Objective To present our experience with endovascular treatment of dural carotid cavernous fistulas (DCCFs) and determine if there is a correlation between clinical symptoms and the Cognard classification system.

Methods We searched our institutional neurovascular database to identify patients treated for DCCFs from January 1995 to May 2015. DCCFs were defined as a vascular shunt between meningeal branches of the internal carotid artery (ICA), external carotid artery (ECA), or both, draining into the cavernous sinus. Clinical symptoms were recorded based on clinical examination. Lesions were classified on angiography using the Cognard and Barrow classification systems. Treatment goal was defined as symptomatic cure based on clinical examination, cure of ophthalmic venous drainage, and cortical venous reflux on angiography.

Results The search revealed 37 patients with DCCFs; 32 DCCFs underwent endovascular treatment. The primary treatment goal was met in 30/32 (94%) lesions, with one neurologic complication (1/34; 3%). We identified 31 Cognard IIa fistulas, 4 Cognard IIa+b, 2 Cognard IIb, 7 Barrow B, 7 Barrow C, and 23 Barrow D fistulas. Eye redness, proptosis, and ocular pain were significantly lower in the Cognard IIb group than in the Cognard IIa and IIa+b groups (p=0.0015). Intracranial hemorrhage was more likely in the Cognard IIb group than in the Cognard IIa and IIa+b groups, with marginal significance (p=0.054). No correlation was seen between symptomatology and the Barrow classification.

Conclusions Endovascular treatment of DCCFs has a high degree of clinical success and a low complication rate. The Cognard system is suitable for grading DCCFs as it correlates with presenting symptomatology and venous drainage patterns. The Barrow classification adds no value in grading DCCFs.