Seizures in Arteriovenous Malformations

Angioarchitectural characteristics

Univariate OR (95 % CI)

Multivariate OR (95 % CI)

Sensitivity (95 % CI)

Specificity (95 % CI)

Nidus size > 3 cm

3.79 (1.47–10.86)

–

61 (42–79)

71 (58–82)

Frontal, parietal, temporal

5.0 (1.50–19.07)

4.52 (0.95–21.47)

91 (82–100)

33 (20–47)

Arterial dilatation

–

100

22 (11–36)

Perinidal angiogenesis

3.71 (1.47–12.05)

–

70 (53–83)

61 (48–75)

Fistulous component

4.11 (1.45–13.20)

–

75 (59–91)

58 (44–71)

Pial long draining vein

14.86 (5.67–66.09)

5.71 (1.32–24.65)

79 (64.91)

80 (69–91)

Pseudophlebitic changes

3.08 (1.21–9.25)

–

61 (46–76)

67 (53–80)

Venous outflow stenosis

6.50 (2.22–32.49)

6.71 (1.99–22.56)

50 (34–69)

87 (76–96)

Venous ectasia

3.00 (1.05–13.75)

–

82 (70–94)

40 (27–56)

AVM arteriovenous malformation, CI confidence interval, NA not applicable, OR Odds Ratio

A scoring system to predict the risk for seizures was established [23] taking into account the three strongest predictors—restriction of venous drainage, location of AVM, and a long pial-draining vein. Each risk factor was given 1 point, thus establishing a scoring system of 0–3 points. A score of 0 had 97 % sensitivity and 93 % negative predictive value whereas a score of 3 had 98 % specificity and 80 % positive predictive value for seizure development. Although radiographic risk factors have been well established, there are limited data available on seizure types and the risk of development of epilepsy in patients with AVMs. Osipov et al. described 92 of 328 patients with AVMs that presented with seizures. Complex partial seizures were identified in 22 %, focal seizures with secondary generalization in 12 %, and generalized tonic-clonic seizures in 65 % of patients [29]. Recently, a prospective population study by Josephson et al. found that patients who did not have a history of seizures and experienced a first unprovoked seizure at presentation or during the follow-up had a 58 % chance of developing epilepsy over 60 person-years of follow-up. Sixty-five percent of patients with AVMs were started on antiepileptic drugs (AEDs) following the incident seizure [3]. There is a paucity of data to suggest any change in risk of rupture of AVM after the occurrence of seizures and the relationship of a high-flow state with seizure occurrence. Salman et al. performed a population-based study to determine the risk for a first-ever seizure in patients with AVMs. They found that the 5-year risk for seizure development was higher (23 %, 95 % CI 9–37 %) in patients presenting with an intracranial hemorrhage secondary to the malformation and a focal neurological deficit as compared to those with incidentally discovered AVMs (8 %, 95 % CI 0–20 %). The increased incidence after intracranial hemorrhage is often attributable to the size and location of the hemorrhage [30].

Imaging

The diagnosis of AVMs is usually based on clinical grounds and supported by imaging studies. Noninvasive head imaging may confirm the presence of large and conspicuous AVMs. CT scan of the head may initially be normal if no evidence of dilated veins or calcification appears (Fig. 7.1). Imaging modalities such as specialized brain MRI, MR angiography, and MR venography are often employed to identify the underlying arteriovenous communication. Besides the location, MR studies assist in evaluating the cytoarchitectural changes associated with large AVMs such as gliosis and involution of adjacent parenchymal tissue (Fig. 7.2). High venous pressure and venous distension and distortion may lead to a mass effect on important structures such as the brain stem. Invasive vascular studies such as catheter angiography reveal more detailed information and provide pathway to the management [31–33]. The angioarchitectural analysis includes the study of various components of the AVMs (arteries, nidus, and veins). It also includes the vascular response to chronic arteriovenous shunting (arterial stenosis, associated aneurysms, venous stenosis, and ectasias) (Fig. 7.3).

Fig. 7.1

CT scan of the head (a) showing dark flow voids in the left sylvian fissure in a young woman presented with headache and focal seizures affecting the language function. Cerebral angiography (b) shows a tangle of abnormal vessels supplied by arterial feeders and often drained by large dilated veins. (Case abstracted from: Neuroimaging in Neurology by Dr. David C. Preston, with permission)

Fig. 7.2

T2-weighted coronal MRI (a) demonstrating a tangle of blood vessels in the left frontal lobe, consistent with an arteriovenous malformation. Sagittal post-gadolinium-enhanced MRI (b) demonstrating an enhancing nidus of blood vessels in the left frontal lobe

Fig. 7.3

Catheter angiography of the right vertebral artery (arterial phase) (a) showing a large posterior AVM with multiple arterial feeders and dysplastic basilar-tip aneurysm measuring 11 mm (black arrow). The venous phase of the same angiogram (b) shows large and dysplastic veins (black arrow head) and dysplastic venous aneurysm (white arrow head)

The diagnosis of AVMs-associated seizures entails using sophisticated imaging techniques as well as electrophysiology to identify the epileptogenic zone. Routine EEG usually does not suffice due to its short duration and prolonged monitoring, or video EEG may be necessary to capture clinical as well as subclinical spells.

