance of cortical venous drainage” was observed in 26 cases: in 4 cases associated with benign presentation, in 11 with an aggressive presentation, and in 12 associated with hemorrhage. The presence of venous aneurysms has been documented 12 times: 1 time associated with a benign presentation, 2 times with an aggressive one, and 8 times with a hemorrhage (Table 5 and Figure 6).

FIGURE 6.

Diagrammatic representation of clinical presentations in patients without cortical reflux (Cognard 1 and 2a) and in patients with cortical reflux (Cognard 2b, 2a + b, 3, and 4).

FIGURE 6.

Diagrammatic representation of clinical presentations in patients without cortical reflux (Cognard 1 and 2a) and in patients with cortical reflux (Cognard 2b, 2a + b, 3, and 4).

TABLE 5.

Distribution of Clinical Presentations in Patients Without Cortical Reflux (Cognard 1 and 2a) and in Patients With Cortical Reflux (Cognard 2b, 2a + b, 3, and 4)

 Benign Presentation Aggressive Presentation Hemorrhage Total 
Absence of CVR (Cognard I e 2a) 13 (92.86%) 1 (7.14%) 0 (0%) 14 (28.57%) 
Presence of CVR (Cognard 2b, 2a+b, 3, 4) 13 (31.42%) 11 (31.42%) 13 (37.14%) 35 (71.43%) 
 Benign Presentation Aggressive Presentation Hemorrhage Total 
Absence of CVR (Cognard I e 2a) 13 (92.86%) 1 (7.14%) 0 (0%) 14 (28.57%) 
Presence of CVR (Cognard 2b, 2a+b, 3, 4) 13 (31.42%) 11 (31.42%) 13 (37.14%) 35 (71.43%) 

CVR: cortical veins reflux.

Analysis of Correlation and Risk Assessment

A correlation analysis was carried out between the various patterns examined: (a) single thrombosis of a dural sinus, (b) double thrombosis of a dural sinus (isolated sinus), (c) restricted cortical venous outflow, (d) pseudophlebitic appearance, and (e) venous aneurysms, with clinical presentation, which are divided into (I) benign, (II) aggressive, and (III) hemorrhage. The evaluation of the association was assessed by the χ2 test and logistic regression analysis with estimation of risk expressed as odds ratio (Table 6).

TABLE 6.

Association Among Examined Patterns, Clinical Presentation, and Risk Estimation

Venous Drainage Pattern Correlation Risk (odd ratio) P 
Single thrombosis of a dural sinus - - .608 
Double thrombosis of a dural sinus (isolated sinus) Hemorrhage + 525% .048* 
Outflow restriction pattern Aggressive presentation + 350% .029* 
Pseudophlebitic appearance of cortical vessels Aggressive presentation + 420% .002** 
Venous aneurysms Hemorrhage + 620% .013** 
Venous Drainage Pattern Correlation Risk (odd ratio) P 
Single thrombosis of a dural sinus - - .608 
Double thrombosis of a dural sinus (isolated sinus) Hemorrhage + 525% .048* 
Outflow restriction pattern Aggressive presentation + 350% .029* 
Pseudophlebitic appearance of cortical vessels Aggressive presentation + 420% .002** 
Venous aneurysms Hemorrhage + 620% .013** 

*statistical significance; **very strong statistical significance.

The single thrombosis of a dural sinus did not show a statistically significant correlation with the benign/aggressive clinical onset and with bleeding. Even regarding specific symptoms such as intracranial hypertension and epilepsy, a significant association was not identified.

For the condition of double thrombosis of a dural sinus (isolated sinus), it was not found a statistically significant correlation with benign/aggressive presentation. It does, however, present a correlation (to the limits of statistical significance P = .048) with the hemorrhagic event, with an increased risk of bleeding estimated around 5 times (+525%). There was no statistically significant association found with epilepsy and intracranial hypertension.

For the restricted outflow pattern, a negative association with benign symptoms (P = .19) and a significant positive association with aggressive symptoms (P = .029) with an increased risk of +350% were observed. There is no association with bleeding or with specific symptoms such as epilepsy and intracranial hypertension.

For the “pseudophlebitic appearance of cortical vessels,” a very strong statistically significant negative correlation with the benign symptoms (P = .002) was found. There is also an association with aggressive symptoms, with a very strong significance (P = .002); the estimated risk of an aggressive course is +420%. There was no significant correlation with the hemorrhagic event. There is also a trend of association, although it does not reach the limits of statistical significance, with the epileptic manifestations (3/4 cases of the whole series).

