Convection-Enhanced Delivery of Muscimol in Patients with Drug-Resistant Epilepsy

Abstract

BACKGROUND

Minimally invasive therapies for drug-resistant epilepsy (DRE) have been advocated. A study of convection-enhanced delivery (CED) of muscimol, a GABA_A receptor agonist, was previously completed in non-human primates.

OBJECTIVE

To investigate the safety and anti-epileptic effects of intracerebral muscimol infusion into the epileptic focus of patients with DRE.

METHODS

In this phase 1 clinical trial, 3 adult patients with DRE underwent CED into the seizure focus of artificial CSF vehicle followed by muscimol for 12 to 24 h each using a crossover design. Basic pathophysiology of the epileptic focus was examined by assessing the infusions’ effects on seizure frequency, electroencephalogram (EEG) spike-wave activity, and power-spectral EEG frequency.

RESULTS

Inter-ictal neurological function remained normal in all patients. Pathological examination of resected specimens showed no infusion-related brain injuries. Seizure frequency decreased in 1 of 3 patients during muscimol infusion but was unchanged in all patients during vehicle infusion. Mean beta frequencies did not differ significantly before, during, or after infusion periods. Infused fluid provided insufficient MRI-signal to track infusate distribution. In the 2 yr after standard epilepsy surgery, 1 patient had temporary reduction in seizure frequency and 2 patients were seizure-free.

CONCLUSION

CED of muscimol into the epileptic focus of patients with DRE did not damage adjacent brain parenchyma or adversely affect seizure surgery outcome. This study did not confirm that intracerebral muscimol infusion effectively suppressed seizures. A surrogate tracer is recommended to track infusion distribution to the epileptic focus and surrounding structures in future studies using CED to suppress the seizure focus.

Drug-resistant epilepsy (DRE) is a common condition with profound health and socioeconomic consequences for those afflicted and their families. Approximately 15 000 to 30 000 people in the US develop DRE annually.¹ Epilepsy is considered “drug-resistant” when patients have not become seizure-free after adequate trials of 2 appropriate antiepileptic drugs (AEDs) that do not cause unacceptable side effects. DRE significantly reduces quality of life, has neuropsychological and behavioral co-morbidities, and increases mortality.¹

Epilepsy surgery is an effective treatment for selected patients. However, when the seizure focus lies within or near areas of eloquent cortical functions, such as speech, movement, or sensation, the risk of functional loss from resection of the epileptic focus may exceed the reward. Recovery from neurologic deficits from removing eloquent cortex is rarely complete and depends on brain plasticity.² Even when surgical resection avoids eloquent cortex, quadrantanopia is common with temporal lobectomy, and infrequently, injury to the middle cerebral, anterior choroidal, or posterior cerebral arteries leads to stroke and hemiparesis, aphasia, and hemianopsia.³ In patients with DRE, a procedure that could suppress or eliminate neuronal excitability of the epileptic focus and preserve surrounding cortex, white matter, and vasculature would be less invasive and potentially safer than open craniotomy.

Epilepsy arises from extremely rapid and synchronized firing of neurons within the brain.⁴ Pharmacological suppression of these excitable neurons is the primary treatment of epilepsy. Oral AED dosing produces a uniform AED level throughout normal and epileptic brain regions, suppressing neurons in non-epileptic brain structures, with potential side effects. Surgical candidates with DRE have an epileptic focus in a single brain region and resistant to suppression with systemic pharmacological treatment. Convection-enhanced delivery (CED) is a method to infuse pharmacological agents into the extracellular (interstitial) space of the brain.^5-11 Delivering an AED directly into the extracellular space of the epileptic focus could suppress its neuronal activity and spare non-infused brain regions. Such a strategy of regional pharmacological therapy in DRE patients has not been attempted.

