Dural Arteriovenous Fistulas With or Without Cerebral Venous Thrombosis: A Cross-Sectional Analysis of 511 Patients

Abstract

BACKGROUND AND OBJECTIVES:

Recent studies suggest a bidirectional relationship of dural arteriovenous fistula (DAVF) with cerebral venous thrombosis (CVT). We aimed to compare the characteristics of patients with DAVF with or without CVT and to analyze the risk factors for the coexistence of CVT in a DAVF population.

METHODS:

A total of 511 adult patients with DAVF were enrolled consecutively in our hospital from February 2019 through November 2022. Demographic data, clinical manifestations, and imaging characteristics were reviewed in detail. The patients with DAVF were divided into two groups: DAVF with CVT (DAVF-CVT) group and without CVT (DAVF alone) group. Univariate logistic regression and multivariate logistic regression were used to analyze the risk factors for the coexistence of CVT and DAVF.

RESULTS:

CVT was found in 19.8% of patients with DAVF. In univariate analysis, compared with the DAVF-alone group, the DAVF-CVT group was more likely to have tinnitus (P = .001), blurred vision (P < .001), visual field loss (P = .001), focal neurological deficits (P = .002), seizures (P = .008), and cognitive impairment (P = .046) and less likely to have spinal cord/brain stem dysfunction (P = .004). In addition, there were significant differences in age (P = .009), sex (P = .019), the occurrence of venous cerebral infarction (P = .001), and DAVF location (P < .001) between the two groups. Furthermore, multivariate analysis showed that blurred vision, venous cerebral infarction, large sinus DAVF, and multiple DAVF were risk factors for the coexistence of CVT in patients with DAVF, with the odds ratio of 2.416 (95% CI 1.267-4.606, P = .007), 6.018 (95% CI 1.289-28.100, P = .022), 5.801 (95% CI 2.494-13.496, P < .001), and 5.640 (95% CI 2.122-14.989, P = .001), respectively.

CONCLUSION:

CVT occurred in approximately one fifth of patients with DAVF. Blurred vision, venous cerebral infarction, large sinus DAVF, and multiple DAVF may be the risk factors for predicting the coexistence of CVT in patients with DAVF.

ABBREVIATIONS:

CS-DAVF

cavernous sinus DAVF

CVT

cerebral venous thrombosis

DAVF

dural arteriovenous fistula

LS-DAVF

large sinus DAVF

MDAVF

multiple DAVF

TDAVF

tentorial DAVF.

Dural arteriovenous fistulas (DAVFs) are pathological anastomoses between dural arteries and venous sinuses or cortical veins, accounting for approximately 10%–15% of all intracranial vascular malformations. The pathogenesis of DAVF is still controversial, which was initially believed to be congenital, but now generally considered to be acquired.¹ Cerebral venous thrombosis (CVT) refers to the formation of thrombus in intracranial veins or venous sinuses because of various causes, which obstructs venous drainage and cerebrospinal fluid circulation, leading to intracranial hypertension and focal brain damage. It is a special subtype of cerebral vascular disease and accounts for approximately 0.5%–1% of all strokes.² The association of the two disease entities seems to be bidirectional. On the one hand, thrombus in the cerebral veins or venous sinuses can block blood flow, increase venous pressure and decrease cerebral perfusion, and cause the opening up of pre-existing physiological arteriovenous shunts or the stimulation of neoangiogenesis, ultimately leading to the formation of DAVF; on the other hand, DAVF may generate secondary CVT, as well. First, DAVF leads to turbulent flow into the venous sinus, which damages the intima and generates luminal thrombosis. Second, DAVF causes venous hypertension and blood stasis, leading to venous thrombosis.^3-5

It has been reported that DAVF is present in approximately 0.9%–13% of patients with CVT, and the occurrence of CVT in patients with DAVF is 31%–39%.^3,4,6-9 Furthermore, there are temporal and anatomic correlations between the two disease entities. DAVF can be diagnosed before, simultaneously with, and subsequent to CVT.^4,6,10,11 In addition, CVT was demonstrated to be primarily located at either the sinus(es) around the DAVF or the downstream venous flow pathways of the DAVF.^4,6

Despite the strong association between DAVF and CVT, there are relatively few studies on the features of patients with DAVF with or without CVT yet. We examined a cohort of 511 patients with DAVFs and correlated the presence of the DAVF with underlying CVT. We aimed to compare the features of patients with DAVF with the presence or the absence of CVT at length and to analyze the risk factors for the coexistence of CVT in a large continuous adult DAVF population.