With regard to venous aneurysms, there exists a significant correlation only with hemorrhagic manifestations (P = .013) with an estimate of bleeding risk of +620%.

Data on associations between different patterns and the clinical presentation with the estimated risk (odds ratio) are summarized in Table 6, which shows the significance of the association according to the Pearson test.

DISCUSSION

Two Distinct Populations in High-Grade dAVFs

The presence of cortical venous reflux is a well-known risk factor for aggressive clinical presentation and for a worse clinical course in patients affected by dAVFs.18 However, recent studies have shown that, despite the presence of cortical venous reflux, the clinical course of these patients can be very different. These reports suggest that the initial presentation may be an important predictor of the natural history of the disease.5,16 dAVFs with cortical venous reflux (therefore, Borden grades 2 and 3 and Cognard 2b, 2a + b, 3, and 4) represent a broad spectrum of pathology and essentially include the majority of cases of dAVFs.

In the series presented in this study, as many as 35 out of 49 cases (73.49%) presented some degree of cortical venous drainage. However, only a part presented aggressive clinical manifestation (31.43%) or bleeding (37.14%), while more than a third of cases (31.42%) had a benign clinical course (Table 5 and Figure 6).

As confirmed by our data, it is possible to identify, within dAVFs with cortical venous reflux, 2 distinct subpopulations according to the severity of the clinical presentation. The presence of cortical venous reflux appears insufficient to define a complex set of conditions that determine the organization of the venous drainage. Venous leptomeningeal system, even if congested, may show or not signs of strain. In addition, the variability of the venous anastomotic network and the different types of organization, depending on the anatomical region concerned, may determine an extreme variability of the impact of the flow on the underlying brain parenchyma and highlight a very different spectrum of clinical presentations.

Venous Angio-Architectural Features and Clinical Presentation

The classic model of “distortion” of the venous system only refers to venous ectasia and the presence of venous aneurysms (Cognard grades 3 and 4). The other extreme is a significant, but otherwise uniform, venous dilation, without focal or aneurysmal ectasia. This, even if it corresponds to a certain degree of overload, cannot be considered properly a sign of strain of the cortical venous system. In other words, there is a spectrum of “grays” between the 2 above-mentioned examples, intermediate conditions that have different clinical courses.

We shifted the focus on other aspects of venous drainage in high-grade dAVfs, analyzing that “gray area's” angio-architectural features in order to discriminate different levels and different degrees of venous strain and the correspondent clinical risk.

As recently pointed out in the recent papers by Baltasavias et al19-24 and Shin et al18, which currently represent the only 2 experiences published in the literature on this topic, other factors may be significant in the angiographic analysis of patients with cortical venous reflux. Following these experiences, we examined both leptomeningeal aspects (such as restricted venous outflow and pseudophlebitic appearance of cortical veins) and specific configurations of the dural sinuses determined by thrombosis (single thrombosis and isolated sinus). Our results confirm that among high-grade dAVFs there are significant differences in venous drainage arrangement that reflect to significant differences in clinical presentation.

Regarding the thrombosis of a sinus, a statistically significant correlation with the clinical presentation was not found. The relationship between dAVFs and thrombosis is certainly strict and the observation of a thrombosis in a dAVFs is a common event, as the fistula itself is most probably a delayed event following a sinus thrombosis. Even if it is by itself not a risk factor for an aggressive course, it might be a predisposing event to a restriction of the venous outflow through a limited cortical territory (restricted cortical venous outflow) or even compartmentalize an area within the dural sinus (double thrombosis or isolated sinus).

The restricted cortical venous outflow is characterized by a single leptomeningeal vein draining the fistula. It is a concept that well describes the absence of a collateral and anastomotic network able to redistribute the arterialized flow through alternative outputs. It embodies a significant overload of the system in a restricted brain area, and it is associated almost systematically with a pseudophlebtic aspect of local cortical veins. This aspect correlates with an aggressive clinical presentation, with an increased risk of +350%.