Treatment

Medical management of seizures secondary to AVMs involves utilizing one or more AED. Currently, there exist no treatment guidelines or recommendations for utilization of antiepileptic medication s for seizures in AVMs. Medical intractability in epilepsy is defined as the failure of two antiepileptic drugs to achieve full seizure control and must necessitate a surgical evaluation [34]. The Spetzler–Martin grading system as highlighted previously has been used widely to classify AVMs into grades based on size, eloquence of adjacent brain, and pattern of venous drainage [9]; a higher grade confers greater surgical risk. Different treatment approaches have been utilized to target an AVM lesion. Most instances, multiple approaches are required to obliterate a lesion. Endovascular treatment includes embolization technology usually coupled with either surgical resection or radiosurgery. Data regarding seizure freedom following either of these treatment approaches are conflicting:

Surgical resection and seizure outcomes

Piepgras et al. followed 110 of 280 patients with preoperative seizures secondary to AVMs for 2 years. The follow-up study revealed that 7 % of patients died, and of the survivors, 83 % were seizure-free while 17 % had intermittent seizures [35]. In another study by Yeh et al., 54 patients with a supratentorial unruptured AVM were followed postoperatively [36, 37]. Heros et al. evaluated 153 patients that underwent complete surgical excision of a cerebral AVM and were followed longitudinally for a mean period of 3.8 years. About 33 % of these patients had a seizure at presentation, 8.2 % of patients that did not have a seizure prior to surgery developed a postoperative seizure, and 7.1 % developed late seizures that were medically managed. About 50 % of those with history of preoperative seizures either improved or were seizure-free at follow-up [38]. In contrast, Murphy and colleagues found no difference in seizure frequency in patients that underwent surgical treatment of AVMs as compared to those who were managed medically [39].

Endovascular treatment and seizure outcomes

Several endovascular approaches to AVM treatment have been established, but few studies addressed seizure outcomes following embolization. Lv et al. treated 109 patients with cerebral AVMs endovascularly. Complete embolization was achieved in four patients. Thirty of 109 patients (27.5 %) experienced seizures before treatment. Seizure types were classified as generalized tonic-clonic seizures (56.7 %), simple partial (20 %), complex partial (3.3 %) seizures, or partial seizures with secondary generalization (20 %). Following endovascular treatment, seizure control over an unclear follow-up period was described as excellent in 21, good in 4, fair in 2, and poor in 3 patients. Eighteen patients required one embolization to achieve seizure freedom whereas others required multiple attempts at embolization [40].

Stereotactic radiosurgery and seizure outcomes

Stereotactic radiosurgery holds promise for AVM treatment, and results pertaining to seizure outcomes have been uplifting as compared to other treatment techniques. Schauble and colleagues followed 65 patients with single or multiple seizures who underwent AVM radiosurgery for more than a year after surgery. Median follow-up period was 48 months. Fifty one percent of patients were observed to be seizure- and aura-free at 3 years. Factors associated with seizure freedom were a low (< 4) Engel Seizure frequency score and small size of AVM [41]. However, studies have reported a delayed increase in seizure frequency in patients undergoing radiosurgery [42].

Multimodal approach and seizure outcome

Recently Josephson et al. [43] conducted the Scottish Intracranial Vascular Malformation Study (SIVMS) to prospectively evaluate and compare patients with cerebral AVMs that underwent conservative management or AVM treatment (surgery, endovascular, stereotactic, or multimodality). Of the 229 patients enrolled in the study, 44 (19 %) presented with epileptic seizures at onset, 68 % underwent AVM treatment, whereas 32 % were managed conservatively. There was no difference demonstrated in the 5-year risk of first or recurrent seizures between the two groups, and seizure at onset did not confer any additional risk for the development of seizures in these patients. Table 7.2 summarizes the risk of seizure recurrence and seizure-free intervals for various treatment modalities employed in their study. The limitations of the above study included possible confounding factors due to nonrandomization of patients. Results of the Randomized Trial of Unruptured Brain Arteriovenous (ARUBA) Malformations are eagerly awaited to confirm the impact of AVM treatment on seizures and other outcomes [44, 45].

Table 7.2

Risk of seizure following medical (conservative) or surgical treatment