METHODS

Study Population

This is a hospital-based cross-sectional study. A total of 558 consecutive adult patients with DAVF, who were hospitalized in our hospital from February 2019 through November 2022, were initially included in the study (Figure 1). The study was approved by the Institutional Review Board of the hospital's ethics committee and received informed consent from all patients. Data of the patients in the inpatient database were reviewed in detail, including demographic data, clinical manifestations, and imaging features. The diagnosis of DAVF and CVT was confirmed by digital subtraction angiography (DSA).^12,13 According to the following inclusion and exclusion criteria, 47 patients were excluded and 511 patients were finally enrolled in the analysis. The patients with DAVF were further divided into two groups: patients with DAVF with CVT (DAVF-CVT) group and without CVT (DAVF alone) group.

FIGURE 1.

Flow diagram of the study. AVM, arteriovenous malformation; CVT, cerebral venous thrombosis; DAVF, dural arteriovenous fistula; DSA, digital subtraction angiography.

Inclusion criteria:

1. At least 18 years old;

2. Confirmed with DAVF by DSA;

3. Signed informed consent to be included in this study.

Exclusion criteria:

1. Complicated with arteriovenous malformation;

2. Complicated with intracranial space-occupying lesions;

3. Incomplete data.

DAVF Location

DAVF was categorized into the following five types according to the location involved: (1) multiple DAVF (MDAVF): the DAVFs involved two or more sites simultaneously or successively¹⁴; (2) cavernous sinus DAVF (CS-DAVF): the DAVFs located in the cavernous sinus; (3) large sinus DAVF (LS-DAVF): the DAVFs located in the transverse sinus, sigmoid sinus, or superior sagittal sinus; (4) tentorial DAVF: including galenic, straight sinus, torcular, tentorial sinus, superior petrosal sinus, and incisural DAVFs (Lawton-Halbach's classification); (5) other DAVF: the DAVFs that did not belong to the above four categories.

DAVF Course

We defined the course of DAVF as the time from the onset of relevant symptoms, such as headache, tinnitus, and visual impairment, to hospitalization, with 48 hours and 30 days as the dividing points for acute, subacute, and chronic disease courses (acute course: <48 hours, subacute course: 48 hours to 30 days, and chronic course: >30 days).

Statistical Analysis

Statistical analysis was performed using the Statistical Package for the Social Sciences 26.0 (IBM Corp.). Continuous variables are represented by median (IQR), and categorical variables are represented by numeric (percentage [%]). The continuous and categorical variables were compared by univariate logistic regression. For the factors with P-values of less than .1, multivariate logistic regression was further performed and odds ratios (ORs) and 95% CI were calculated. A P-value of less than .05 was considered to be statistically significant.

RESULTS

Demographic Characteristics

The characteristics of the 511 adult patients with DAVF are summarized in Table, including 410 (80.2%) cases of DAVF alone and 101 (19.8%) cases of DAVF-CVT.

TABLE.