Particularly interesting are the data on the “isolated sinus” (by a dual thrombosis within a dural sinus). First, it should be noticed that this is not an uncommon phenomenon (seen in 7/35 cases, 20%) and in about half of the observations it was identified at the transverse-sigmoid sinus. The presence of a double sinus thrombosis compartmentalizes a territory at high pressure that diverts the flow through one or more bridging veins to the cortical venous system. De facto, a direct shunt with cortical veins exists even if a dural sinus is interposed. This possibility is not formally recognized by current classifications and it is a strong determinant of a hemorrhagic clinical presentation, especially in the transverse/sigmoid district. Our data confirm a correlation (P = .048) with the hemorrhagic event with an increased risk of about 5 times (+525%). The presence of an isolated sinus is also a predisposing event to conditions of venous outflow restriction, especially if the sinus isolated territory is not large and, therefore has a unique outflow through a single bridging vein. The cause of the development of an isolated sinus is poorly understood. It may be the result of sinus thrombosis due to the turbulent flow in its proximal and distal ends or may be due to partial recanalization of a thrombosed sinus.

For the pseudophlebitic appearance of cortical veins, a positive association with aggressive symptoms was found, without an increase in bleeding risk. This anatomo-functional picture was observed relatively frequently, 27/35 cases (77.14%). Our data show that patients with cortical venous reflux that do not show such a feature display a rate of aggressive presentation markedly lower than those who show this aspect. Even if at the limit of statistical significance, it should be noted that such aspect was identified in 3/4 of epilepsy cases. The venous tension is probably a strong determinant toward an aggressive clinical behavior. This phenomenon is quite reasonable, since the overload of the leptomeningeal venous system is conversely an expression of cerebral parenchyma distress. In other words, cortical venous strain can be focal, expressed by venous aneurysm, or more regional and functional, expressed by venous congestion and stagnation manifested by the appearance of several pseudophlebitic vessels within a relatively large parenchymal area. While venous aneurysms directly correlate with bleeding risk (with a strong statistical significance P = .013 and a risk of +620%), the pseudophlebitic appearance of cortical veins is mainly correlated with aggressive but nonhemorrhagic manifestations (strong significance P = .002 and risk of +420%).

Anatomical Site and Clinical Behavior

In our series, we observed different behavior in relation to different anatomical sites of the cases examined. A clear tendency toward a benign presentation was observed in cases of cavernous sinus dAVFs, which is, in agreement with the literature, an anatomical site generally associated with a relatively positive clinical course.2 No case of bleeding was observed and more than half of patients had a benign clinical presentation (conjunctival chemosis, ocular pulse, exophthalmos, decreased visual acuity not associated with glaucoma). These data are related to the venous anatomy of the cavernous sinus which presents multiple drainage options (ophthalmic veins, petrosal sinus, pterygoid plexus, petrosal sinus, intercavernous sinuses, middle cerebral vein, uncal, and prepontine veins).21 Interestingly, the review of the single cases of our series demonstrated that a benign clinical presentation is generally observed when multiple venous outputs are functional, while in 4 out of the 6 cases with aggressive presentation it was possible to observe a restriction of venous drainage, exclusively through the ophthalmic vein, for the presence of associated thrombosis. In these 4 patients, the clinical presentation was related to ocular hypertension, with secondary glaucoma and reduced visual acuity, often associated with cranial nerve palsies. However, even if the small sample size does not allow to demonstrate a statistically significant relationship between the 2 events, there is definitely an association trend among presence of aggressive ocular symptoms and drainage exclusively through the ophthalmic vein (Figure 7).

FIGURE 7.

Comparison between 2 different drainage patterns in cavernous sinus dAVFs (red arrows). On the left, drainage is supplied by multiple venous outputs, particularly from the ophthalmic vein (blue arrows), the uncal vein and superior petrosal sinus (green arrows). On the right, because of a petrous sinus thrombosis (yellow arrow), the only possible outflow is through the ophthalmic vein. The first cases (left) showed only signs of ocular congestion, the second case (right) showed a reduction in visual acuity, ocular hypertension, and retinal hemorrhages.

FIGURE 7.

Comparison between 2 different drainage patterns in cavernous sinus dAVFs (red arrows). On the left, drainage is supplied by multiple venous outputs, particularly from the ophthalmic vein (blue arrows), the uncal vein and superior petrosal sinus (green arrows). On the right, because of a petrous sinus thrombosis (yellow arrow), the only possible outflow is through the ophthalmic vein. The first cases (left) showed only signs of ocular congestion, the second case (right) showed a reduction in visual acuity, ocular hypertension, and retinal hemorrhages.

A definitely different behavior was observed in tentorial dAVFs where all the patients had an aggressive or a hemorrhagic presentation (3 cases, 2 aggressive, and 1 hemorrhagic presentations). It is important to point out that proper tentorial dAVFs are considered those arising close to tentorial hiatus; the clinical data matches with the anatomical regional conformation, as there are no proper dural sinuses in the region and the drain is exclusively conveyed by leptomeningeal veins toward the region of Galen, which is considered a site at risk; the risk of intracranial hemorrhage reported in the literature varies from 60% to 74% in this site4 (Figures 5 and 8).