Recent approaches to DRE indicate that physicians and patients desire therapies like stereotactic laser ablation that are less invasive, safer, and more selective than present therapies.¹² To provide a basis in humans for using regional pharmacotherapy to treat DRE, we performed a pilot study of prolonged micro-infusion of muscimol into the epileptic focus. Muscimol was chosen because short muscimol infusions were delivered safely to the brain of humans with movement disorders and muscimol has a reversible GABA-receptor-mediated effect that inhibits neuronal activity and could suppress the epileptic focus.^13,14

METHODS

Study Design

This phase 1 clinical study introduced the investigational new drug, muscimol, into the human brain by CED. The study was designed to obtain information about muscimol's pharmacological actions, side effects associated with increasing doses, and early evidence of effectiveness, and support the design of well-controlled, scientifically valid, phase 2 studies.¹⁵ The study design after seizure localization is shown in Figure 1. The protocol was approved by the institutional review board and adheres to CONSORT guidelines. All participants gave informed consent.

FIGURE 1.

Study design flow chart after seizure focus localization. In total, 3 patients were enrolled in the study. MST = multiple subpial transections. DEC = depth electrode-catheter.

Participant Recruitment

Adult patients who were candidates for standard surgical care of DRE were eligible for this trial. Candidates were 18 yr of age or older, had partial or complex partial seizures (CPS) for more than 2 mo despite medical therapy, and had a seizure focus in a single region of 1 cerebral hemisphere. Recruitment was through self- and physician-referral. Three patients participated.

Screening

Patients were evaluated by institutional neurology staff and received standard testing as inpatients or outpatients, including: (1) history, physical, and neurologic examination; (2) chest X-ray; (3) EKG; (4) blood tests (CBC, platelet count, and differential; PT, PTT; acute care panel; AED levels); (5) urine for urinalysis, culture, and pregnancy test; and (6) head MRI with and without contrast. Patients then underwent continuous video-EEG monitoring with scalp electrodes as inpatients. AEDs were slowly withdrawn to increase the rate of seizures. At least 3 ictal events were captured to localize the seizure focus.

Patient Population

Patient 1 was a 44-yr old right-handed woman with a 10-yr history of left index finger twitching. Over time, involuntary twitching episodes lasted longer, had shorter rests between episodes (5-10 min), and spread throughout the left hand, producing epilepsia partialis continua. The left upper extremity became weak (4/5 strength) proximally and distally, hyper-reflexic, and the left hand clumsy. She took carbamazepine extended-release capsules, 400 mg 3 times daily, and clonazepam 1 mg every morning and 0.5 mg 3 to 4 times daily as needed without much relief. Previous treatment with divalproex, phenytoin, carbamazepine, gabapentin, levetiracetam, lamotrigine, oxcarbazepine, topiramate, and intravenous immunoglobulin and plasmapheresis for possible Rasmussen's encephalitis, failed to relieve her seizures.

Patient 2 was a 26-yr old right-handed man whose seizures began at age 3. He had a history of febrile seizures. He had 3 to 4 seizures per month, each presenting with headache, followed by dizziness, staring, lip and tongue automatisms, and secondary generalization. He took carbamazepine extended-release capsules, 300 mg twice daily. Phenobarbital and other AEDs were ineffective.

Patient 3 was a 34-yr old right-handed man with an 11-yr history of complex-partial and generalized tonic-clonic (GTC) seizures. By study enrollment, CPSs occurred daily and GTC seizures every 6 wk, despite treatment with levetiracetam and gabapentin. Previous treatment with divalproex, phenytoin, and topiramate was ineffective. He had chronic right shoulder pain from recurrent shoulder dislocation during seizure activity.

MRI Examinations

All patients underwent brain MRI pre- and post-intracerebral depth electrode-catheter (DEC) placement. T2-weighted and FLAIR MRI of the brain was performed during vehicle and muscimol infusions to detect infusate in the brain around the catheter.¹⁶

Epileptic Focus Localization

After hospital admission, Patient 1 underwent placement of a right frontoparietal subdural electrode grid and then underwent 2 consecutive days of video-EEG recording, which identified the seizure focus and localized it to the right precentral gyrus (Figure 2). The patient entered the study and the following day returned to the operating room for DEC placement into the seizure focus (right precentral gyrus). Ictal and interictal activity was recorded for 5 d before starting sequential 12-h infusions of vehicle and muscimol solutions. After the infusions, patient 1 returned to the operating room for subdural grid and DEC removal and multiple subpial transections (MST) under local anesthesia with awake monitoring of left-sided motor function.

