RESEARCH—HUMAN—CLINICAL STUDIES RESEARCH—HUMAN—CLINICAL STUDIES
Endovascular Treatment of Cavernous Sinus Dural Arteriovenous Fistula With Ipsilateral Inferior Petrosal Sinus Occlusion: A Single-Center Experience Jong Kook Rhim, MD* Young Dae Cho, MD* Jeong Jin Park, MD* Jin Pyeong Jeon, MD* Hyun-Seung Kang, MD, PhD‡ Jeong Eun Kim, MD, PhD‡ Won-Sang Cho, MD, PhD‡ Moon Hee Han, MD, PhD*‡ Dep ar tments of *R adiol og y and ‡Neurosurgery, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Korea Correspondence: Young Dae Cho, MD, Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, 28 Yongon-dong, Jongno-gu, Seoul, Korea 110-744. E-mail: [email protected] Received, October 28, 2014. Accepted, February 19, 2015. Published Online, March 31, 2015. Copyright © 2015 by the Congress of Neurological Surgeons.
BACKGROUND: Although a transvenous route via the ipsilateral inferior petrosal sinus (IPS) is preferred in treating cavernous sinus dural arteriovenous fistula (CSdAVF), this option may be limited if an occluded ipsilateral IPS undermines microcatheter delivery to the cavernous sinus. OBJECTIVE: To describe our experience with endovascular treatment of CSdAVF complicated by ipsilateral IPS occlusion. METHODS: From January 2003 through September 2014, a total of 49 CSdAVFs with ipsilateral IPS occlusion were identified in 49 patients, who then underwent endovascular treatment. Clinical and radiologic data were retrospectively reviewed. RESULTS: Either transvenous (n = 38) or transarterial (n = 11) access was initially elected, the latter reserved for single-hole or dominant arterial feeder fistulas. Access via occluded ipsilateral IPS was usually attempted (n = 34) by transvenous approach, with a 54.3% success rate. Anterior (n = 3) or posterior (n = 1) facial vein was alternatively used. Direct surgical exposure of ophthalmic vein (n = 3) or radiosurgery (n = 4) was performed for access failure or unsuccessful occlusion by other means. In 46 fistulas (93.9%), complete occlusion was achieved, with no procedure-related morbidity or mortality. Postprocedural symptom improvement was noted in all but 2 patients, who separately experienced paradoxical worsening of cranial nerve palsy and access failure. CONCLUSION: In patients with CSdAVF and ipsilateral IPS occlusion, various treatment strategies may be applied (given angioanatomic suitability), resulting in excellent procedural and short-term follow-up results. Reopening of an occluded IPS is reasonable as an initial access attempt. KEY WORDS: Cavernous sinus, Dural arteriovenous fistula, Embolization, Inferior petrosal sinus, Shunt Neurosurgery 77:192–199, 2015
DOI: 10.1227/NEU.0000000000000751
C
avernous sinus (CS) dural arteriovenous fistula (CSdAVF) is defined as an abnormal arteriovenous connection involving dura mater, within or near the walls of the CS.1 Such fistulas are fed chiefly by multiple dural branches from external and internal carotid arteries, draining into the CS.2 Although the natural course of a CSdAVF is relatively benign, with potential to spontaneously regress,3 highrisk lesions involving venous reflux require ABBREVIATIONS: CS, cavernous sinus; CSdAVF, cavernous sinus dural arteriovenous fistula; IPS, inferior petrosal sinus
192 | VOLUME 77 | NUMBER 2 | AUGUST 2015
www.neurosurgery-online.com
treatment because of ocular symptoms (chemosis, proptosis, periorbital pain, and eyelid swelling), cranial nerve deficits, venous hypertension (by cortical reflux), etc. Treatment of dAVF has recently improved, thanks to advances in devices and novel embolic material, but still entails 2 differing endovascular modalities routed through arteries or veins. Choice of modality depends on suitable angioanatomic conditions, with embolization via transvenous approach typically preferred for excellent outcomes. A transvenous route through the ipsilateral inferior petrosal sinus (IPS) provides a relatively straight course and is the shortest route to the CS, so it is generally preferred. However, the ipsilateral IPS is
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
CAVERNOUS SINUS DURAL ARTERIOVENOUS FISTULA WITH IPSILATERAL INFERIOR PETROSAL SINUS OCCLUSION
sometimes occluded, thus impairing microcatheter delivery to the CS. Here, we present our experience in treating patients with CSdAVF and ipsilateral IPS occlusion, focusing on alternative treatment strategies and methods of assessing radiologic and clinical outcomes by treatment modality.