FIGURE 8.

Tentorial dAVF; Cognard 3; the blood supply is provided by marginal tentorial artery (Bernasconi and Cassinari). Multiple points of fistula are evident (red arrows) into the Galen venous district (blue arrow), with significant signs of overload of the deep venous system (green arrows).

FIGURE 8.

Tentorial dAVF; Cognard 3; the blood supply is provided by marginal tentorial artery (Bernasconi and Cassinari). Multiple points of fistula are evident (red arrows) into the Galen venous district (blue arrow), with significant signs of overload of the deep venous system (green arrows).

The data concerning anterior cranial fossa dAVFs are also interesting. In our series, their clinical behavior is bimodal: about half had a relatively benign clinical presentation (57.14%), while the remaining cases (42.85%) had a hemorrhagic event. Baltasavias et al19,20,22 have recently presented a similar pattern in their series; they suggest that lesions of the anterior cranial fossa have a diverse venous anatomy and should be divided into 2 basic groups: a shunt in the foremost third of anterior skull base (orbitofrontal region) generally drain through the bridging veins to the superior sagittal sinus, while more direct posterior shunts (lamina cribrosa) flow directly to the olfactory region and then to the deep venous system and to the basal vein of Rosenthal19,20,22 (Figure 9). Although both examples at a first glance might appear similar, the venous drain architecture is very different and corresponds to a distinct clinical behavior. In our series, we had 3 cases in which a hemorrhagic event was observed, and the shunt was localized at the lamina cribrosa in all cases; the remaining cases, for which a relatively benign course was observed, were instead characterized by a more anterior fistula with a venous drainage afferent to the superior sagittal sinus.

FIGURE 9.

Different types of venous drainage of anterior cranial fossa dAVFs; on the left, the shunt is in the orbitofrontal region (red arrow) with drainage through the bridging veins afferent to the superior sagittal sinus (blue arrow). On the right, the shunt occurs more posteriorly, at the level of the lamina cribrosa (red arrow), and the drainage takes place through the leptomeningeal veins of the olfactory region (blue arrow), which, in this case, is the only draining vein of the fistula (outflow restriction pattern). This corresponds to an important qualitative difference of venous drainage.

FIGURE 9.

Different types of venous drainage of anterior cranial fossa dAVFs; on the left, the shunt is in the orbitofrontal region (red arrow) with drainage through the bridging veins afferent to the superior sagittal sinus (blue arrow). On the right, the shunt occurs more posteriorly, at the level of the lamina cribrosa (red arrow), and the drainage takes place through the leptomeningeal veins of the olfactory region (blue arrow), which, in this case, is the only draining vein of the fistula (outflow restriction pattern). This corresponds to an important qualitative difference of venous drainage.

More variable is the clinical presentation of dAVFs of the transverse/sigmoid sinus. This is the prevalent location in our series (14 cases, 28.57% of the total), and presents benign symptoms in 50% of cases, aggressive in 14.28%, and hemorrhage in 35.71%. In this anatomical site, the condition of “isolated sinus” is often the discriminating factor between a benign course and an aggressive or hemorrhagic event. In 4 out of 5 cases, bleeding was observed associated to a condition of “isolated sinus”; the fistula functionally behaves as a direct shunt with leptomeningeal veins even if a dural sinus interposed. Double thrombosis (proximal and distal to the sinus shunt) isolates a compartment with high blood pressure that diverts arterialized flow directly to cortical veins, de facto bypassing the normal sinus circulation (Figures 2and 10).

FIGURE 10.

Transverse-sigmoid sinus dAVF with a double thrombosis determining an isolated sinus configuration; Cognard 2a + b. The fistula is supplied by multiple afferences by the occipital artery, by petrous branches of the middle meningeal artery, by meningeal branches of the PICA, and, to a lesser extent, by branches from the ascending pharyngeal artery (red arrows). Drainage occurs in the middle-distal end of the transverse sinus, which appears “isolated” from the normal circle, by a dual thrombosis (yellow arrows) proximal and distal to the fistulous point. The isolated sinus territory is small. From the isolated transverse sinus, a marked reflux to pial dilated veins with aneurysmal dilatation can be observed (blue arrows). Drainage is then carried out through the straight sinus and through the contralateral transverse/sigmoid sinus (green arrows).

FIGURE 10.

