Abstracts

Rationale: Brain AVM patients experience seizures in 20 to 30% of cases. 1 out of 100 unprovoked first seizures is caused by AVMs. AVMs are considered to be models for epileptogenic lesions as a result of perilesional gliosis, hemorraghe, steal phenomena, and iron deposits. The aim of this study was to define predictors for seizure outcome in a large prospective data base with specific focus on the role of intervention. Methods: Between 1982 and 2007, the University of Toronto Brain AVM Study Group treated 1106 patients with brain AVMs, of which 333 (30.1%) experienced seizures (SZ-Group). 155 patients met inclusion criteria for this study including a complete set of clinical and neuroradiological data. Mean ( SD) follow up was 7.35 5.43 years. Treatment consisted of surgical resection, radiosurgery or embolization, either alone or in combination. A total of 49 variables were examined; outcome was determined by Engel seizure outcome scale. These data were compared for statistical relevance to 50 AVM patients with no seizures (Non-SZ-group) over the course of the disease. Mean follow-up was 5.04 4.8 years. Results: The SZ-group was diagnosed with AVM at age 35 13 years compared to the Non-Sz-group with age 39 18 years. Infratentorial, occipital and deep white matter location were found in 46.0% in the Non-Sz-group versus 8.5% in the Sz-Group (P<0.01). Small sized AVMs occured in 68.1% of the Non-Sz-group vs. 43.8% in the Sz-group (p <0.01). The number of treatment procedures was significantly higher in Sz-group (P=0.01). 58.1% of the Sz-group became seizure free, and 30.0% stayed seizure-free after weaning off antiepileptic drugs (AED). Predictors for a good seizure outcome were generalized tonic-clonic seizures (P<0.01), intervention after single sz (P<0.01), and complete obliteration of the AVM (P<0.01). The Non-Sz group presented significantly more often with bleeding events (P<0.01). Factors significantly reducing chance of a favorable outcome (P<0.01) were focal seizures, particularly complex-partial seizures, transition from GTCS to CPS or SPS, and embolization in single modality treatment. Conclusions: Our data strongly suggest that an early diagnosis of AVM after single sz presentation and a subsequent rigorous treatment approach (leading to complete obliteration of the AVM) leads to the overall best seizure outcome. We speculate that delayed diagnosis may lead to secondary epileptogenesis and unfavourable seizure outcome.