TABLE. General Characteristics of Study Population (n = 49)a
METHODS Patient Population In a 12-year period (January 2003-September 2014), 123 patients with 142 CSdAVFs (bilateral, n = 19) underwent endovascular treatment at our institute. A total of 49 CSdAVFs (34.5%) complicated by ipsilateral IPS occlusion at the time of treatment were included in this study (women, 34; men, 15; mean age, 57.2 6 10.8 years). General characteristics of the cohort are summarized in the Table. Therapeutic alternatives were discussed with both neurological/neurosurgical and neurointerventional teams in a multidisciplinary decision-making process, and informed consent was obtained from all patients after careful consultation. The study protocol was approved by our Institutional Review Board.
Endovascular Procedures Endovascular procedures were largely performed under local anesthesia (n = 41), with general anesthesia used for only 8 patients. Before treatment, angioarchitectural explorations were conducted via the Integris V (Philips Medical System, Best, the Netherlands) or the Innova IGS 630 (GE Healthcare, Wauwatosa, Wisconsin) biplane system, including bilateral internal and external carotid angiographies to assess feeding arteries, location and extent of fistulous sites, and venous parameters (drainage paths and patterns). After thorough investigation of angiographic anatomy, initial access routes were determined. All lesions were categorized as focal or diffuse type, depending on whether fistulous sites (mural channels) were limited in scope or diffusely involved areas of the CS. In addition, Cognard classifications were assigned. Antiplatelet preparations were not routinely prescribed in advance of procedures, but after femoral sheath placement, systemic heparinization was achieved with single 2000-IU injections. For all transvenous embolizations, pushable fibered coils were used, first introducing 5F angiocatheters into external carotid arteries by way of main feeding arteries and then navigating 6F guiding catheters into jugular veins (unless direct ophthalmic vein access was used). Microcatheters for delivery of pushable coils were placed in the CS near fistulous sites. After confirmation of angioanatomic AVF configurations by selective angiography, embolizations were performed using pushable coils. Most transarterial embolizations involved use of n-butyl cyanoacrylate (Cordis Neurovascular Inc, Miami Lakes, Florida), except in 2 patients in whom Onyx 18 (Covidien, Mansfield, Massachusetts) or pushable coil was elected. For transarterial approach, 6F guiding catheters were placed in the external carotid artery through main feeding arteries, and microcatheters for delivery of embolic materials were navigated into feeding arteries (as far as feasible) under microguidewire guidance. After confirmation of angioanatomic configurations and exclusion of any dangerous anastomoses by selective angiography, embolizations were performed.
