Effects of New-Generation Antiepileptic Drug Prophylaxis on Delayed Neurovascular Events After Aneurysmal Subarachnoid Hemorrhage

Neuroelectric disruptions such as seizures and cortical spreading depolarization may contribute to the development of delayed cerebral ischemia (DCI) after aneurysmal subarachnoid hemorrhage (SAH). However, effects of antiepileptic drug prophylaxis on outcomes remain controversial in SAH. The authors investigated if prophylactic administration of new-generation antiepileptic drugs levetiracetam and perampanel was beneficial against delayed neurovascular events after SAH. This was a retrospective single-center cohort study of 121 consecutive SAH patients including 56 patients of admission World Federation of Neurological Surgeons grades IV − V who underwent aneurysmal obliteration within 72 h post-SAH from 2013 to 2021. Prophylactic antiepileptic drugs differed depending on the study terms: none (2013 − 2015), levetiracetam for patients at high risks of seizures (2016 − 2019), and perampanel for all patients (2020 − 2021). The 3rd term had the lowest occurrence of delayed cerebral microinfarction on diffusion-weighted magnetic resonance imaging, which was related to less development of DCI. Other outcome measures were similar among the 3 terms including incidences of angiographic vasospasm, computed tomography-detectable delayed cerebral infarction, seizures, and 3-month good outcomes (modified Rankin Scale 0 − 2). The present study suggests that prophylactic administration of levetiracetam and perampanel was not associated with worse outcomes and that perampanel may have the potential to reduce DCI by preventing microcirculatory disturbances after SAH. Further studies are warranted to investigate anti-DCI effects of a selective α-amino-3-hydroxy-5-methyl-4-isoxazole propionate receptor antagonist perampanel in SAH patients in a large-scale prospective study.


Introduction
Outcomes of subarachnoid hemorrhage (SAH) caused by a ruptured intracranial aneurysm remain poor in spite of medical advances [1].The outcome of SAH may rather worsen year by year as the population ages [2,3].Although both early brain injury, which is a deleterious effect of acute SAH on the brain, and delayed cerebral ischemia (DCI) have determined the outcome of aneurysmal SAH [4], the most important modifiable prognostic factor is DCI and the resultant cerebral infarction on computed tomography (CT) scan and magnetic resonance (MR) imaging [5].However, DCI is still a challenging condition consisting of various mechanisms including cerebral vasospasm and microcirculatory disturbance [6][7][8].
Recently, neuroelectric disruptions such as cortical spreading depolarization (CSD) and epileptiform activity have been revealed to be an important contributor to the development of DCI after SAH [9].CSDs and epileptiform activities may have similar toxic effects such as an increase in metabolic demand, arteriolar constriction via inverse neurovascular coupling, microthrombosis formation with platelet activation, and neuroinflammatory reactions, all of which may cause DCI [9,10].Because CSDs and epileptiform activities trigger each other to augment their actions and to worsen the metabolic supply-demand mismatch [9], some antiepileptic drugs may antagonize CSDs as well as epileptiform discharges and resultantly prevent DCI [11].In fact, recent experimental studies reported that some antiepileptic drugs had neuroprotective effects against post-SAH pathologies [12,13].However, it has been controversial if prophylactic administration of antiepileptic drugs has beneficial effects on outcomes in patients with aneurysmal SAH [14].At least, the use of old-generation antiepileptic drugs may not be warranted, because they have a high incidence of adverse events and many significant drug-drug interactions, which may aggravate outcomes [15].The aim of this retrospective study was thus to investigate if prophylactic administration of new-generation antiepileptic drugs such as levetiracetam and perampanel has preventive effects against DCI and delayed cerebral infarction on CT scan and MR imaging.