Transverse-sigmoid sinus dAVF with a double thrombosis determining an isolated sinus configuration; Cognard 2a + b. The fistula is supplied by multiple afferences by the occipital artery, by petrous branches of the middle meningeal artery, by meningeal branches of the PICA, and, to a lesser extent, by branches from the ascending pharyngeal artery (red arrows). Drainage occurs in the middle-distal end of the transverse sinus, which appears “isolated” from the normal circle, by a dual thrombosis (yellow arrows) proximal and distal to the fistulous point. The isolated sinus territory is small. From the isolated transverse sinus, a marked reflux to pial dilated veins with aneurysmal dilatation can be observed (blue arrows). Drainage is then carried out through the straight sinus and through the contralateral transverse/sigmoid sinus (green arrows).

Limitations of Current Classifications

The Cognard classification assumes that only types 3 and 4 dAVFs (fistulas with direct duro-pial shunt without interposition of a dural sinus) may show signs of venous strain that determines cortical vein ectasia.7 However, as highlighted in a number of cases of the present series, there are particular conditions in which signs of cortical vein overload/ectasia can also occur in dAVFs of the dural sinuses (grades 2b, 2a + b); these special cases, which are currently not acknowledged by current classification, occur, for example, if an “isolated sinus” is observed; this condition realizes the same hemodynamic effect of a direct shunt without a sinus interposed. Consequently, if certain contingencies occur, 2 lesions belonging to different Cognard degrees (such as a 2b and a 3) and theoretically subjected to a different risk can have similar risk determinants, or even higher in lesion of lower grade (Figure 11). Our data support the importance of focusing on other aspects of the venous drainage that better describe the conformation of the venous system (eg, presence of an isolated sinus) and the strain of the venous drainage (eg, pseudophlebitic aspect, venous outflow restrictions, aneurysms, etc.). These aspects are probably more accurate determinants and better correlate with clinical course.

FIGURE 11.

Comparison between 2 different classes of Cognard: a degree 2a+b (above) and a grade 3 (below). Although belonging to 2 different classes (with different proportional risk) the complexity of the 2 dAVFs does not match the classification. The Cognard 2a+b, fed by branches of the middle meningeal artery and occipital artery (red arrows), configures a condition of “isolated sinus” for a dual thrombosis at the transverse sinus and at torcular (yellow arrows). The venous flow is redirected through the straight sinus to the region of the internal cerebral veins and thalamo-striated veins, which show a pseudoflebitic appearance (blue arrows). The dAVFs clinical manifestation was thalamic bleeding. According to the Cognard classification, this dAVf belongs to a lower class of risk compared to the following example that is a direct duro-pial fistula at the smaller sphenoid wing (fed by branches from middle meningeal artery, deep temporal artery, anterior ethmoid, and maxillary artery—red arrows). Cognard classification: grade 3. This dAVF shows, instead, a long venous drainage to the left posterior parietal region to the superior sagittal sinus. Venous drainage appears “brisk” and shows uniform expansion throughout the whole course of the vein, without signs of strain. The clinical presentation was benign. According to the classification of Cognard, only direct duro-pial shunts (grade 3) can show leptomeningeal venous ectasia. The above example (2a+b) is, however, a case not fully cataloged in the current classifications, as it displays venous ectasia even if it is not a direct duro-pial fistula. In these 2 examples, the grading does not reflect the real class of risk.

FIGURE 11.

Comparison between 2 different classes of Cognard: a degree 2a+b (above) and a grade 3 (below). Although belonging to 2 different classes (with different proportional risk) the complexity of the 2 dAVFs does not match the classification. The Cognard 2a+b, fed by branches of the middle meningeal artery and occipital artery (red arrows), configures a condition of “isolated sinus” for a dual thrombosis at the transverse sinus and at torcular (yellow arrows). The venous flow is redirected through the straight sinus to the region of the internal cerebral veins and thalamo-striated veins, which show a pseudoflebitic appearance (blue arrows). The dAVFs clinical manifestation was thalamic bleeding. According to the Cognard classification, this dAVf belongs to a lower class of risk compared to the following example that is a direct duro-pial fistula at the smaller sphenoid wing (fed by branches from middle meningeal artery, deep temporal artery, anterior ethmoid, and maxillary artery—red arrows). Cognard classification: grade 3. This dAVF shows, instead, a long venous drainage to the left posterior parietal region to the superior sagittal sinus. Venous drainage appears “brisk” and shows uniform expansion throughout the whole course of the vein, without signs of strain. The clinical presentation was benign. According to the classification of Cognard, only direct duro-pial shunts (grade 3) can show leptomeningeal venous ectasia. The above example (2a+b) is, however, a case not fully cataloged in the current classifications, as it displays venous ectasia even if it is not a direct duro-pial fistula. In these 2 examples, the grading does not reflect the real class of risk.