The Characteristics of Patients With DAVF With and Without CVT

Variable	Total n = 511	DAVF-alone n = 410	DAVF-CVT n = 101	Univariate analysis		Multivariate analysis
				P value	OR (95% CI)	P value	OR (95% CI)
Demographic characteristics
Northern China, n (%)	391 (76.52)	321 (78.29)	70 (69.31)	.058*	0.626 (0.386, 1.015)	.294	0.727 (0.400, 1.319)
Age, years, median (IQR)	53.0 (44.0, 63.0)	54.0 (45.0, 63.0)	50.0 (38.0, 60.0)	.009**	0.978 (0.961, 0.994)	.610	0.995 (0.974, 1.015)
Male, n (%)	344 (67.32)	286 (69.76)	58 (57.43)	.019**	1.710 (1.093, 2.674)	.171	1.469 (0.847, 2.546)
BMI, median (IQR)a	24.49 (22.74, 26.57)	24.49 (22.81, 26.46)	24.30 (22.55, 27.05)	.795	0.991 (0.928, 1.059)
Clinical characteristics
History, n (%)
Hypertension	173 (33.86)	140 (34.15)	33 (32.67)	.779	0.936 (0.589, 1.487)
Diabetes mellitus	74 (14.48)	61 (14.88)	13 (12.87)	.608	0.845 (0.444, 1.607)
Dyslipidemia	227 (44.42)	182 (44.39)	45 (44.55)	.976	1.007 (0.650, 1.560)
Hyperuricemiab	70 (14.34)	60 (15.42)	10 (10.10)	.181	0.616 (0.303, 1.252)
Hyperhomocysteinemiac	51 (13.60)	42 (14.09)	9 (11.69)	.584	0.807 (0.374, 1.739)
Stroke	36 (7.05)	28 (6.83)	8 (7.92)	.701	1.174 (0.518, 2.659)
Craniocerebral trauma	29 (5.68)	20 (4.88)	9 (8.91)	.122	1.908 (0.841, 4.326)
Craniotomy	13 (2.54)	10 (2.44)	3 (2.97)	.762	1.1224 (0.331, 4.534)
Local infection	5 (0.98)	4 (0.98)	1 (0.99)	.989	1.015 (0.112, 9.181)
Disease duration, n (%)
Acute (<48 h)	28 (5.48)	27 (6.59)	1 (0.99)	Reference		Reference
Subacute (48 h-30 d)	113 (22.11)	97 (23.66)	16 (15.84)	.156	4.454 (0.565, 35.111)	.317	2.975 (0.352, 22.133)
Chronic (>30 d)	370 (72.41)	286 (69.76)	84 (83.17)	.044**	7.930 (1.062, 59.226)	.229	3.615 (0.446, 29.306)
Clinical symptoms, n (%)
Headache	211 (41.29)	167 (40.73)	44 (43.56)	.605	1.123 (0.723, 1.744)
Dizziness	106 (20.74)	90 (21.95)	16 (15.84)	.177	0.669 (0.374, 1.199)
Intracranial murmur	22 (4.31)	15 (3.66)	7 (6.93)	.154	1.961 (0.778, 4.945)
Tinnitus	121 (23.68)	84 (20.49)	37 (36.63)	.001**	2.244 (1.402, 3.592)	.726	1.108 (0.624, 1.968)
Hearing loss	11 (2.15)	7 (1.71)	4 (3.96)	.175	2.374 (0.681, 8.272)
Chemosis	65 (12.72)	53 (12.93)	12 (11.88)	.778	0.908 (0.466, 1.772)
Exophthalmos	61 (11.94)	46 (11.22)	15 (14.85)	.315	1.38 (0.736, 2.587)
Blurred vision	74 (14.48)	44 (10.73)	30 (29.70)	.000**	3.515 (2.071, 5.966)	.007**	2.416 (1.267, 4.606)
Visual field loss	17 (3.33)	8 (1.95)	9 (8.91)	.001**	4.916 (1.847, 13.084)	.397	1.650 (0.518, 5.295)
Focal neurological deficits	34 (6.65)	20 (4.88)	14 (13.86)	.002**	3.138 (1.525, 6.456)	.324	1.595 (0.631, 4.031)
Cranial nerve dysfunction	47 (9.20)	39 (9.51)	8 (7.92)	.621	0.818 (0.370, 1.810)
Spinal cord/brain stem dysfunction	63 (12.33)	60 (14.63)	3 (2.97)	.004**	0.179 (0.055, 0.582)	.096	0.345 (0.098, 1.209)
Impaired consciousness	39 (7.63)	28 (6.83)	11 (10.89)	.172	1.667 (0.800, 3.475)
Seizures	31 (6.07)	19 (4.63)	12 (11.88)	.008**	2.775 (1.300, 5.924)	.412	1.455 (0.593, 3.568)
Cognitive impairment	18 (3.52)	11 (2.68)	7 (6.93)	.046**	2.701 (1.020, 7.153)	.400	1.65 (0.514, 5.291)
Psycho-behavioral abnormality	3 (0.59)	2 (0.49)	1 (0.99)	.562	2.040 (0.183, 22.722)
Head and neck mass	7 (1.37)	6 (1.46)	1 (0.99)	.716	0.673 (0.080, 5.656)
Asymptomatic	34 (6.65)	31 (7.56)	3 (2.97)	.110	0.374 (0.112, 1.250)
Accessory examination characteristics
Fibrinogen, g/Ld	3.08 (2.63, 3.59)	3.09 (2.70, 3.64)	2.87 (2.43, 3.47)	.064*	0.791 (0.618, 1.014)	.914	1.010 (0.771, 1.324)
D-dimer, μg/mLe	0.34 (0.22, 0.75)	0.34 (0.22, 0.75)	0.35 (0.22, 0.75)	.547	0.933 (0.746, 1.168)
Intracranial hemorrhage, n (%)	105 (20.55)	87 (21.22)	18 (17.82)	.450	0.805 (0.459, 1.412)
Cerebral venous infarction, n (%)	12 (2.35)	4 (0.98)	8 (7.92)	.001**	8.731 (2.575, 29.609)	.022**	6.018 (1.289, 28.100)
DAVF location, n (%)
Other DAVF	137 (26.81)	129 (31.46)	8 (7.92)	Reference		Reference
LS-DAVF	151 (29.55)	98 (23.90)	53 (52.48)	.000**	8.721 (3.964, 19.185)	.000**	5.801 (2.494, 13.496)
CS-DAVF	63 (12.33)	58 (14.15)	5 (4.95)	.578	1.390 (0.436, 4.432)	.705	0.788 (0.230, 2.702)
TDAVF	102 (19.96)	91 (22.20)	11 (10.89)	.168	1.949 (0.754, 5.037)	.202	1.936 (0.702, 5.340)
MDAVF	58 (11.35)	34 (8.29)	24 (23.76)	.000**	11.382 (4.698, 27.576)	.001**	5.640 (2.122, 14.989)
CVT location, n (%)
Superior sagittal sinus	—	—	30 (15.31)
Straight sinus	—	—	9 (4.60)
Torcular	—	—	8 (4.08)
Left transverse sinus	—	—	39 (19.90)
Right transverse sinus	—	—	29 (14.80)
Left sigmoid sinus	—	—	33 (16.84)
Right sigmoid sinus	—	—	23 (11.74)
Left internal jugular vein	—	—	8 (4.08)
Right internal jugular vein	—	—	9 (4.60)
Other	—	—	8 (4.08)