Patient Population
Between October 2013 and December 2021, 220 spontaneous SAH patients were transported to our hospital by ambulance or treated urgently.Among the patients, two patients underwent ventricular drainage and died without aneurysmal obliteration due to failed improvement from too severe neurological impairments, while no operation was performed in 69 patients because of devastating neurological conditions, the request of patient's family, or SAH of unknown etiology (Fig. 1).A prospectively maintained SAH database at our hospital was searched for patients who underwent clipping or coiling for ruptured cerebral aneurysms at days 0 (onset day of SAH) to 3 and had MR imaging at least twice, including within a few days after the aneurysmal obliteration and at days 14 to 21 post-SAH.During the study period, MR imaging to detect microinfarction was included in our treatment protocol: as a result, it was not implemented only in too severe or unstable cases, and therefore, 6 of 12 patients with inadequate MR imaging died prematurely due to severe neurological deficits.Medical records of 121 consecutive SAH patients (29 males and 92 females), 35 to 92 years of age (mean 65.4 years), were retrieved and retrospectively reviewed.The ethical committee of our institute approved this study and waived the need for informed consent.The inclusion criteria were ≥ 20 years of age at onset, SAH on CT scans, saccular aneurysm as the cause of SAH confirmed on digital subtraction angiography (DSA) or 3-dimensional CT angiography, aneurysmal obliteration within 72 h of onset, and MR imaging at days 1-4 (after aneurysmal obliteration) and days 14-21.Excluded from the study were SAH patients by dissecting, mycotic and arteriovenous malformation-related aneurysms, and other vascular anomalies or diseases.The study included 56 patients (46.3%) of admission World Federation of Neurological Surgeons (WFNS) grades IV-V [16], 55 (45.5%) patients of modified Fisher CT grade Fig. 1 A flow chart showing the included and excluded patients with aneurysmal subarachnoid hemorrhage (SAH) in this study 4 [17], and 103 patients (85.1%) of ruptured anterior circulation aneurysms including 35 patients (28.9%) of middle cerebral artery (MCA) aneurysms.The DCI-diagnostic and DCI-treatment protocol was unchanged during the study period, with the exception of the antiepileptic drug regimes.

Treatment Protocol
Aneurysmal clipping (33 patients, 27.3%) or coiling (88 patients, 72.7%) was performed as judged by the attending neurosurgeons to be appropriate for the individual patient.Ventricular, cisternal, and/or lumbar spinal drainage catheters were placed to manage acute hydrocephalus and persistent ventriculomegaly or to promote SAH clearance according to the preference of attending neurosurgeons as previously described [18].All patients were treated with intravenous injections of a Rho kinase inhibitor fasudil hydrochloride to prevent cerebral vasospasm for 14 days from the day after clipping or coiling.In addition, oral or enteral cilostazol, a selective inhibitor of phosphodiesterase type III and clinically available anti-platelet with pleiotropic actions [19], was administered to prevent DCI from 1-day post-clipping or post-coiling according to the 300 mg/ day cilostazol protocol, which was approved by the ethical committee of our hospital: a daily dosage was 200 mg (100 mg twice, every 12 h) between 1-day post-clipping or post-coiling and day 3, was increased to 300 mg (100 mg three times, every 8 h) between day 4 and days 10 to 28, and then returned to 200 mg (100 mg twice) [18].The cilostazol protocol was dosed down, modified, or discontinued at the discretion of the attending neurosurgeon, when uncontrolled headache, sinus tachycardia, heart failure, or pleural effusion occurred: such cases were recorded as cilostazol protocol violations.The administration of cilostazol was discontinued at the time of discharge in principle.Additional treatments were included to administer sufficient intravenous fluid volume and enough nutrition, to prevent meningitis, pneumonia, and hypoxia, and to correct anemia and hypoproteinemia.

