Extended resection of hemosiderin fringe is better for seizure outcome: A study in patients with cavernous malformation associated with refractory epilepsy
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.115070
Source of Support: None, Conflict of Interest: None
Background: Cerebral cavernous malformation (CCM) is frequently associated with intractable focal epilepsy. Epileptogenicity is usually attributed to the hemosiderin deposits. Extent of resection is a crucial issue for achieving good seizure outcome. Aim: To assess whether seizure outcome is related to the extended resection (ER) of surrounding hemosiderin fringe brain tissue. Materials and Methods: Between April 2000 and April 2008, 132 patients with CCM and refractory epilepsy were scheduled for surgery based on the high-resolution magnetic resonance imaging (MRI) findings and intensive video-electroencephalogram (EEG) monitoring. All patients underwent pre- and post-operative MRI. Based on MRI findings patients were grouped into: ER group (ER, hemosiderin completely removed) and lesionectomy group (LE, hemosiderin not/partially removed). Post-operative seizure outcome was compared between the two groups based on Engel and the International League Against Epilepsy outcome scales. Results: At 1-year follow-up of the 86 patients in the ER group, 54 (74.4%) achieved seizure free outcome and in the LE group of the 46 patients, 20 (59.5%) achieved seizure-free outcome. At 5-year follow-up, 59.5% (25/42) of patients in ER group and 27.8% (5/18) of patients in LE group achieved seizure-free outcome. ER was not associated with increased neurological morbidity. Conclusions: Our study suggests that complete removal of hemosiderin fringe brain tissue surrounding CCMs may improve short-term and long-term seizure outcome.
Keywords: Cerebral cavernous malformation, hemosiderin, outcome, refractory epilepsy
Seizures are the most common presenting symptom of supratentorial cerebral cavernous malformations (CCMs) and occur in 35-70% of symptomatic patients. Approximately, 40% of these patients have medically refractory epilepsy. ,,, Epilepsy related to CCM accounts for 4% of all refractory focal epilepsy.  Microsurgical resection of CCM is the best treatment option in patients with medically refractory epilepsy. ,,,, However, there is a controversy whether excision of the hemosiderin fringe typically surrounding CCMs is necessary to achieve seizure freedom. The limitations of the studies reporting seizure outcome following the resection of CCM along with the excision of hemosiderin fringe are diverse clinical characteristics of the patients and small sample size. We review our experience with 132 patients who underwent microsurgical resection of CCM whether excision of the hemosiderin fringe surrounding CCMs had a better seizure outcome.
Patient population included 259 consecutive patients who were treated for epilepsy related to CCM between April 2000 and April 2008. Inclusion criteria were: (a) Well-established refractory epilepsy defined as for the criteria proposed by the International League Against Epilepsy (ILAE);  (b) demonstration of supratentorial CCM, single or multiple, by computed tomography or magnetic resonance imaging (MRI); (c) presence of hemosiderin rim surrounding CCM; (d) availability of both pre- and post-operative MRIs; (e) histological confirmation of CCM; (f) resection of CCMs by microsurgical technique; and (g) at least, 1-year of follow-up. Patients excluded from the study were those who had gross lobar resections and hippocampectomy. The study population was divided into 2 groups: (a) Extended resection group (ER, hemosiderin completely removed) and (b) lesionectomy group (LE, hemosiderin not/partially removed).
Clinical variables analyzed included: Age at presentation, gender, age at seizure onset, duration of seizures before surgery, type of seizures, localization and size of CCM. Pre- and post-operative MRIs were obtained, In addition to multiplanar spin echo sequences with T1- and T2-weighted images, gradient-echo T2-weighted images and susceptibility weighted imaging were carried out to delineate perilesional hemosiderin better. The MR images were assessed by two experienced Neuroradiologists for testing presence of hemosiderin surrounding the cavernoma before and after surgery. In addition seizure outcome, we also looked for any surgical sequelae and cavernoma bleeding during the follow-up.
The short-term outcome measure was seizure status at 1-year after surgery and was assessed by using Engel's classification and the ILAE: Engel IA and ILAE1-seizure-free; Engel IB-IIIB and ILAE 2-5 - drastic reduction in the frequency and severity of seizures; and Engel IVA-IVB and ILAE 5 no improvement or worsening. We also assessed the long-term seizure outcome at 5 years after surgery.