Angiographic Outcome and Follow-up States Immediate angiographic results after endovascular embolization were classified by degree of shunt as follows: complete occlusion (no shunt),
NEUROSURGERY
Characteristics
Cases, n (%)
Age, mean (range), y Sex Male Female Type Idiopathic Traumatic Surgical Initial symptoms Ocular Cranial nerve palsy Headache Mixed Cognard classification IIa IIa 1 IIb III IV V Side of IPS occlusion Ipsilateral Bilateral Drainage vein Ophthalmic vein SMCV Contralateral IPS Pterygoid plexus SPS Fistula type Focal Diffuse Initial route Ipsilateral IPS Transfacial Transarterial Reopening of occluded IPS (n = 34) Yes No Immediate procedural outcome Complete Nearly complete Partial Access failure Follow-up outcome Cured Residual shunt Symptom improvement Yes No change Aggravation
30-80 (57.2) 15 (30.6) 34 (69.4) 45 (91.8) 2 (4.1) 2 (4.1) 27 6 5 11
(55.1) (12.2) (10.2) (22.4)
39 6 1 2 1
(79.6) (12.2) (2) (4.1) (2)
29 (59.2) 20 (40.8) 34 7 3 1 4
(69.4) (14.3) (6.1) (2) (8.2)
9 (18.4) 40 (81.6) 34 (69.4) 4 (8.2) 11 (22.4) 19 (55.9) 15 (44.1) 34 8 5 2
(69.4) (16.3) (10.2) (4.1)
46 (93.9) 3 (6.1) 47 (96.0) 1 (2.0) 1 (2.0)
a
IPS, inferior petrosal sinus; SMCV, superficial middle cerebral vein; SPS, superior petrosal sinus.
near-complete occlusion (small residual shunt with marked reduction in volume and velocity), and partial occlusion (large residual shunt with slight reduction or no change in volume and velocity).
VOLUME 77 | NUMBER 2 | AUGUST 2015 | 193
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
RHIM ET AL
In instances of near-complete or partial occlusion, ipsilateral carotid compression was recommended for at least 2 weeks. Clinical follow-up at 1 and 3 months after treatment was advised if fistulas were completely occluded. Only in patients with aggravated clinical symptoms were further imaging studies such as digital subtraction angiography or magnetic resonance angiography recommended. In those with near-complete or partial occlusion after treatment, follow-up digital subtraction angiography was recommended 1 month after procedures to confirm progressive occlusion or to opt for further treatment if necessary. Clinical outcomes were gauged by degree of symptomatic improvement after treatment as follows: improvement, no change, and aggravation. Assessments were done during hospitalization periods and in outpatient clinics at 1 and 3 months after treatment.
RESULTS Patient Population Most instances of CSdAVFs were idiopathic in nature (n = 45), with occasional traumatic (n = 2) or surgical (n = 2) causes. The most common complaints were orbital symptoms such as proptosis and conjunctival injection (n = 44), followed by diplopia resulting from cranial nerve palsy (n = 31), headache (n = 6), and tinnitus (n = 6). Some patients (n = 20) experienced multiple symptoms (.2) in combination, and 2 patients complained of all symptoms. There were no instances of intracranial hemorrhage. Right-sided (n = 24, 49.0%) and leftsided (n = 25, 51.0%) fistulas were similar in frequency, but diffuse-type CSdAVFs predominated (diffuse, 40; focal, 9). By Cognard classification, type IIa was most common (n = 39), followed by IIa 1 IIb (n = 6), IV (n = 2), III (n = 1), and V (n = 1). Ophthalmic veins (n = 34) usually provided drainage, followed by the superficial middle cerebral vein (n = 7), superior petrosal sinus
(n = 4), contralateral IPS (n = 3), and pterygoid plexus (n = 1). Among patients with ophthalmic venous drainage, anterior facial drainage (n = 26) decisively surpassed the posterior route (n = 2). In 6 instances, both routes of facial venous drainage were evident (Table). Procedural and Follow-up Results Transvenous Route of Initial Access Of the 49 patients with CSdAVF and ipsilateral IPS occlusion, transvenous access was elected as the initial route in 38 (77.6%). A schematic of transvenous treatment strategies is shown in Figure 1. In these 38 patients, attempts to reopen occluded ipsilateral IPS usually were made (n = 34, 89.5%). The success rate in accessing fistulas of the CS by microcatheter was 54.3% (n = 19). Embolization was curative in all fistulas accessible by microdevices, with the exception of 1 partially treated lesion. Here, a supplemental contralateral IPS approach later proved effective. Access failure in 15 patients prompted searches for alternative transvenous routes to accommodate key angioanatomic features such as patterns/patency of draining veins and calibers/routes of feeding arteries. Of these 15 patients, 11 chose other venous routes, ordered by frequency as follows: anterior facial vein (n = 5), contralateral IPS (n = 4), and posterior facial vein (n = 2). Occlusion succeeded in 7 patients but failed in 4 patients, who eventually opted for transarterial embolization (n = 2), radiosurgery (n = 1), or direct exposure of ophthalmic vein (n = 1). Another route via facial vein was elected in 4 patients with prominent facial venous drainage (Figure 2). Access through anterior (n = 3) and posterior (n = 1) facial veins was attempted, resulting in complete (n = 2) or partial (n = 1) occlusion in 3 of
FIGURE 1. Algorithm for the initial transvenous approach in the study population. Ant, anterior; IPS, inferior petrosal sinus; Post, posterior; OV, ophthalmic vein.