Prophylactic Antiepileptic Drugs
The policy of using antiepileptic drugs differed depending on the study terms, which were divided into the following three terms.In the first treatment term (October 2013 to December 2015), antiepileptic drugs were not administered prophylactically and were used only after the onset of seizures.In the second treatment term (January 2016 to December 2019), in the light of a report that seizure burden was associated with poor outcomes after SAH [20], antiepileptic drugs were given prophylactically only in patients at high risks of seizures such as those with significant hematoma requiring surgical removal, decompressive craniectomy, and/or a ruptured MCA aneurysm [21].A new-generation antiepileptic drug levetiracetam was prophylactically administered in the second treatment term, because levetiracetam was most commonly used in Japan at that time.In the third term (January 2020 to December 2021), all patients received another new-generation antiepileptic drug perampanel prophylactically from the day of admission or the first day after aneurysmal obliteration.This was because our experimental study showed perampanel to be neuroprotective [13] and because perampanel monotherapy was approved in Japan at that time.
The prophylactic administration of antiepileptic drugs was discontinued at the time of discharge in principle, but in cases where the risk of epilepsy was considered to be high or when seizures occurred during hospitalization, the drug was continuously administered for 3 months or longer at the discretion of the attending neurosurgeons.

Outcome Measures
Primary outcome measures were incidences of delayed cerebral infarction on CT scans or MR imaging, and functional outcomes assessed by the modified Rankin Scale (mRS) at discharge and 3 months post-SAH.Delayed cerebral infarction was defined as a new infarction on either CT scans or diffusion-weighted MR imaging, which was unrelated to aneurysm obliteration and therefore did not appear within a few days after aneurysmal obliteration.All patients underwent CT scans at admission, on the day after aneurysmal obliteration, at least every 4-5 days until day 14, and then once a week until discharge.MR imaging and angiography were taken at least twice, including days 1-4 (after aneurysmal obliteration) and days 14-21.MRI imaging was acquired using two 1.5 T MR imaging scanners (Ingenia 1.5 T with dS head coil and Achieva 1.5 T with an 8-channel SENSE head coil; Philips Medical System, The Netherlands), and diffusion-weighted images were evaluated in an axial plane using 5-mm slice thickness in the first treatment term or 3-mm slice thickness in the second and third treatment terms.
Secondary outcome measures consisted of incidences of DCI, angiographic vasospasm, intraarterial infusions of fasudil hydrochloride (chemical angioplasty), percutaneous transluminal angioplasty (PTA), seizure, and chronic hydrocephalus.In addition, the incidence of reversible high-signal lesions in the splenium of the corpus callosum on diffusionweighted MR imaging was investigated, although the precise significance remains unknown [22].DCI was defined as any neurological deterioration presumed related to cerebral ischemia that persisted for ≥ 1 h independent of imaging findings, after rigorous exclusion of other potential causes of clinical deterioration [5].DCI was treated with hypertensive hypervolemic therapy, and chemical angioplasty and/or PTA were performed for angiographic vasospasm if indicated.All patients also underwent both CT perfusion and N-isopropyl-p-[(123) I]-iodoamphetamine single-photon emission CT around days 6-10, and in poor-grade or sedated patients, the DCI treatment was initiated based on the findings of regional cerebral blood flow reduction as well as angiographic vasospasm.Angiographic vasospasm was defined as ≥ 50% reduction in the baseline vessel diameter of cerebral arteries demonstrated by DSA, 3-dimensional CT angiography, or MR angiography, which were performed at least at days 6-8, days 14-21, and at the onset of clinical symptoms [18]: proximal vasospasm involved the internal carotid artery, the first segment (A1, M1, and P1) of anterior cerebral artery, MCA and posterior cerebral artery, or vertebrobasilar artery, while distal vasospasm developed only downstream of the first segments (M2, M3, A2, A3, P2, and P3).Electroencephalography was performed only as clinically indicated, and any seizures detected were included in the secondary outcome.Chronic hydrocephalus was diagnosed and treated with ventriculoperitoneal shunting (shuntdependent hydrocephalus) when a clinical deterioration with no detectable causes other than hydrocephalus occurred after day 14, and the ventricular size progressively increased (Evans index ≥ 0.30) [23].