FIGURE 2.

Patient 1–Seizure focus. MRI showed the focus of right UE epilepsia partialis continua in the cortex and revealed right precentral gyrus atrophy and decreased signal on T2 (white arrow).

Before entering the trial, Patient 2 had increased signal intensity in the left hippocampus on FLAIR MRI consistent with mesial temporal sclerosis (Figures 3A and 3B). The left temporal horn was slightly larger than the right. Patient 2’s video-EEG identified 7 complex-partial seizures, all typical of the patient's seizures, localized to the left anterior and mid-temporal regions. The left hippocampus had 13% lower uptake of fluorodeoxyglucose than the right on ¹⁸F-FDG PET. These findings together diagnosed left mesial temporal sclerosis without requiring invasive monitoring.¹⁷ The patient entered the trial and the DEC was placed stereotactically into the seizure focus (left hippocampus) solely for the research purposes of the trial. Ictal and interictal activity was recorded for 5 d before starting sequential 24-h infusions of vehicle and muscimol solutions. The next day, the patient underwent left anterior temporal lobectomy, amygdalohippocampectomy, and DEC removal. The patient tolerated the infusions and procedures well.

FIGURE 3.

Patient 2–Seizure focus. Coronal A and axial B FLAIR MRI revealed left mesial temporal sclerosis (MTS; white arrows).

Before admission to our institution, Patient 3 had had a comprehensive workup which localized a seizure focus to the left anterior temporal brain region. However, MRI did not show definite signs of mesial temporal sclerosis. The patient entered the trial and underwent left frontotemporal craniotomy for placement of left temporal subdural grids and strips for surface-electrode recording and a left hippocampal DEC for depth-electrode recording. After withdrawal of AEDs, 5 CPS with secondary generalization occurred during a 5-d recording period, originating near medial temporal surface and hippocampal depth electrodes, consistent with mesial temporal sclerosis. He then underwent vehicle and muscimol infusions over 2 consecutive days. One day after completion of the infusions, Patient 3 returned to the operating room for left anterior temporal lobectomy, amygdalohippocampectomy, and DEC removal, which were well-tolerated.

Evolution of representative seizure EEG-recordings for Patients 1, 2, and 3 are shown in Figure 4A-4C, respectively.

FIGURE 4.

Evolution of ictal events during pre-infusion recording period. EEG recordings showing seizure evolution in Patient #1 A, Patient #2 B, and Patient #3 C. Showing the EEG recording for the entirety of the seizure was not practical. The double black lines represent a break in the EEG recording in order to show both the beginning and end of the ictal event.

Depth Electrode-Catheter Placement

The DEC (AdTech, Oak Creek, Wisconsin) was a modified Silastic depth electrode whose central stylet was replaced by an inner cannula for infusion. The infusion cannula consisted of polyimide-coated synthetic fused silica. Its tip was placed in each patient's seizure focus (Figure 5A).^17,18 The outer diameter of the Silastic part was 2 mm and of the polyimide part was 1.1 mm. The length of the DEC was customized to the expected depth of insertion. The DEC was included by the FDA in the study IND. The Stealth frameless stereotactic system (Medtronic, Dublin, Ireland) guided placement of the DEC. Framelink software (Medtronic) and the Leksell (Elekta AB, Stockholm, Sweden) stereotactic system were used in Patients 2 and 3 who underwent hippocampal DEC placement. The infusion part of the DEC cannula passed through the center of the scalp incision with the infusion port resting immediately above the scalp incision. The electrode leads were tunneled through the adjacent scalp. The DEC was capped with the introducing stylet in place during the recording period before infusion. The cap and stylet (Figure 5A) were replaced by the infusion cannula (Figure 5A) for the infusions. Dressings and topical antibiotic cream were applied to the DEC scalp exit site to prevent infection and trauma to the DEC and surrounding scalp wound. Postoperative MRI confirmed DEC placement in Patient 1’s epileptic focus and intraoperative MRI confirmed intrahippocampal DEC placement in Patients 2 and 3. The DEC obtained baseline depth electrode recordings for 5 d before infusion initiation.

FIGURE 5.