194 | VOLUME 77 | NUMBER 2 | AUGUST 2015
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
CAVERNOUS SINUS DURAL ARTERIOVENOUS FISTULA WITH IPSILATERAL INFERIOR PETROSAL SINUS OCCLUSION
FIGURE 2. Various transvenous routes based on angioanatomic dictates (A) occluded inferior petrosal sinus (IPS), (B) contralateral patent IPS, (C) anterior facial vein, (D) posterior facial vein, and (E) surgically exposed ophthalmic vein.
the 4 patients. Transarterial embolization was added in the latter instance. In the remaining patient, access failure after facial vein and occluded IPS attempts was remedied by direct surgical exposure of ophthalmic vein, which proved curative (Figure 1). In summary, nearly all patients undergoing transvenous embolization as the initial access route were successfully treated. However, residual shunt persisted in 2 patients, including 1 in whom access was never achieved. Although transvenous access through a single or multiple channels failed in 7 patients, there were no procedurerelated complications. Ultimately, transarterial embolization was performed in 6 of these patients, achieving successful (or progressive) occlusion in 5 patients. The last patient was cured through radiosurgery. Most patients registered symptom improvement, although 1 patient suffered paradoxical worsening of cranial nerve palsy, and another was plagued by access failure. Transarterial Route of Initial Access Transarterial access was elected as initial route in 11 patients in whom fistulas were single-hole type (n = 4) or had single or
NEUROSURGERY
dominant arterial feeder(s) (n = 7; Figure 3). Feeding arteries in these instances, ordered by frequency, were as follows: ascending pharyngeal (n = 4), accessory meningeal (n = 4), middle meningeal (n = 3), and internal carotid artery dural branch (n = 2). Embolization in 2 patients involved 2 feeding arteries. n-Butyl cyanoacrylate was primarily used for embolization (n = 9), followed by Onyx (n = 1) and detachable coil (n = 1). Complete occlusion was achieved in 6 patients, with partial occlusion in the other 7; and 2 patients experienced procedural complications. One suffered arterial injury during catheterization (treated by embolization), and asymptomatic distal branch occlusion of MCA occurred in the other as a result of migration of glue through a dangerous anastomosis. Of the 5 patients with residual shunt after transarterial embolization, transvenous access via occluded ipsilateral IPS was attempted in 2 patients, and 3 underwent radiosurgery. The former patients required either supplemental transarterial embolization or embolization through direct surgical exposure of ophthalmic vein because of access failures. Complete occlusion was achieved in most patients,
VOLUME 77 | NUMBER 2 | AUGUST 2015 | 195
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
with 1 instance of partial occlusion, and all patients subjected to transarterial embolization showed symptom improvement. An algorithm for transarterial approach as the first-line treatment is presented in Figure 4. Overall Outcome Immediately after endovascular management, whether by transvenous, transarterial, or a combined approach, outcomes were as follows: complete occlusion, 34; near-complete occlusion, 8;
Endovascular Treatment of Cavernous Sinus Dural Arteriovenous Fistula With Ipsilateral Inferior Petrosal Sinus Occlusion: A Single-Center Experience Jong Kook Rhim, MD* Young Dae Cho, MD* Jeong Jin Park, MD* Jin Pyeong Jeon, MD* Hyun-Seung Kang, MD, PhD‡ Jeong Eun Kim, MD, PhD‡ Won-Sang Cho, MD, PhD‡ Moon Hee Han, MD, PhD*‡ Dep ar tments of *R adiol og y and ‡Neurosurgery, Seoul National University Hospital, Seoul National University College of Medicine, Seoul, Korea Correspondence: Young Dae Cho, MD, Department of Radiology, Seoul National University Hospital, Seoul National University College of Medicine, 28 Yongon-dong, Jongno-gu, Seoul, Korea 110-744. E-mail: [email protected] Received, October 28, 2014. Accepted, February 19, 2015. Published Online, March 31, 2015. Copyright © 2015 by the Congress of Neurological Surgeons.