BMI, body mass index; CS-DAVF, cavernous sinus DAVF; CVT, cerebral venous thrombosis; DAVF, dural arteriovenous fistula; LS-DAVF, large sinus DAVF; MDAVF, multiple DAVF; OR, odds ratio; TDAVF, tentorial DAVF.

^a4 missing data in the DAVF-alone group.

^b21 missing data in the DAVF-alone group and two in the DAVF-CVT group.

^c112 missing data in the DAVF-alone group and 24 in the DAVF-CVT group.

^d9 missing data in the DAVF-alone group and two in the DAVF-CVT group.

^e194 missing data in the DAVF-alone group and 41 in the DAVF-CVT group.

* P < .1; ** P < .05

The patients came from almost all areas of the country except Hong Kong, Macau, Taiwan, and Shanghai, in which less than one fifth were from Beijing. In both groups, cases from the north accounted for the majority (78.29% and 69.31%, respectively). Further univariate analysis showed a marginal difference in regional distribution between the two groups (P = .058). Figure 2 shows the age and sex distribution of the two groups, with most of the patients aged between 41 and 60 years, accounting for 51.46% and 48.51%, respectively. patients with DAVF-CVT were significantly younger than patients with DAVF alone, with the median age of 50.0 years (IQR 38.0, 60.0) and 54.00 years (IQR 45.0, 63.0), respectively (P = .009). Besides, although the proportion of males in both groups was higher than that of females, the male proportion of DAVF-CVT was significantly lower than that of DAVF alone (57.43% and 69.76%, respectively, P = .019). There was no significant difference in body mass index between the two groups (P = .795). However, it is worth noting that the proportion of overweight and obesity (BMI ≥ 24) was high in both groups, accounting for 59.02% and 60.39%, respectively.

FIGURE 2.

Sex and age distribution of patients with DAVF with and without CVT: A, DAVF-alone group and B, DAVF-CVT group. CVT, cerebral venous thrombosis; DAVF, dural arteriovenous fistula.