Depth-electrode catheter. A, Depth electrode-catheter (attached to green cap) and stylet (attached to red cap) used for infusion in all 3 patients. B, Surgical placement of the depth electrode-catheter. C, Post-op T1-weighted MRI in the parasagittal plane showing depth electrode-catheter placement (white arrow).

Intracranial EEG Recording and Beta Frequency Averaging

Intracranial DECs anatomically targeted the epileptic focus in each patient, based on pre-surgical epilepsy planning and clinical profiles described above. Intracranial DEC data were continuously recorded using a sampling frequency of at least 500 Hz and bandpass filters ranging from 1 to 500 Hz (TWin Software; Grass Technologies, West Warwick, Rhode Island). Intracranial EEG data using monopolar and bipolar montages at varying filter and gain settings were reviewed by 2 independent epileptologists: one for clinical purposes and the other for post-acquisition data analysis. To determine average beta frequency (normal range: 13-30 Hz)¹⁹ and if it were affected by muscimol, DEC recordings were sampled at least 30 random times in each patient in each period before, during, and after muscimol infusion and analyzed using Grass Technologies Frequency software.

Convective Infusion of Artificial CSF Vehicle and Muscimol

All patients underwent convective muscimol infusion into their epileptic focus (cortex in Patient 1; hippocampus in Patients 2 and 3). Muscimol and vehicle solutions were continuously infused at 1.4 μL/min (84 μL/h) based on calculations that this flow rate would preferentially target the gray matter more than surrounding white matter tracts. All patients received artificial CSF vehicle followed by muscimol solution (Patients 1 and 2: 7.1 ng/μL, 0.0625 mM; Patient 3: 14.2 ng/μL, 0.125 mM). Patient 1 had 12-h infusions of vehicle and of muscimol. Patients 2 and 3 had 24-h infusions of vehicle and of muscimol.

The muscimol infusate concentration was much less than the 1 μg/μL concentration previously delivered by bolus injection into the globus pallidus interna and thalamus of humans to inhibit these structures during movement disorder surgery.^14,20 Infusion models indicated that steady-state concentrations within the infused brain would be achieved within 800 min and that muscimol concentration would fall to 0.0125 mM (12.5 μM; 10% of the infused concentration) at a distance (r) of 0.4 to 0.5 cm from the infusion point.^6,10 The total daily muscimol dose administered (7.2-28.6 μg) was well below the oral dose of 5-10 mg/d previously tolerated without systemic toxicity in another study.¹³

Surgical & Post-surgical Care

Operative treatment occurred 1 to 3 d after the infusions, Patient 1 underwent MST to the epileptic focus in the motor cortex. Patients 2 and 3 underwent left anterior temporal lobectomy and amygdalohippocampectomy to remove the epileptic focus. Surgical specimens were obtained from Patients 2 and 3, but not Patient 1, and examined for infusion-related histological changes. Following surgery, patients were monitored in the Surgical Intensive Care Unit (1-2 d) and then in the department's nursing unit (4-8 d) before discharge.

Outcomes Measures

The primary outcome measure was seizure frequency. Secondary outcome measures included EEG spike-wave activity (spikes/minute) and power-spectral EEG frequency (Hz).

Blinding

The study team was blinded to the infusate identity (artificial CSF vehicle vs muscimol). Individuals reviewing video-EEG and performing neurological examinations were also blinded to the infusate identity.

Statistical Analysis

Statistical analysis was performed with commercial software (JMP 13.0, SAS Institute, Cary, North Carolina).

RESULTS

Convective Infusion of Artificial CSF Vehicle and Muscimol

Patients 1 and 2 tolerated the infusion of muscimol at a concentration of 7.1 ng/μL (0.0625 mM) and Patient 3 tolerated the next highest concentration of muscimol, 14.2 ng/μL (0.125 mM), well.

Imaging

In Patient 1, MRI demonstrated cortical laminar sclerosis in the right precentral gyrus as the DRE-evocative lesion (Figure 2) and confirmed accurate DEC placement (Figure 5B and 5C). FLAIR signal did not increase during infusion of vehicle or muscimol.