BACKGROUND: Although a transvenous route via the ipsilateral inferior petrosal sinus (IPS) is preferred in treating cavernous sinus dural arteriovenous fistula (CSdAVF), this option may be limited if an occluded ipsilateral IPS undermines microcatheter delivery to the cavernous sinus. OBJECTIVE: To describe our experience with endovascular treatment of CSdAVF complicated by ipsilateral IPS occlusion. METHODS: From January 2003 through September 2014, a total of 49 CSdAVFs with ipsilateral IPS occlusion were identified in 49 patients, who then underwent endovascular treatment. Clinical and radiologic data were retrospectively reviewed. RESULTS: Either transvenous (n = 38) or transarterial (n = 11) access was initially elected, the latter reserved for single-hole or dominant arterial feeder fistulas. Access via occluded ipsilateral IPS was usually attempted (n = 34) by transvenous approach, with a 54.3% success rate. Anterior (n = 3) or posterior (n = 1) facial vein was alternatively used. Direct surgical exposure of ophthalmic vein (n = 3) or radiosurgery (n = 4) was performed for access failure or unsuccessful occlusion by other means. In 46 fistulas (93.9%), complete occlusion was achieved, with no procedure-related morbidity or mortality. Postprocedural symptom improvement was noted in all but 2 patients, who separately experienced paradoxical worsening of cranial nerve palsy and access failure. CONCLUSION: In patients with CSdAVF and ipsilateral IPS occlusion, various treatment strategies may be applied (given angioanatomic suitability), resulting in excellent procedural and short-term follow-up results. Reopening of an occluded IPS is reasonable as an initial access attempt. KEY WORDS: Cavernous sinus, Dural arteriovenous fistula, Embolization, Inferior petrosal sinus, Shunt Neurosurgery 77:192–199, 2015
DOI: 10.1227/NEU.0000000000000751
C
avernous sinus (CS) dural arteriovenous fistula (CSdAVF) is defined as an abnormal arteriovenous connection involving dura mater, within or near the walls of the CS.1 Such fistulas are fed chiefly by multiple dural branches from external and internal carotid arteries, draining into the CS.2 Although the natural course of a CSdAVF is relatively benign, with potential to spontaneously regress,3 highrisk lesions involving venous reflux require ABBREVIATIONS: CS, cavernous sinus; CSdAVF, cavernous sinus dural arteriovenous fistula; IPS, inferior petrosal sinus
192 | VOLUME 77 | NUMBER 2 | AUGUST 2015
www.neurosurgery-online.com
treatment because of ocular symptoms (chemosis, proptosis, periorbital pain, and eyelid swelling), cranial nerve deficits, venous hypertension (by cortical reflux), etc. Treatment of dAVF has recently improved, thanks to advances in devices and novel embolic material, but still entails 2 differing endovascular modalities routed through arteries or veins. Choice of modality depends on suitable angioanatomic conditions, with embolization via transvenous approach typically preferred for excellent outcomes. A transvenous route through the ipsilateral inferior petrosal sinus (IPS) provides a relatively straight course and is the shortest route to the CS, so it is generally preferred. However, the ipsilateral IPS is
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
CAVERNOUS SINUS DURAL ARTERIOVENOUS FISTULA WITH IPSILATERAL INFERIOR PETROSAL SINUS OCCLUSION
sometimes occluded, thus impairing microcatheter delivery to the CS. Here, we present our experience in treating patients with CSdAVF and ipsilateral IPS occlusion, focusing on alternative treatment strategies and methods of assessing radiologic and clinical outcomes by treatment modality.