Statistical Analysis
For statistical analyses, commercially available software (SPSS version 27, IBM Corp.) was used.Variables were recorded as categorical or continuous variables.Categorical variables were reported as percentages and analyzed using chi-square or Fisher exact tests, as appropriate.For 3-group comparisons of proportions, z-tests with Bonferroni-adjusted post hoc tests were used.Continuous variables were expressed as means ± standard deviation and/or medians (interquartile ranges).After confirming that each population being compared followed a normal distribution using Shapiro-Wilk W tests, statistical differences were analyzed using unpaired t-tests for 2-group comparisons or oneway analysis of variance (ANOVA) and the Tukey-Kramer multiple comparison procedure for 3-group comparisons.When data were not normally distributed, the data were compared using Mann-Whitney U tests between 2 groups or ANOVA and the Dunnet multiple comparison procedure among 3 groups.Pearson's or Spearman's rank correlation coefficient was used to determine a correlation between 2 variables related to 3-month poor outcomes (mRS 3-6).Any variable with a P value of < 0.05 on univariate analyses was entered into multivariate unconditional logistic regression models using the 3-month dichotomous mRS outcome (good or poor) as the dependent variable to find independent association with poor outcomes, although only the variable with the smallest P value was used as a candidate variable among similar clinical variables that were intercorrelated.Adjusted odds ratios with 95% confidence intervals were calculated, and independence of variables was tested using the likelihood ratio test on reduced models.A P value < 0.05 was considered significant.

Baseline Characteristics
Thirty-one, 59, and 31 patients were treated in the 1st, 2nd, and 3rd treatment terms, respectively (Fig. 1).The 3rd term group had a higher incidence of modified Fisher CT grade 1 compared with the 2nd term group (Online Resource: Table S1), but the highest incidence of ventricular drainage for acute hydrocephalus among the 3 treatment terms (Online Resource: Table S2).The 1st term group was less frequently treated with decompressive craniectomy to prevent or control the elevation of intracranial pressure compared with the other 2 term groups.
Antiepileptic drugs were administered in one patient (3.2%) in the 1st term, 17 patients (28.8%) in the 2nd term, and 31 patients (100%) in the 3rd term.One patient in the 1st term had seizure at day 2 and then took carbamazepine (100 mg twice enterally).In the second term, 13 patients received levetiracetam (500 mg twice enterally or orally) prophylactically, and 4 of them were given an additional dose of another antiepileptic drug enterally (carbamazepine, 100 mg twice, 1 patient; perampanel, 2 mg once, 1 patient; and lacosamide, 50 mg twice, 2 patients) after the onset of seizure: in addition, levetiracetam (500 mg twice enterally, 3 patients) or lacosamide (50 mg twice enterally, 1 patient) was administered after the onset of seizure, although no prophylactic antiepileptic drugs were administered.In the 3rd term, all of 31 patients received perampanel prophylactically, which was started with 2 mg once orally or enterally and increased to 4 mg 2 weeks later: 2 of them were given an additional dose of levetiracetam (500 mg twice enterally, 1 patient) or lacosamide (50 mg twice enterally, 1 patient) after the onset of seizure.

Comparisons of Outcome Measures Among the 3 Treatment Terms
As to the primary outcome measures, the incidence of delayed cerebral infarction on CT scans and outcomes at discharge or 3 months post-SAH were not different among the 3 terms.However, the incidence of delayed cerebral microinfarction on diffusion-weighted MR imaging, which was undetectable on CT scans, was the lowest in the 3rd term, although the number of the lesions was not different among the 3 terms (Table 1).The highest number of delayed cerebral microinfarctions on diffusion-weighted MR imaging was seen in a patient in the 2nd term who had no prophylactic administration of antiepileptic drugs and was associated with DCI (Fig. 2).
In the secondary outcome measures, the incidence of DCI in the 3rd term was significantly lower compared with the 2nd term on Fisher exact tests (P < 0.05), but this significance was lost after correction for multiple comparisons (Table 2).Other secondary outcome measures were similar among the 3 terms.Nonconvulsive seizures were diagnosed in one patient in the 1st term, 4 patients in the 2nd term, and 2 patients in the 3rd term.In 2 patients in the 3rd term, convulsions occurred after discharge from our hospital with discontinuation of antiepileptic drugs.Other seizures were observed during hospitalization.
When variables were compared between patients with and without delayed cerebral microinfarction on diffusion-weighted MR imaging, which was undetectable on CT scans, univariate analyses showed that no use of perampanel and the occurrence of DCI were significantly related to the development of delayed cerebral microinfarction on diffusion-weighted MR imaging (Online Resource: Table S3).