In Patient 2, MRI showed unilateral MTS (Figure 3A and 3B). The Leksell stereotactic system accurately placed the DEC within the hippocampus (Figure 6A and 6B), as confirmed by MRI (Figure 6C-6F). FLAIR MRI detected increased signal around the DEC after DEC placement, during monitoring (Figure 6G), during infusion period (Figure 6H and 6I), and after infusion (Figure 6J). FLAIR signal did not increase during infusion of vehicle (Figure 6H), muscimol (Figure 6I), or afterwards (Figure 6J).

FIGURE 6.

Depth electrode-catheter–Surgical setup and placement. A, Patient in the prone position with the Leksell frame and coordinator indicator box in place and the frame secured to the iMRI table. B, Surface coils were placed on the sides of the head. C, Occipital approach for placement of the depth electrode-catheter. D, MP-RAGE image in the left parasagittal plane confirmed that the depth electrode catheter (white arrow heads) passed inferior to the lateral ventricle (black arrow). MP-RAGE image in the axial E and coronal F planes confirmed placement of the tip of the electrode-catheter in the hippocampus (white arrow) medial to the temporal horn. MRI after depth electrode-catheter placement and during muscimol infusions G–J: G, FLAIR MRI 1 d post-op, H, during vehicle infusion, I, during muscimol infusion, and J, 24 h post-muscimol infusion.

In Patient 3, the diagnosis of MTS was made by video-EEG and electrocorticography since MRI was non-diagnostic. Hippocampal DEC placement was confirmed by MRI (Figure 7A-7C). MRI did not detect the presence of infused fluid during vehicle (Figure 7D) or muscimol infusion (Figure 7E).

FIGURE 7.

Patient 3–Pre-, intra-, and post-infusion MRI. Occipital approach for placement of the depth electrode-catheter A and B: Intraoperative MP-RAGE imaging in the parasagittal A and axial B planes confirmed placement of the depth electrode-catheter in the hippocampus. MRI after depth electrode-catheter placement and during vehicle and muscimol infusion C–E: T2-weighted MRI C, 1-d post-op, D, during vehicle infusion, E, and during muscimol infusion.

Seizure Frequency, EEG Spike-Wave Activity, and Power-Spectral EEG Frequency

In Patient 1, seizures averaged 6 (range 0-21) per day before infusion. The DEC recorded persistent spike and wave epileptiform discharges every 1 to 3 s throughout this period (Figure 8A-8C). Four seizures occurred during the 12-h artificial CSF vehicle infusion. No seizures occurred during the 12-h muscimol infusion. Epileptiform discharges were not reduced during muscimol infusion. The patient had 13 seizures during the 2.5-d post-infusion recording period. Seizure duration before infusion averaged 189 s (range 19-473 s), during vehicle infusion 124 s (range 39-337 s), and post-muscimol infusion 45 s (range 7-103 s). Upon power-spectral EEG analysis, mean beta frequencies during pre-, intra-, or post-infusion periods did not vary significantly (Table 1). Patient 1 had much less frequent seizures for several months after subpial transection before seizure activity returned to pre-operative levels later in the year (Engel Classification: 4B).²¹

FIGURE 8.

Patient #1: EEG before, during, and after muscimol infusion. Sample depth-electrode recording of inter-ictal discharges in the periods before A, during B, and after C muscimol infusion. Patient #1’s epileptiform discharges did not differ significantly before, during, or after infusion of muscimol.

TABLE 1.

EEG Mean Beta Frequencies (Hz)

Mean Beta Frequency (Hz)
	Pre-infusion	Vehicle Infusion	Muscimol Infusion	Post-infusion
Patient 1	17.2 (± 1.05)	18.4 (± 1.35)	17.8 (± 1.54)	17.5 (± 1.37)
Patient 2	16.0 (± 1.49)	15.8 (± 0.97)	16.3 (± 2.11)	16.8 (± 1.72)
Patient 3	19.1 (± 1.69)	17.7 (± 1.16)	16.6 (± 1.25)	17.0 (± 1.19)

Values = mean (± standard deviation).