As suggested by Batsavias et al,20,22 one other important aspect of the system of Cognard7 is the discrimination between retrograde and antegrade flow within a dural sinus. Flow inversion is determined by a thrombotic event proximal to the shunt site. Nevertheless, dural sinuses have different anatomies, positions, and interconnections. The significance of a retrograde flow may vary a lot depending on the sinus involved. For example, the presence of retrograde flow in the transverse sinus, in many cases, it may simply mean an increase in regular and anterograde flow in the contralateral transverse sinus.20-22 Our data support this evidence: in our series, the presence of a thrombosis of a dural sinus does not correlate with an increased clinical or hemorrhagic risk.

Besides, the Borden system8,9, which is probably the most common classification system, completely ignores some essential aspects of cerebral venous congestion, and focuses solely on presence/absence of cortical reflux (grades Borden 2 and 3), without discriminating if signs of strain are present.

Angiographic Evaluation and Risk Assessment

This study focuses on specific angiographic aspects that, as confirmed by our data, are commonly observed in high-grade dAVFs. The angiographic evaluation of a dAVF should be very careful in the evaluation of a fistulous site, but also to detect presence of thromboses in the dural sinuses. Above all, the delayed parenchymal and venous phase should be meticulously reviewed at DSA as some shading aspects of venous strain can be identified at this phase; these are genuine indicators of focal or diffused venous overload over brain parenchyma. The correct stratification of risk is of fundamental importance in setting the therapeutic strategy and timing.3,6

Many points of the “original” dAVF classifications by Borden and Cognard are relevant and helpful but the angiographic aspect of the venous drainage (especially at the DSA delayed parechymal phase) could improve the accuracy of predicting the dAVF with worst natural history. In order to incorporate these observations in a practical setting, dAVFs could be first divided according to the absence or presence of cortical vein reflux. Absence of cortical vein reflux = low risk. Presence of cortical vein reflux = high risk. Clinical risk, in those displaying cortical vein reflux, could be further stratified according to angio-architectural features predisposing to an “aggressive clinical presentation—A” (restriction of outflow and pseudophlebitic pattern) adding a point for each characteristic (A0, A1, or A2, if present none, 1, or 2 at the same time); the same could be applied to the “bleeding risk factors—B” (dual thrombosis and venous aneurysms), thus obtaining B0, B1, or B2 according to the DSA venous phase. At this point, it is possible to have a global score able to quickly identify the overall risk and propensity for aggressive presentation (A) and bleeding (B) (Figure 12).

FIGURE 12.

Diagrammatic representation of dAVF stratification of risk at DSA (with focus on delayed parenchymal and venous phase). dAVFs could be first divided according to the absence or presence of cortical vein reflux. Absence of cortical vein reflux = low risk. Presence of cortical vein reflux = high risk. In the latter case, the aggressive/non-hemorrhagic A risk can be evaluated if at DSA a restriction of outflow or a pseudophlebitic pattern, or both parameters are present. Equally, bleeding risk B can be identified if a DSA displays a dual thrombosis (isolated sinus), presence of venous aneurysms, or both features are present. CVR, cortical vein reflux.

FIGURE 12.

Diagrammatic representation of dAVF stratification of risk at DSA (with focus on delayed parenchymal and venous phase). dAVFs could be first divided according to the absence or presence of cortical vein reflux. Absence of cortical vein reflux = low risk. Presence of cortical vein reflux = high risk. In the latter case, the aggressive/non-hemorrhagic A risk can be evaluated if at DSA a restriction of outflow or a pseudophlebitic pattern, or both parameters are present. Equally, bleeding risk B can be identified if a DSA displays a dual thrombosis (isolated sinus), presence of venous aneurysms, or both features are present. CVR, cortical vein reflux.

CONCLUSIONS

This study confirms the recently published observations on high-grade dAVFs; within this category, it is possible to identify 2 distinct subgroups with different clinical behavior, even if both present cortical venous reflux. Given the presence of cortical venous reflux, a series of intermediate conditions exist between the simple occurrence of a cortical reflux and the presence of venous aneurysms and ectasia. These intermediate degrees of venous congestion correlate well with the clinical risk; the simple definition of cortical reflux, as indicated in the current classifications, appears insufficient to identify a wide range of conditions that ultimately determine the organization of the cortical venous drainage.