TABLE. General Characteristics of Study Population (n = 49)a
METHODS Patient Population In a 12-year period (January 2003-September 2014), 123 patients with 142 CSdAVFs (bilateral, n = 19) underwent endovascular treatment at our institute. A total of 49 CSdAVFs (34.5%) complicated by ipsilateral IPS occlusion at the time of treatment were included in this study (women, 34; men, 15; mean age, 57.2 6 10.8 years). General characteristics of the cohort are summarized in the Table. Therapeutic alternatives were discussed with both neurological/neurosurgical and neurointerventional teams in a multidisciplinary decision-making process, and informed consent was obtained from all patients after careful consultation. The study protocol was approved by our Institutional Review Board.
Endovascular Procedures Endovascular procedures were largely performed under local anesthesia (n = 41), with general anesthesia used for only 8 patients. Before treatment, angioarchitectural explorations were conducted via the Integris V (Philips Medical System, Best, the Netherlands) or the Innova IGS 630 (GE Healthcare, Wauwatosa, Wisconsin) biplane system, including bilateral internal and external carotid angiographies to assess feeding arteries, location and extent of fistulous sites, and venous parameters (drainage paths and patterns). After thorough investigation of angiographic anatomy, initial access routes were determined. All lesions were categorized as focal or diffuse type, depending on whether fistulous sites (mural channels) were limited in scope or diffusely involved areas of the CS. In addition, Cognard classifications were assigned. Antiplatelet preparations were not routinely prescribed in advance of procedures, but after femoral sheath placement, systemic heparinization was achieved with single 2000-IU injections. For all transvenous embolizations, pushable fibered coils were used, first introducing 5F angiocatheters into external carotid arteries by way of main feeding arteries and then navigating 6F guiding catheters into jugular veins (unless direct ophthalmic vein access was used). Microcatheters for delivery of pushable coils were placed in the CS near fistulous sites. After confirmation of angioanatomic AVF configurations by selective angiography, embolizations were performed using pushable coils. Most transarterial embolizations involved use of n-butyl cyanoacrylate (Cordis Neurovascular Inc, Miami Lakes, Florida), except in 2 patients in whom Onyx 18 (Covidien, Mansfield, Massachusetts) or pushable coil was elected. For transarterial approach, 6F guiding catheters were placed in the external carotid artery through main feeding arteries, and microcatheters for delivery of embolic materials were navigated into feeding arteries (as far as feasible) under microguidewire guidance. After confirmation of angioanatomic configurations and exclusion of any dangerous anastomoses by selective angiography, embolizations were performed.