Effects of Antiepileptic on 3-Month Outcomes
Univariate analyses showed that patients with 3-month poor outcomes (mRS 3-6) were associated with older age, worse admission WFNS grade, worse modified Fisher CT grade, acute hydrocephalus, significant hematoma requiring surgical removal, cerebrospinal fluid drainage, cilostazol protocol violation, antiepileptic drug, reversible splenial lesions, delayed cerebral infarction on CT, seizure, and shuntdependent hydrocephalus (Online Resource: Table S4).

Table 1 Aneurysm occlusion modality and primary outcome measures in 3 treatment terms
The number of delayed cerebral microinfarctions on diffusion-weighted magnetic resonance imaging (DWI), which is undetectable on computed tomography (CT), is expressed as mean ± standard deviation and median (interquartile range) and is compared with one-way analysis of variance: no significant differences among the terms.Other data indicate the number of cases (% of total case per group) and are compared with z-tests with Bonferroni-adjusted post hoc tests: a P < 0.05 versus 1st and 2nd terms hydrocephalus with the smallest P value on univariate analyses, respectively, were used for subsequent multivariate analyses.Multivariate analyses revealed that independent factors related to 3-month poor outcomes were older age, admission WFNS grade V, acute hydrocephalus, significant hematoma requiring surgical removal, and cilostazol protocol violation (Table 3).The administration of antiepileptic drugs, levetiracetam and perampanel, had no significant relationships with the outcomes.Two patients had adverse effects of antiepileptic drugs: one patient had levetiracetam-induced myelosuppression, which improved with its discontinuation and symptomatic treatment, and another patient had visual hallucination at night due to perampanel, which disappeared after the discontinuation of administration.