A direct duro-pial shunt does not necessarily provide a hemodynamic impact on brain parenchyma greater than sinus-based shunts; specific configurations of the dural sinuses, such as the presence of a single or double thrombosis, or particular conditions of leptomeningeal reflux, such as the presence of a “restriction of venous drainage” or a “pseudophlebitic appearance of cortical vessels,” effectively correlate with an aggressive/nonhemorrhagic clinical presentation or with the hemorrhagic risk. It should be also underlined that these observations are far from uncommon.

The angiographic aspects of venous drainage presented in this study may prove useful to identify categories of patients with a high risk and to support the choice between an immediate or elective treatment. These indicators may prove also effective in angiographic follow-up, since their presence is a sign of strain, focal or diffused, of the leptomeningeal venous system.

Disclosure

The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

REFERENCES

1.
Hacein-Bey
L
,
Konstas
AA
,
Pile-Spellman
J
.
Natural history, current concepts, classification, factors impacting endovascular therapy, and pathophysiology of cerebral and spinal dural arteriovenous fistulas
.
Clin Neurol Neurosurg
 .
2014
;
121
:
64
-
75
.
2.
Signorelli
F
,
Della Pepa
GM
,
Sabatino
G
et al
Diagnosis and management of dural arteriovenous fistulas: a 10 years single-center experience
.
Clin Neurol Neurosurg
 .
2015
;
128
:
123
-
129
.
3.
Ghobrial
GM
,
Marchan
E
,
Nair
AK
et al
Dural arteriovenous fistulas: a review of the literature and a presentation of a single institution's experience
.
World Neurosurg
 .
2013
;
80
(
1-2
):
94
-
102
.
4.
Daniels
DJ
,
Vellimana
AK
,
Zipfel
GJ
,
Lanzino
G
.
Intracranial hemorrhage from dural arteriovenous fistulas: clinical features and outcome
.
Neurosurg Focus
 .
2013
;
34
(
5
):
E15
.
5.
van Dijk
JM
,
terBrugge
KG
,
Willinsky
RA
,
Wallace
MC
.
Clinical course of cranial dural arteriovenous fistulas with long-term persistent cortical venous reflux
.
Stroke
 .
2002
;
33
(
5
):
1233
-
1236
.
6.
Gupta
A
,
Periakaruppan
A
.
Intracranial dural arteriovenous fistulas: a review
.
Indian J Radiol Imaging
 .
2009
;
19
(
1
):
43
-
48
.
7.
Cognard
C
,
Gobin
YP
,
Pierot
L
et al
Cerebral dural arteriovenous fistulas: clinical and angiographic correlation with a revised classification of venous drainage
.
Radiology
 .
1995
;
194
(
3
):
671
-
680
.
8.
Borden
JA
,
Wu
JK
,
Shucart
WA
.
A proposed classification for spinal and cranial dural arteriovenous fistulous malformations and implications for treatment
.
J Neurosurg
 .
1995
;
82
(
2
):
166
-
179
.
9.
Borden
JA
,
Wu
JK
,
Shucart
WA
.
Correction: dural arteriovenous fistulous malformations
.
J Neurosurg
 .
1995
;
82
(
4
):
705
-
706
.
10.
Washington
CW
,
McCoy
KE
,
Zipfel
GJ
.
Update on the natural history of cavernous malformations and factors predicting aggressive clinical presentation
.
Neurosurg Focus
 .
2010
;
29
(
3
):
E7
.
11.
Zipfel
GJ
,
Shah
MN
,
Refai
D
,
Dacey
RG
Jr
,
Derdeyn
CP
.
Cranial dural arteriovenous fistulas: modification of angiographic classification scales based on new natural history data
.
Neurosurg Focus
 .
2009
;
26
(
5
):
E14
.
12.
Soderman
M
,
Pavic
L
,
Edner
G
,
Holmin
S
,
Andersson
T
.
Natural history of dural arteriovenous shunts
.
Stroke
 .
2008
;
39
(
6
):
1735
-
1739
.
13.
Duffau
H
,
Lopes
M
,
Janosevic
V
et al
Early rebleeding from intracranial dural arteriovenous fistulas: report of 20 cases and review of the literature
.
J Neurosurg
 .
1999
;
90
(
1
):
78
-
84
.
14.
Satomi
J
,
van Dijk
JM
,
Terbrugge
KG
,
Willinsky
RA
,
Wallace
MC
.
Benign cranial dural arteriovenous fistulas: outcome of conservative management based on the natural history of the lesion
