The learning curve effect on embolization of cerebral dural arteriovenous fistula single-center experience in...

Cerebral dural arteriovenous fistulas (DAVFs) are rare intracranial vascular lesions but can cause significant morbidity and mortality.

To analyze the effect of the center's experience on DAVF embolization efficacy and safety.

From May 2008 to October 2014, 57 embolization procedures were attempted on 48 patients (37 men and 11 women; median age: 63.9 years) for DAVF in a single center. DAVF presented with cortical venous reflux in 44/48 cases (91.7%) and hemorrhagic manifestation in 21/48 cases (43.75%). Angiographic occlusion quality, whether complete or incomplete (efficacy), and neurological complications (safety) were recorded. The patient population was divided into four consecutive quartiles during the inclusion period to assess the progress profile. Efficacy and safety outcomes were compared with Fisher's test.

A logistic regression was performed to explore a learning curve phenomenon, showing a significant association between the chronological rank in the cohort and embolization efficacy (P=0.007). Significant differences were found between first and last quarter (P=0.036). The endovascular technique involved an arterial injection of Onyx^® in 36/48 cases (75%), administered via the middle meningeal artery in 25/36 cases (69.5%). The complete occlusion rate improved significantly from 33.3% for the first quartile of the population, to 75.0% for the 2nd and 3rd quartiles and 83.3% for the last quartile. Neurological complications were found in 7/48 patients (14.6%), the rate decreased by 41.7% to 16.7%, without statistically difference.

The efficacy and safety of DAVF embolization improved with the experience gained at the center, suggesting the existence of a learning curve.