Discussion
This study showed that the administration of new-generation antiepileptic drugs, levetiracetam and perampanel, was not associated with a higher rate of poor outcomes after SAH.Among independent prognostic factors, the only modifiable factor was the violation of cilostazol dose-up (300 mg/day) treatment, suggesting the treatment's efficacy [18].However, the prophylactic use of cilostazol as a replacement for nimodipine is neither routine practice in other parts of the world, nor recommended in the American Heart Association guideline [21], needing further studies.Although perampanel administration had no significant association with 3-month outcomes, it decreased CT-undetectable delayed cerebral microinfarction on diffusion-weighted MR imaging and appeared to suppress the development of DCI.Thus, perampanel is thought to be effective for post-SAH microcirculatory disturbance to some extent and is worthy of further investigation.DCI occurs via a complex interplay among several concurrent processes including cerebral (angiographic) vasospasm, cerebral microcirculatory, and neuroelectric disturbances including CSDs [6,9].It was reported that the development of late-onset epileptiform discharges might cause DCI by triggering CSDs, or to a lesser extent directly [24].CSD clusters are considered to induce inverted vasoconstrictive neurovascular coupling and then cerebral microcirculatory disturbances in compromised brain tissues after severe SAH, causing DCI [9,10].Although seizure and CSD are biologically distinct phenomena, they are interlinked and can trigger each other, worsening metabolic supply-demand mismatch and leading to the progression of DCI after SAH [9].Seizures and CSDs may also induce neuroinflammation, and vice versa, setting the stage for a viscous cycle among seizures, CSDs, and inflammatory reactions [9,25].Inflammatory reactions are important mediators of DCI development [26], and cross talk is known among neuroinflammation, arteriolar spasm, and microthrombi [6].Thus, although CSD itself may cause cortical infarction, the cross talk via seizures and neuroinflammation may induce microinfarction in the subcortical region and basal ganglia as observed in Fig. 2, possibly by microvasospasm and microthrombi [6,9].As seizures produce a condition of metabolic derangements and localized electrolyte imbalance, they may contribute to the development of cerebral vasospasm as well as microvasospasm in a similar manner as CSDs [6,27].A study showed the development of pathologically low-frequency activities (worsening of focal slowing) on continuous electroencephalographic monitoring a few days before the documentation of angiographic vasospasm [28], and another study reported that angiographic vasospasm was associated with seizures from onset of SAH to hospital discharge [27].
According to a literature review of aneurysmal SAH, simple partial or tonic-clonic seizures with or without secondary generalization are the most commonly reported seizure type, and the incidence of clinically evident seizures is 3-26% at onset, 1-28% within the first 2 weeks, and 1-35% after 2 weeks of SAH [29][30][31].In addition, nonconvulsive seizures or status epilepticus are reported in 3-18% of SAH patients admitted to neurological intensive care unit, especially in patients with poor neurological status or neurological deterioration [29,32,33].Seizures may be triggered by neuronal damage caused by the blood itself and secondary brain injuries [29], and the wide range in frequency of reported seizures can be explained by differences in the patient selection and the administration of prophylactic antiepileptic drugs: worse neurological grades, larger amounts of intracranial blood, the presence of cerebral edema, and a ruptured MCA aneurysm are generally associated with more frequent seizures [29,31,34].With increased use of continuous electroencephalographic monitoring, there may be higher detection rates of electrographic and nonconvulsive seizures or status epilepticus, leading to more treatment with antiepileptic drugs.However, it remains unknown if antiepileptic treatment can reduce the incidence or severity of seizures, CSDs, cerebral vasospasm, or DCI after SAH.
Occurrence of post-SAH seizure itself or longer or more frequent seizures has been suggested to lead to worse outcomes [20], although such a correlation may merely reflect the severity of SAH.Currently, there is insufficient evidence to support the routine use of antiepileptic drugs for the primary or secondary prevention of seizures after SAH [29,35].However, both the Neurocritical Care Society and the American Heart Association guidelines weakly recommend the use of antiepileptic drugs in the immediate post-SAH period (3-7 days from ictus), although routine long-term use of antiepileptic drugs is not recommended by either guideline [21,36].According to a recent survey in North America, the vast majority of practitioners prophylactically used antiepileptic drugs for all SAH patients [35].Interestingly, most of those who responded to the survey did not support a randomized controlled trial of antiepileptic drugs for SAH patients [35], probably reflecting a higher degree of concern about the perceived risks of seizures including aneurysmal rebleeding and secondary brain injuries [37].On the other hand, a number of retrospective studies have shown that prophylactic usage of antiepileptic drugs (mostly phenytoin) does not improve seizure outcomes, but is associated with worse functional outcomes and higher rates of complications including cerebral vasospasm, DCI, delayed cerebral infarction, cognitive dysfunction, and drug-induced fever [14,29,38,39].There are also some evidences from animal and human studies