.
J Neurosurg
 .
2002
;
97
(
4
):
767
-
770
.
15.
Satomi
J
,
Satoh
K
.
[Epidemiology and etiology of dural arteriovenous fistula]
.
Brain Nerve
 .
2008
;
60
(
8
):
883
-
886
.
16.
Willinsky
R
,
Goyal
M
,
terBrugge
K
,
Montanera
W
.
Tortuous, engorged pial veins in intracranial dural arteriovenous fistulas: correlations with presentation, location, and MR findings in 122 patients
.
AJNR Am J Neuroradiol
 .
1999
;
20
(
6
):
1031
-
1036
.
17.
Strom
RG
,
Botros
JA
,
Refai
D
et al
Cranial dural arteriovenous fistulae: asymptomatic cortical venous drainage portends less aggressive clinical course
.
Neurosurgery
 .
2009
;
64
(
2
):
241
-
247
;
discussion 247-248
.
18.
Shin
NY
,
Kwon
YS
,
Ha
SY
,
Kim
BM
,
Kim
DI
,
Kim
DJ
.
Venous angioarchitectural features of intracranial dural arteriovenous shunt and its relation to the clinical course
.
Neuroradiology
 .
2013
;
55
(
9
):
1119
-
1127
.
19.
Baltsavias
G
,
Spiessberger
A
,
Hothorn
T
,
Valavanis
A
.
Cranial dural arteriovenous shunts. Part 4. Clinical presentation of the shunts with leptomeningeal venous drainage
.
Neurosurg Rev
 .
2015
;
38
(
2
):
283
-
291
;
discussion 291
.
20.
Baltsavias
G
,
Roth
P
,
Valavanis
A
.
Cranial dural arteriovenous shunts. Part 3. Classification based on the leptomeningeal venous drainage
.
Neurosurg Rev
 .
2015
;
38
(
2
):
273
-
281
;
discussion 281
.
21.
Baltsavias
G
,
Parthasarathi
V
,
Aydin
E
,
Al Schameri
RA
,
Roth
P
,
Valavanis
A
.
Cranial dural arteriovenous shunts. Part 1. Anatomy and embryology of the bridging and emissary veins.
Neurosurg Rev
 .
2015
;
38
(
2
):
253
-
263
;
discussion 263-254
.
22.
Baltsavias
G
,
Kumar
R
,
Avinash
KM
,
Valavanis
A
.
Cranial dural arteriovenous shunts. Part 2. The shunts of the bridging veins and leptomeningeal venous drainage
.
Neurosurg Rev
 .
2015
;
38
(
2
):
265
-
271
;
discussion 272
.
23.
Baltsavias
G.
Aberrant venous drainage pattern in a medial sphenoid wing dural arteriovenous fistula: commentary
.
World Neurosurg
 .
2014
;
82
(
3-4
):
e563
-
565
.
24.
Baltsavias
G
,
Valavanis
A
.
Endovascular treatment of 170 consecutive cranial dural arteriovenous fistulae: results and complications
.
Neurosurg Rev
 .
2014
;
37
(
1
):
63
-
71
.

COMMENTS

The authors present a retrospective review of their series of 49 cases of high-grade dural arteriovenous fistulas (dAVFs) that had an angiographic appearance of cortical venous reflux. They correlated the venous drainage patterns with the overall clinical presentation. The authors observed a benign presentation in 42%, an aggressive presentation in 42%, and hemorrhage in 14%. As a result, they extrapolated 2 distinct subpopulations within these high-grade lesions: those patients with an aggressive, non-hemorrhagic clinical presentation, and those with hemorrhage. The subset of patients with a non-aggressive clinical presentation exhibited characteristics of outflow restriction and pseudo-phlebitic cortical vessels, whereas those with hemorrhage showed dual thrombosis and venous aneurysms.

Classifications by Borden and Cognard have placed a premium on identifying those patients with dAVFs with cortical venous drainage.1,2 Treating this subset of high risk lesions ubiquitously as 1 cohort may lack the nuance needed for the treatment of a very complex vascular disease. The authors do very well in elucidating specific venous architecture to more completely define this pathology in their patient cohort. This will certainly aid not only in our understanding of dAVFs and the gradations of clinical risk but also in deepening our understanding of the eclectic nature cerebral venous physiology.

Vernard S. Fennell

Elad I. Levy

Buffalo, New York