Statistical analyses were performed with the χ2 test and Mann-Whitney U test. Statistical significance was defined as P < 0.05.
Demographic and clinical data
Of the 259 patients recruited from the database, 127 patients were excluded from the analysis because: 114 patients did not meet the criteria for medically intractable epilepsy; 5 patients had gross lobar resections; 6 patients had had hippocampectomy and 2 patients lost for follow-up. Of 132 reaming patients, the mean age was 39.3 years (range: 12-69 years), 64 (48.5%) were female and 68 (51.5%) and mean age of seizure onset was 37.0 years (range: 8-65 years). Seizure type was simple partial seizure in 45 patients, complex partial seizure in 40 patients and 47 patients had secondary generalized tonic-clonic seizures. In 84 patients, electroencephalogram (EEG) revealed a unifocal finding with a topographic relation to the CCM and 48 patients had multifocal findings.
The most frequent location was temporal (n = 51) followed by parietal (n = 35), frontal (n = 31) and occipital (n = 8). Multiple cavernomas were observed in seven patients. The maximal cavernoma diameter ranged from 0.4 to 5.3 cm (mean, 2.9 cm), in 74 patients the maximum diameter was <1.5 cm and in 68 it was ≥1.5 cm. On pre-operative MRI evidence of fresh or old hemorrhage was found in 39 (29.5%) patients. Hemosiderin deposits surrounding the CCM were detected by MRI in all patients. The CCM was completely removed in all patients. Hemosiderin was completely removed in 86 (65.2%) patients and partial or no removal in 46 (34.8%).
There were no significant differences in the baseline characteristics between both groups [Table 1].
Seizure outcome and follow-up
At 1-year follow-up all patients were examined 1-year and at 5-year follow-up 72 patients were lost for follow-up. Of the remaining 60 patients, 42 patients were in the ER group, 18 patients were in the LE group.
At 1-year follow-up, of the 86 patients in the ER group, 64 (74.4%) patients remained free of seizures; 14 (16.3%) had drastic reduction in the seizure frequency and severity of seizures; in 8 (9.3%) no improvement was seen. Of the 46 in LE group, 25 (54.3%) patients remained seizure free, 9 (19.6%) had drastic reduction in the frequency and severity of seizures and 12 (26.1%) patients had no improvement. At 5-year follow-up, of the 42 patients in the ER group, 25 (59.5%) patients had seizure free outcome, 13 (31%) had drastic reduction in the frequency and severity of seizures and 4 (9.5%) had no improvement.
Of the 18 patients in the LE group, 5 (27.8%) had seizure free outcome, 10 (55.5%) had drastic reduction in the frequency and severity of seizures and no improvement in 3 (16.7%) [Figure 1]. Seizure outcome was significantly difference between the two groups, patients with complete excision of hemosiderin fringe doing better both at 1-year and 5-year follow-up (74.4% vs. 54.3% in 1-year follow-up, χ2 (1) =5.497, P = 0.02; 59.5% vs. 27.8% in 5 years follow-up, χ2 (1) =5.653, P = 0.02).
To identify whether ER was more useful than pure LE, we took location of CCM as "center effect" and employed the Cochran-Mantel-Haenszel Chi-squared test, yet we could not find ER temporal lobe lesions was more helpful. (χ2CMH = 3.9195, P = 0.15) [Table 2].
Post-operative patients remained on the antiepileptic drugs (AEDs), which they were taking, usually for 6 months. If there was no seizure recurrence during this period, the patients were advised to decrease the dosage or discontinue the medication.
Neurological outcome and post-operative deficits
Pre-operative neurologic examination was normal in 115 patients. Three patients had partial motor aphasia and 6 patients had ataxia, which improved after operation. Eight patients had hemiparesis and 3 had improved and 5 remained unchanged after operation.
In the immediate post-operative period, 15 (11.4%) patients developed new neurological symptoms. Of these 15 patients, in 5 (33.3%) patients, some minor neurological deficits persisted at follow-up. Ten patients had some neurological deficits, in 8 (9.3%) patients in the ER group and in 2 (4.3%) patients in LE group at 1 year follow-up (5 hemiparesis, 1 sensory disturbances and 3 had mild dysphasia). The difference was not significant in between the two groups (P = 0.31).