Patients 2 and 3 had infrequent epileptiform activity without appreciable difference in discharge frequency between the pre-, intra-, and post-infusion periods (Patient 2: Figure 9A-9C, Patient 3: Figure 10A-10C). Upon power-spectral EEG analysis, mean beta frequencies during pre-, intra-, or post-infusion periods did not vary significantly (Table 1). Patient 2 did not have seizures during vehicle or muscimol infusion. Patient 3 had 2 seizures during vehicle infusion and 2 seizures during muscimol infusion. Patients 2 and 3 were seizure-free during the 2-yr follow-up period after left anterior temporal lobectomy and amygdalohippocampectomy (Engel Classification: 1A).²¹ Seizure frequency quantification for all patients is shown in Table 2.

FIGURE 9.

Patient #2: EEG before, during, and after muscimol infusion. Sample depth-electrode recording of inter-ictal discharges in the periods before A, during B, and after C muscimol infusion. Patient #2’s epileptiform discharges did not differ significantly before, during, or after infusion of muscimol.

FIGURE 10.

Patient #3: EEG before, during, and after muscimol infusion. Sample depth-electrode recording of ictal and inter-ictal discharges in the periods before A, during B, and after C muscimol infusion. The sample recordings before A and during B muscimol infusion show the development of a seizure. Patient #3 did not experience any seizures after muscimol infusion C. Antiepileptic medications were tapered off before recordings began and restarted after the infusions were complete.

TABLE 2.

Seizure Frequencies

Seizures During Each Recording Period
	Pre-infusion*	Vehicle Infusion	Muscimol Infusion	Post-infusion*
Patient 1	6.29 (± 6.58)	4	0	4.33 (± 3.06)
Patient 2	0.4 (± 0.8)	0	0	0 (± 0)
Patient 3	0.3 (± 0.64)	2	2	0 (± 0)

*Mean number of seizures per 24 h. Values = mean (± standard deviation).

Neurological Examination During and After Infusion & Pathological Studies

Interictal neurological function remained stable in Patient 1 and normal in Patients 2 and 3 during and after the infusions. Patient 1 had no tissue specimen because her seizure focus was in the hand area of the motor cortex.

Pathological examination of left temporal lobe neocortex from Patient 2 revealed minimal perivascular lymphocytes. The hippocampus showed evidence of hippocampal sclerosis and localized parenchymal loss associated with histiocytes and perivascular lymphocytes, consistent with a catheter tract. No infusion-related brain injuries were noted.

Examination of the left temporal lobe of Patient 3 revealed minimal perivascular lymphocytes, likely associated with prior subdural grid placement. The hippocampal specimen showed leptomeningeal inflammation and perivascular lymphocytes. The needle tract or gliosis was not observed. No infusion-related brain injuries were found.

DISCUSSION

This first-in-human phase 1 clinical trial delivered muscimol to the epileptic focus using CED. This trial demonstrated safety of muscimol infusion for up to 24 h in patients with DRE. Only Patient 1 experienced reduced seizure frequency during muscimol infusion compared to vehicle infusion. Patient 2 did not experience a seizure during either vehicle or muscimol infusion. Patient 3 had 2 seizures during both vehicle and muscimol infusion.

In this study, we used CED to deliver muscimol into the brain where it could interact with GABA_A receptors on epileptic focus neurons. Muscimol (3-hydroxy-5-aminomethylisoxazole) is a GABA_A receptor agonist that has been used to study the role of GABA within animal and human brains. GABA is the primary inhibitory neurotransmitter in the brain, present in 60% to 70% of all brain synapses. In the brain, muscimol is degraded much more slowly than GABA, producing a more sustained effect than GABA itself. Orally administered muscimol is rapidly metabolized in the periphery, preventing intact muscimol from reaching the brain, limiting its effect to its metabolites.^22,23