Clinical Characteristics

There was no significant difference in the history of hypertension, diabetes mellitus, dyslipidemia, hyperuricemia, hyperhomocysteinemia, stroke, craniocerebral trauma, craniotomy, and local infection between the two groups (all P > .05). Among patients with DAVF-CVT and DAVF alone, 3 (2.97%) and 31 (7.56%) cases were asymptomatic and sought medical treatment after routine physical examination. There was a significant difference in the aspect of disease duration between the two groups although most of the cases showed a chronic process in both the groups (DAVF alone vs DAVF-CVT was 69.76% vs 83.17%). Univariate analysis demonstrated that compared with the acute process, the chronic process conferred an increased risk of thrombosis, with the OR of 7.930 (95% CI 1.062, 59.226, P = .044). The clinical symptom spectrum of the two groups is displayed in Figure 3. The most common symptom in DAVF-CVT was headache (43.56%), followed by tinnitus (36.63%) and blurred vision (29.70%), whereas in the DAVF-alone group, the most common symptom was headache (40.73%), followed by dizziness (21.95%) and tinnitus (20.49%). Furthermore, univariate analysis showed that patients with DAVF-CVT were more likely to have tinnitus (P = .001), blurred vision (P < .001), visual field loss (P = .001), focal neurological deficits (P = .002), seizures (P = .008), and cognitive impairment (P = .046) and less likely to have spinal cord/brain stem dysfunction (P = .004). Other clinical symptoms, including headache, dizziness, intracranial murmur, hearing loss, chemosis, exophthalmos, cranial nerve dysfunction, impaired consciousness, psycho-behavioral abnormality, and head and neck mass, were comparable between the two groups.

FIGURE 3.

Clinical symptom spectrum of patients with DAVF with and without CVT. CVT, cerebral venous thrombosis; DAVF, dural arteriovenous fistula.

Assistant Examination Characteristics

There was no significant association between the levels of fibrinogen or D-dimer and the presence of CVT in patients with DAVF. In addition, there was a significant difference in the occurrence of venous cerebral infarction between DAVF-CVT and DAVF-alone patients (7.92% and 0.98%, P = .001), but the occurrence of intracranial hemorrhage was comparable (17.82% and 21.22%, P = .450). Most importantly, there was a significant difference in DAVF location between the two groups. The most common DAVF locations in the DAVF-CVT group and the DAVF-alone group were LS-DAVF (52.48%) and other DAVF (31.46%), respectively, whereas the least common DAVF locations were CS-DAVF (4.95%) and MDAVF (8.29%), respectively. Univariate logistic regression demonstrated that compared with other DAVF, LS-DAVF and MDAVF were more prone to combine with CVT, with a 7.721-fold (95% CI 2.964-18.185) and 10.382-fold (95% CI 3.698, 26.576) increased risk, respectively (both P < .001). What is more, in 101 patients with DAVF-CVT, a total of 196 sinus thromboses occurred, of which the left transverse sinus (19.90%) was the most common location, followed by the left sigmoid sinus (16.84%).

Multivariate Logistic Regression

Multivariate logistic regression showed that the significant associations of sex, age, disease duration, tinnitus, visual field loss, focal neurological deficits, seizures, cognitive impairment, and spinal cord/brain stem dysfunction with the presence of CVT disappeared; however, the significant associations of blurred vision, venous cerebral infarction, LS-DAVF, and MDAVF with the coexistence of CVT in patients with DAVF remained, with the OR of 2.416 (95% CI 1.267-4.606, P = .007), 6.018 (95% CI 1.289-28.100, P = .022), 5.801 (95% CI 2.494-13.496, P < .001), and 5.640 (95% CI 2.122-14.989, P = .001), respectively.

DISCUSSION

In our hospital-based cohort of 511 consecutive adult patients with DAVF, the prevalence of CVT was approximately one fifth (19.8%). We found that after adjusting for confounding factors, blurred vision, venous cerebral infarction, and LS-DAVF and MDAVF were the risk factors for the coexistence of CVT in patients with DAVF.