that the administration of certain antiepileptic drugs (phenytoin, valproate, carbamazepine, and phenobarbital) may lower the chance of a good functional recovery by the antiepileptic treatment itself, particularly with phenytoin, rather than seizure activity [29,38,39].Considering higher incidences of side effects (especially valproate) and potential drug-drug interactions (especially phenytoin and carbamazepine with a marked reduction in plasma concentrations of an anti-DCI drug nimodipine approved by the Food and Drug Administration), the use of these older antiepileptic drugs is not recommended in SAH patients [39].Thus, newer antiepileptic drugs such as levetiracetam and lacosamide with fewer side effects and less drug-drug interactions are more frequently used recently, but higher initial dosing of these renal excreted drugs may be needed to maintain therapeutic levels and remain seizure-free due to augmented renal clearance in SAH patients: reportedly a total levetiracetam daily dose of 2000 − 4000 mg might be required, while lacosamide has no information about the required dose or its efficacy in SAH patients [15,39].A recent study demonstrated no significant association between levetiracetam administration (unknown dose and duration) and outcomes after SAH [37].According to a single-center, randomized trial of a brief (3-day) course of levetiracetam (1000 mg, twice a day) versus extended treatment (until hospital discharge), extended levetiracetam treatment had more drug-related adverse events and was an independent determinant for poor outcomes [34].However, effects of levetiracetam on delayed cerebral microinfarction have never been investigated after SAH.
Clinical effects of perampanel have never been investigated in aneurysmal SAH patients as far as we know.However, a recent study reported that α-amino-3-hydroxy-5-methyl-4-isoxazole propionate (AMPA) receptors were activated in brain capillary endothelial cells and neurons after experimental SAH and that a selective AMPA receptor antagonist perampanel suppressed post-SAH occurrence of blood-brain barrier disruption as well as epileptic spikes, resulting in improved neurological function in mice [13].In a phase 2, randomized, double-blind, placebo-controlled trial, an inhibitor of postsynaptic density-95 protein NA-1, which blocks activation of both AMPA and N-methyl-Daspartate (NMDA) receptors, significantly decreased the occurrence of microembolic infarction associated with endovascular coiling of unruptured and ruptured cerebral aneurysms and improved 30-day outcomes in patients with less severe SAH [40].Thus, taken together with the findings in the present study, an AMPA receptor antagonist perampanel is expected to exert neuroprotective effects in aneurysmal SAH patients.Newer antiepileptic drugs generally have fewer and less severe side effects, and there would be no major issue if the prophylactic administration is discontinued at discharge; however, when the prophylaxis is given to all SAH patients, it would be necessary to carefully evaluate the risk of psycho-cognitive side effects.Further studies will determine the optimal dosing protocol, including dosage and duration, of the new-generation antiepileptic drug perampanel.
This study has some limitations.First, this is a relatively small-scale retrospective study with potential selection bias.In particular, levetiracetam was used only in patients at high risk of seizures, whereas perampanel was used in all patients.Therefore, it cannot be denied that levetiracetam may have had similar effects to perampanel if administered to all patients.In addition, a thrombo-embolic origin of microinfarction cannot be fully excluded, although MR imaging was taken 1 to 3 days after aneurysmal obliteration in all patients, and the opportunities of DSA and chemical angioplasty were comparable among the groups in this study.Second, nimodipine has the best evidence for the treatment of DCI [21] and may inhibit CSDs [41,42].However, nimodipine was not used in this study, because the drug was not approved in Japan.Thus, the combined effects of the new-generation antiepileptic drug and nimodipine remain unknown.Third, CSDs and electroencephalographic finding were not evaluated, and the direct effects of new-generation antiepileptic drugs on them remain unknown in this study.Fourth, cognitive function was not assessed in this study, potentially having contributed to the lack of significant differences in outcome.New-generation antiepileptic drugs, especially perampanel, have not been systematically studied in trials, and therefore, further studies including large-scale prospective studies are needed to assess its risks and benefits for prophylactic use after SAH, including the effects on epileptiform abnormalities on electroencephalogram, seizures, DCI, and functional outcomes.

Conclusions
This study showed that prophylactic administration of newgeneration antiepileptic drugs, especially perampanel, may have a beneficial effect against post-SAH delayed microcirculatory disturbances in terms of decreased incidences of CT-undetectable but diffusion-weighted MR imagingdetected microinfarction and DCI.New-generation antiepileptic drugs, especially perampanel, may be a promising new therapeutic strategy for preventing and treating neuroelectric disturbances, one of the causes of DCI, and deserve further studies.

Table 3
Multivariate logistic regression with 3-month poor outcome (modified Rankin Scale 3-6) as a binary end pointAge is used as a continuous variable.References are as follows: admission World Federation of Neurological Surgeons (WFNS) grade I for admission WFNS grade, and no antiepileptic drug for levetiracetam and perampanel.CT computed tomography