Introduction/Purpose A carotid cavernous fistula (CCF) is a spontaneous or acquired connection between the carotid artery and cavernous sinus. These lesions classically present with chemosis, exopthalmos, and ocular bruit secondary to shunting of arterialized blood into the cavernous sinus and ophthalmic veins. Endovascular treatment of CCFs has evolved over time to include both transarterial and transvenous embolization using detachable balloons, coils, and liquid embolic agents. The present series comprises our 16-year institutional experience in the endovascular treatment of CCF.

Materials and Methods We reviewed our prospectively-maintained clinical database for patients who were evaluated for endovascular treatment for CCF between 12/1995 and 12/2011. Clinical and demographic data were extracted for 95 patients (40 direct CCF, 55 indirect CCF) from outpatient and inpatient medical records, operative notes, radiographic reports, and cerebral angiograms.

Results Of the 40 patients with direct CCF, 37 were successfully treated using only endovascular techniques (93%). Two patients underwent craniotomy, and one was managed conservatively. We observed an overall 10% morbidity and 2.5% mortality rate. From 12/1995 to 3/2004, detachable balloons were used in 18 of 22 cases of direct CCF. For patients treated with detachable balloons, 50% required carotid occlusion. Trans-arterial coil embolization, both with and without adjunctive balloon remodeling/stent assistance, was undertaken in 14 of 18 patients with direct CCF treated after 3/2004. In 10 patients undergoing transarterial coil embolization (71%), the internal carotid artery was preserved, whereas 4 (29%) required carotid occlusion. Of the 22 patients treated in the detachable balloon era, 4 (18%) exhibited residual fistula, 3 of which subsequently thrombosed. By comparison, of the 15 patients successfully treated in the modern era, 1 (7%) exhibited residual fistula. Transvenous approaches were utilized in 14 cases. In 11 cases, transarterial and transvenous approaches were utilized in tandem. Of the 55 indirect CCF cases, 42 were successfully treated using only endovascular techniques (78%), with transvenous embolization as the first-line treatment. The overall morbidity rate was 7.2%, and there were no deaths. The remaining 12 patients were not treated endovascularly secondary to inadequate fistula access. Treatment for these patients included manual compression (n=5), craniotomy (n=3), radiosurgery (n=3), and observation (n=1). Venous approaches included transfemoral access through the inferior petrosal sinus (n=19) or transfacial vein (n=8), and direct cutdown (n=7) or percutaneous (n=3) access to the ophthalmic veins. Onyx was used for stand-alone embolization in a single case. Of the 42 patients successfully-treated using endovascular procedures, 7 (17%) exhibited residual fistula following treatment, one of which went on to thrombose. Transarterial approaches were employed in 10 patients with indirect CCF, and three cases required both transvenous and transarterial approaches.

Conclusion Advances in endovascular technology have inspired an evolution in the treatment of CCF. For direct CCF, the lack of availability of detachable balloons has led to the adoption of trans-arterial coil embolization with adjunctive stent/balloon assistance to facilitate parent vessel preservation. For indirect CCF, advances in techniques of venous access have enabled successful treatment of lesions with restricted venous outflow.

Competing interests None.