In this study, among 511 adult patients with DAVF, the median age was 53.0 (IQR 44.0, 63.0) years and 344 (67.32%) cases were males. The proportion of male patients was higher than that in previous studies.³ The reasons for this discrepancy were unclear but may be related to our hospital being a tertiary referral medical center for neurovascular diseases. The condition was relatively more severe and complex because of referral bias, and previous studies have shown that DAVF accompanied by aggressive neurological symptoms was more common in men than in women.¹⁵ In addition, of these patients with DAVF, only one fifth also had CVT, which was lower than other studies. Among 69 patients with DAVF at National Taiwan University Hospital (1995-2002) and 178 adult patients with DAVF at University Medical Center Amsterdam (2007-2020), the prevalence of CVT was 39% and 31%, respectively.^3,4 However, the sample size of the former study was relatively small. In the latter study, most patients with DAVF underwent routine follow-up imaging, which increased the chance of incidentally finding asymptomatic CVT.³

There were significant differences in age and sex between patients with DAVF with and without CVT in univariate analysis. Patients with DAVF-CVT appeared to be younger and more likely to be female. Similarly, in a CVT population from the International CVT Consortium registries, patients with DAVF were older and more frequently male.⁶ This may be because DAVF mostly affects middle-aged patients with no sex preference, whereas CVT mainly affects young individuals and females.¹⁶ However, the differences in age and sex between the two groups disappeared in further multivariate analysis. As for body mass index, there was no significant difference between the two groups. However, it is necessary to note that the proportion of overweight and obesity was rather high in both groups, accounting for approximately 60%.

The clinical symptoms of DAVF and CVT are similar to some extent, both of which can be manifested as headache, neuro-ophthalmological symptoms, seizures, and focal neurological deficits. However, it is unclear whether there is a difference in clinical symptoms between patients with DAVF with CVT and those without CVT. In this study, we found that patients with DAVF-CVT more frequently exhibit symptoms such as tinnitus, blurred vision, visual field loss, focal neurological deficits, seizures, and cognitive impairment and less frequently exhibit spinal cord/brain stem dysfunction. However, only the significant association of blurred vision with the coexistence of CVT remained after the adjustment of confounding factors; that is, blurred vision occurred more frequently in the DAVF-CVT group, with a 1.416-fold (95% CI 0.267-3.606) increased risk. For patients with CVT, visual impairment can result from papilledema (intracranial hypertension) and focal lesions (cerebral infarction or hemorrhage).¹⁷ For patients with DAVF, venous reflux into the orbital veins, especially the superior ophthalmic vein, and the resulting orbital venous hypertension can also lead to visual impairment.^18-20 Because CVT was mainly located at the sinuses around DAVF or the downstream venous flow pathways of DAVF,⁴ the presence of CVT may further impede venous drainage and exacerbate venous reflux and intracranial hypertension in patients with DAVF, which may explain the significant association between blurred vision and the presence of CVT in patients with DAVF.

D-dimer is one of the important indicators for CVT diagnosis.^21-24 However, in our study, the level of D-dimer was not associated with the presence of CVT in patients with DAVF. The result is consistent with a small study including 18 patients with DAVF, which showed no significant difference between the mean value of D-dimer and sinus occlusion.²⁵ In addition, we found no significant association of plasma fibrinogen levels with the presence of CVT in patients with DAVF, as well.

Both CVT and DAVF are at the risk of developing aggressive symptoms (venous infarction or hemorrhage). We found that venous cerebral infarction was a risk factor for predicting the coexistence of thrombosis in patients with DAVF. For patients with DAVF, arterialized flow can lead to an increased venous pressure, impede antegrade venous drainage, cause retrograde venous drainage, and thus result in chronic passive congestion and venous infarction.²⁶ As mentioned above, the coexistence of thrombosis may further aggravate venous hypertension, making it more likely to cause venous infarction.

Most importantly, on analyzing the relationship between the DAVF location and the CVT presence in patients with DAVF, we found that compared with other DAVF, LS-DAVF and MDAVF were more prone to combine with CVT, with a 4.801-fold and 5.640-fold increased risk, respectively. The result differs from a study by Tsai et al,⁴ which showed no significant difference in DAVF location between the DAVF-alone group and the DAVF-CVT group. However, the small sample size (69 cases) might have skewed the results.