CCM, a benign vascular lesion, is found in 0.4-0.5% of the population and it is being increasingly detected with the advent of MRI. ,,, Histologically, it characterized by abnormally enlarged blood vessels consisting of a single layer of epithelium without any brain tissue. ,, Common symptoms include: Epileptic seizures, microhemorrhages, focal deficits and headache. Seizures are the most frequent presenting sign in patients with symptomatic supratentorial lesions. ,,
Therapeutic options in patients with CCM presenting with epilepsy include conservative management and surgical intervention. AED are often the initial treatment and satisfactory seizure control may be achieved. However, surgery may become necessary in patients with medically intractable seizures. There is a controversy in the operative treatment of supratentorial CCMs whether to excise the hemosiderin fringe parenchyma surrounding the lesion. ,
In our study, the short-term and long-term seizure outcome was better in patients in whom hemosiderin fringe has been removed completely (P = 0.02 and P = 0.02, respectively). More patients had seizure free outcome with complete excision of hemosiderin fringe than with LE alone (74.4% vs. 54.3% at 1-year follow-up and 59.5% vs. 27.8% at 5-year follow-up).
Most of studies reported improved seizure outcomes with ER of hemosiderin fringe. ,,, In the study by Baumann et al.  seizure freedom at 3-year follow-up was 42% in patients in whom hemosiderin had been removed completely and 14% in patients in whom hemosiderin was partially or not removed. These results were further supported by the observations in the study by Hammen et al.  However, others have demonstrated no additional benefit over LE. , In the study by Zevgaridis et al.  seizure free was 86.9% in patients who had LE and 85.7% in patients who had LE and plus resection of the hemosiderin-stained tissue. The limitations of this study were; small sample size and inclusion of patients with infrequent seizures.
The reason for ER of the hemosiderin-stained tissue is that CCMs may not be the epileptogenic foci as there is no brain tissue in CCMs. Furthermore, the mass effect does not explain the high epileptogenicity of the lesion because other large size lesions such as malignant tumors are less commonly associated with medically refractory epilepsy.  The basic mechanism of epileptogenesis in CCM has not been fully understood. The assumption is that the substrate is chronic, clinically silent microhemorrhages. ,,, The hemoglobin breakdown products such as hemosiderin in the perilesional parenchyma result in abnormal presence of intracellular iron, which may lead to epileptogenesis by different mechanisms. Iron exposure in the adjacent tissue and the reactive glial proliferation is known to be epileptogenic. Under this premise, it is obvious that LE alone for intractable epilepsy is usually not sufficient and resection of both the CCM and the surrounding epileptogenic tissue is probably necessary.
Our study suggests that the long-term seizure outcome in patients with refractory epilepsy due to CCM is likely to be poor. In our study, the overall seizure-freedom at 1-year and at 5-year follow-up was 67.4% and 50% respectively. The reasons for this long-term decrease in seizure free outcome in a subset of patients are not entirely clear. Baumann et al.  felt that it may be related to early discontinuation AED medication. In our series, a lot of patients either decreased the dosage or discontinued the medication 6 months after operation. Some studies suggests that patient with good post-operative seizure control require long-term AED therapy, may be with fewer drugs and in lower dosage. , It quite possible that early discontinuation of AED therapy may be a reason for long-term post-operative poor seizure free outcome.
This study did not found any significant difference in seizure outcome between patients with temporal lobe location and location in other lobes. Unlike some previous studies, our data did not support the view that ER of temporal lobe lesions is more effective. , Studies with large sample size are needed to clarify these uncertainties. Our study showed ER of hemosiderin fringe surrounding CCMs may not be associated with increased neurological morbidity. If the lesion is involving or closer to an eloquent brain region, we do not recommend ER.
Although our study is retrospective, to our knowledge, it provides the largest comparative study between LE with or without ER of hemosiderin fringe in patients with cerebral cavernomas. The results of our study may have limited by statistical power, however, the study still provides some important clues. Probably a prospective and randomized study with the large sample size is needed to clarify some of the uncertainties.
[Table 1], [Table 2]