Angiographic Outcome and Follow-up States Immediate angiographic results after endovascular embolization were classified by degree of shunt as follows: complete occlusion (no shunt),
NEUROSURGERY
Characteristics
Cases, n (%)
Age, mean (range), y Sex Male Female Type Idiopathic Traumatic Surgical Initial symptoms Ocular Cranial nerve palsy Headache Mixed Cognard classification IIa IIa 1 IIb III IV V Side of IPS occlusion Ipsilateral Bilateral Drainage vein Ophthalmic vein SMCV Contralateral IPS Pterygoid plexus SPS Fistula type Focal Diffuse Initial route Ipsilateral IPS Transfacial Transarterial Reopening of occluded IPS (n = 34) Yes No Immediate procedural outcome Complete Nearly complete Partial Access failure Follow-up outcome Cured Residual shunt Symptom improvement Yes No change Aggravation
30-80 (57.2) 15 (30.6) 34 (69.4) 45 (91.8) 2 (4.1) 2 (4.1) 27 6 5 11
(55.1) (12.2) (10.2) (22.4)
39 6 1 2 1
(79.6) (12.2) (2) (4.1) (2)
29 (59.2) 20 (40.8) 34 7 3 1 4
(69.4) (14.3) (6.1) (2) (8.2)
9 (18.4) 40 (81.6) 34 (69.4) 4 (8.2) 11 (22.4) 19 (55.9) 15 (44.1) 34 8 5 2
(69.4) (16.3) (10.2) (4.1)
46 (93.9) 3 (6.1) 47 (96.0) 1 (2.0) 1 (2.0)
a
IPS, inferior petrosal sinus; SMCV, superficial middle cerebral vein; SPS, superior petrosal sinus.
near-complete occlusion (small residual shunt with marked reduction in volume and velocity), and partial occlusion (large residual shunt with slight reduction or no change in volume and velocity).
VOLUME 77 | NUMBER 2 | AUGUST 2015 | 193
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
RHIM ET AL
In instances of near-complete or partial occlusion, ipsilateral carotid compression was recommended for at least 2 weeks. Clinical follow-up at 1 and 3 months after treatment was advised if fistulas were completely occluded. Only in patients with aggravated clinical symptoms were further imaging studies such as digital subtraction angiography or magnetic resonance angiography recommended. In those with near-complete or partial occlusion after treatment, follow-up digital subtraction angiography was recommended 1 month after procedures to confirm progressive occlusion or to opt for further treatment if necessary. Clinical outcomes were gauged by degree of symptomatic improvement after treatment as follows: improvement, no change, and aggravation. Assessments were done during hospitalization periods and in outpatient clinics at 1 and 3 months after treatment.
RESULTS Patient Population Most instances of CSdAVFs were idiopathic in nature (n = 45), with occasional traumatic (n = 2) or surgical (n = 2) causes. The most common complaints were orbital symptoms such as proptosis and conjunctival injection (n = 44), followed by diplopia resulting from cranial nerve palsy (n = 31), headache (n = 6), and tinnitus (n = 6). Some patients (n = 20) experienced multiple symptoms (.2) in combination, and 2 patients complained of all symptoms. There were no instances of intracranial hemorrhage. Right-sided (n = 24, 49.0%) and leftsided (n = 25, 51.0%) fistulas were similar in frequency, but diffuse-type CSdAVFs predominated (diffuse, 40; focal, 9). By Cognard classification, type IIa was most common (n = 39), followed by IIa 1 IIb (n = 6), IV (n = 2), III (n = 1), and V (n = 1). Ophthalmic veins (n = 34) usually provided drainage, followed by the superficial middle cerebral vein (n = 7), superior petrosal sinus
(n = 4), contralateral IPS (n = 3), and pterygoid plexus (n = 1). Among patients with ophthalmic venous drainage, anterior facial drainage (n = 26) decisively surpassed the posterior route (n = 2). In 6 instances, both routes of facial venous drainage were evident (Table). Procedural and Follow-up Results Transvenous Route of Initial Access Of the 49 patients with CSdAVF and ipsilateral IPS occlusion, transvenous access was elected as the initial route in 38 (77.6%). A schematic of transvenous treatment strategies is shown in Figure 1. In these 38 patients, attempts to reopen occluded ipsilateral IPS usually were made (n = 34, 89.5%). The success rate in accessing fistulas of the CS by microcatheter was 54.3% (n = 19). Embolization was curative in all fistulas accessible by microdevices, with the exception of 1 partially treated lesion. Here, a supplemental contralateral IPS approach later proved effective. Access failure in 15 patients prompted searches for alternative transvenous routes to accommodate key angioanatomic features such as patterns/patency of draining veins and calibers/routes of feeding arteries. Of these 15 patients, 11 chose other venous routes, ordered by frequency as follows: anterior facial vein (n = 5), contralateral IPS (n = 4), and posterior facial vein (n = 2). Occlusion succeeded in 7 patients but failed in 4 patients, who eventually opted for transarterial embolization (n = 2), radiosurgery (n = 1), or direct exposure of ophthalmic vein (n = 1). Another route via facial vein was elected in 4 patients with prominent facial venous drainage (Figure 2). Access through anterior (n = 3) and posterior (n = 1) facial veins was attempted, resulting in complete (n = 2) or partial (n = 1) occlusion in 3 of
FIGURE 1. Algorithm for the initial transvenous approach in the study population. Ant, anterior; IPS, inferior petrosal sinus; Post, posterior; OV, ophthalmic vein.