Based on our findings, we have some suggestions. First, for patients with DAVF, especially LS-DAVF and MDAVF, or those with blurred vision or venous cerebral infarction, clinicians should be vigilant for CVT. It is important to look for signs of venous thrombosis on DSA or high-resolution magnetic resonance thrombosis imaging, if necessary. Second, for patients with CVT, attention should also be paid to the presence of DAVF, especially if the following conditions occur: (1) unknown cause of thrombosis and unobvious effect of anticoagulant therapy, (2) recurrent symptoms or new symptoms (such as headache or tinnitus), (3) persistent and unrelieved intracranial hypertension, and (4) obvious intracranial varicose veins on imaging.^5,27-29 Last but not least, stenosis or thrombosis of the venous outflow is one of the main mechanisms for the worsening of venous drainage, and thus, we suggest that approaches that can improve venous drainage and reduce venous hypertension, such as balloon angioplasty or stenting and anticoagulant therapy, may be beneficial for the treatment of DAVFs.³⁰ Previous studies have shown that balloon angioplasty or stent placement can reduce the grade of fistula and even close the fistula.^31,32 However, there are no comparative studies on anticoagulant therapy as a conservative treatment of DAVF. In theory, anticoagulation therapy can promote the recanalization of venous sinus or cortical veins, prevent secondary venous thrombosis, and reduce blood viscosity, thereby improving venous drainage, reducing intracranial pressure, and ameliorating clinical symptoms; however, the bleeding risk of anticoagulants should also be noted. Well-designed prospective studies are needed to demonstrate the strengths and weaknesses of anticoagulants in these patients.

Limitations

Several limitations warrant comment to this study, however. First of all, because of the retrospective study design, we were unable to explore the effect of hypercoagulable status or other thromboembolic factors on the coexistence of CVT in patients with DAVF. Second, because our hospital is a tertiary referral medical center for cerebral vascular disorders, the condition is relatively more severe and complex because of referring bias. Third, the study is cross-sectional and cannot describe a causal relationship. Further prospective multicenter cohort studies are necessary to delineate this issue. Last but not least, in this study, no Borden or Cognard grading of DAVF was evaluated, nor was the association of DAVF grading or retrograde venous drainage with CVT investigated. The correlation between DAVF grading or retrograde venous drainage and venous stenosis or thrombosis warrants further study.

CONCLUSION

This patient cohort showed that the prevalence of CVT in Chinese adult DAVF was 19.8% and blurred vision, venous cerebral infarction, and LS-DAVF and MDAVF might be the risk factors for the coexistence of CVT in patients with DAVF. We suggest that for these patients, it is important to search the signs of thrombosis on DSA or high-resolution magnetic resonance thrombosis imaging maps.

Funding

This study received funding from the Beijing Natural Science Foundation (7212047) and the National Natural Science Foundation Grants (82171297).

Disclosures

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

COMMENTS

The manuscript titled “Dural Arteriovenous Fistulas with or without Cerebral Venous Thrombosis: a cross-sectional analysis of 511 patients” aims to investigate the correlation between cerebral venous thrombosis (CVT) and dural arteriovenous fistulas (DAVFs). Previous studies have hinted at a potential link between CVT and DAVFs. Through a retrospective analysis of this substantial patient cohort, the authors reported that patients presenting with both CVT and DAVFs more frequently exhibited blurry vision, venous cerebral infarction, the involvement of a large sinus (transverse, sigmoid, and sagittal) by the DAVF, and the existence of multiple DAVFs.

While the study's quality and coherence are commendable, its clinical utility remains somewhat limited. Despite its scientific significance, the findings do not provide definitive guidance for diagnosing concurrent CVT in patients with DAVFs. Notably, in the multivariable analysis, the only neurological symptom significantly correlated with CVT was blurred vision, which is a vague and nonspecific symptom. This raises the crucial question of whether all patients with DAVFs should undergo magnetic resonance venography or computed tomography venography or if these tests should be reserved for patients with involvement of large sinuses, blurred vision, or cerebral infarction. I would argue that most patients with DAVFs should undergo catheter digital subtraction angiography because it provides critical information about the angio-architecture of the fistula and the patency of the main venous sinuses. This information is of utmost importance because the management of CVT may necessitate anticoagulation, contingent on the acuity of the thrombosis.

It is worth noting that the association between DAVFs and CVT was also reported by the International Cerebral Venous Thrombosis Consortium.^1a However, in their study, the qualifying event was sinus thrombosis rather than the presence of a DAVF. The presence of a DAVF in this larger consortium was relatively low (2.4%, 29 out of 1218 cases) and was predominantly associated with chronic CVT, older age, and male gender. This broader data set suggests that DAVFs are relatively uncommon in patients with CVT.

Edgar A. Samaniego

Iowa City, Iowa, USA