Unilateral muscimol delivery did not produce cognitive side effects nor result in inferior clinical outcomes from subsequent standard surgical procedures for DRE. For this trial, optimal safety and effectiveness of epileptic focus suppression with muscimol depended on limiting muscimol's effect to the epileptic focus. We were unable to track the muscimol infusate, which had been possible in a previous animal study.¹⁶ In the non-human primate pre-clinical trial of muscimol infusion into the hippocampus, infused fluid increased T2 and FLAIR signal around the DEC catheter tip, but in this human study increased T2 and FLAIR signal did not show the infusate.¹⁶ In future trials, addition of a surrogate tracer to the muscimol solution would allow MRI-tracking of muscimol distribution to the epileptic focus and surrounding brain parenchyma, which could prompt modifications to the infusion rate, duration, and site to better target epileptic tissue, maximize treatment effects, and reduce side effects on surrounding tissue.²⁴ Extending the polyimide catheter tip further beyond the Silastic part of the DEC could reduce leak-back of muscimol solution and increase intra-parenchymal delivery around the infusion tip.²⁵ To avoid the upper part of the DEC passing through a scalp incision, the DEC design could be modified to allow both the extracranial infusion tubing and electrode lead component to exit the scalp through separate needle holes outside the scalp incision. Despite its imperfect design, there were no infections related to the DEC, presumably because of the relatively short recording periods and local wound care.

The risk of depth electrode placement within the hippocampus is inherent in the standard evaluation for epilepsy surgery of patients in whom the epileptic focus cannot be localized non-invasively. Placement of a DEC with a central lumen for infusion carried similar risks to a depth electrode because the non-metallic outer portion of the DEC was the same biocompatible plastic (Silastic) as the depth electrode.¹⁰ The perfusion portion of the DEC was also biocompatible. More FLAIR changes were seen after DEC placement than expected after standard depth-electrode placement, presumably because placing the DEC and fixating it with methyl methacrylate involved more catheter manipulation than placing a standard depth electrode and fixating it with an anchor bolt (AdTech, Oak Creek, Wisconsin).

The DEC tip entered the most inferior part of the hippocampus in Patient 3 placing its most anterior electrode on the inferior margin of the hippocampus (Figure 7A and 7B). Its eccentric placement in the hippocampus may have reduced recording of epileptiform activity. The DEC electrodes did record seizure activity during the monitoring period before infusion. The recording electrodes were at least as close to the hippocampus as depth electrodes placed in the temporal horn of the lateral ventricle, which was shown to localize hippocampal seizure foci.²⁶

In 4 patients with focal DRE, During and Spencer implanted bilateral depth electrodes containing micro-dialysis catheters into the hippocampus and amygdala. They studied safety and tested for 10 to 16 d.¹⁷ As in our study, their electrode-catheters did not produce significant adverse effects nor complicate subsequent surgical procedures or clinical management.

The apparent safety of CED in our study supports future clinical trials using CED to deliver agents into the epileptic focus that selectively modulate or destroy seizure focus neurons while sparing surrounding normal tissue.

CONCLUSION

To investigate the safety and possible effectiveness of intracerebral muscimol infusion in patients with DRE, we performed a Phase 1 clinical trial, the first of its kind in this area of study. Muscimol was safely infused into the brain by CED in 3 patients with DRE. Seizures were reduced during muscimol infusion in 1 of 3 patients. This study did not confirm that intracerebral muscimol infusion was effective in suppressing seizures. Infusate tracking by MRI was unsuccessful and we recommend using a surrogate tracer in future trials. A clinical trial with larger patient enrollment would be necessary to establish the maximal tolerated dose and infusion duration of muscimol administered by CED into the epileptic focus of patients with DRE.

Disclosures

This research was supported by the Intramural Research Program of the National Institute of Neurological Disorders and Stroke at the National Institutes of Health. The authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.

Notes

Preliminary data contained in this study was presented at the 2018 Annual Scientific Meeting of the American Association of Neurological Surgeons in New Orleans, Louisiana, from April 28-May 2.

COMMENT

This manuscript reports the first-in-human, Phase 1 study of prolonged CED of muscimol (a GABA agonist drug) into epileptic foci. The investigators obtained initial safety data, but without clear effectiveness, as intended by a phase 1 trial. The authors should be commended for their work in the neuromodulatory subfield of epilepsy surgery. The greatest challenge, as highlighted, will be the demonstration of efficacy in this very challenging group of patients with medically intractable epilepsy. A larger study to compare adequate dose and maximal efficacy will likely be the next step, and I encourage the authors to continue this work.

Jorge Gonzalez-Martinez

Cleveland, Ohio