194 | VOLUME 77 | NUMBER 2 | AUGUST 2015
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
CAVERNOUS SINUS DURAL ARTERIOVENOUS FISTULA WITH IPSILATERAL INFERIOR PETROSAL SINUS OCCLUSION
FIGURE 2. Various transvenous routes based on angioanatomic dictates (A) occluded inferior petrosal sinus (IPS), (B) contralateral patent IPS, (C) anterior facial vein, (D) posterior facial vein, and (E) surgically exposed ophthalmic vein.
the 4 patients. Transarterial embolization was added in the latter instance. In the remaining patient, access failure after facial vein and occluded IPS attempts was remedied by direct surgical exposure of ophthalmic vein, which proved curative (Figure 1). In summary, nearly all patients undergoing transvenous embolization as the initial access route were successfully treated. However, residual shunt persisted in 2 patients, including 1 in whom access was never achieved. Although transvenous access through a single or multiple channels failed in 7 patients, there were no procedurerelated complications. Ultimately, transarterial embolization was performed in 6 of these patients, achieving successful (or progressive) occlusion in 5 patients. The last patient was cured through radiosurgery. Most patients registered symptom improvement, although 1 patient suffered paradoxical worsening of cranial nerve palsy, and another was plagued by access failure. Transarterial Route of Initial Access Transarterial access was elected as initial route in 11 patients in whom fistulas were single-hole type (n = 4) or had single or
NEUROSURGERY
dominant arterial feeder(s) (n = 7; Figure 3). Feeding arteries in these instances, ordered by frequency, were as follows: ascending pharyngeal (n = 4), accessory meningeal (n = 4), middle meningeal (n = 3), and internal carotid artery dural branch (n = 2). Embolization in 2 patients involved 2 feeding arteries. n-Butyl cyanoacrylate was primarily used for embolization (n = 9), followed by Onyx (n = 1) and detachable coil (n = 1). Complete occlusion was achieved in 6 patients, with partial occlusion in the other 7; and 2 patients experienced procedural complications. One suffered arterial injury during catheterization (treated by embolization), and asymptomatic distal branch occlusion of MCA occurred in the other as a result of migration of glue through a dangerous anastomosis. Of the 5 patients with residual shunt after transarterial embolization, transvenous access via occluded ipsilateral IPS was attempted in 2 patients, and 3 underwent radiosurgery. The former patients required either supplemental transarterial embolization or embolization through direct surgical exposure of ophthalmic vein because of access failures. Complete occlusion was achieved in most patients,
VOLUME 77 | NUMBER 2 | AUGUST 2015 | 195
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited
with 1 instance of partial occlusion, and all patients subjected to transarterial embolization showed symptom improvement. An algorithm for transarterial approach as the first-line treatment is presented in Figure 4. Overall Outcome Immediately after endovascular management, whether by transvenous, transarterial, or a combined approach, outcomes were as follows: complete occlusion, 34; near-complete occlusion, 8;
