Abstract
Background and Objectives
Neuromodulation therapies are approved for the treatment of focal epilepsy based on data from randomized controlled trials (RCTs). After approval of a responsive direct brain stimulation device (The RNS System for focal epilepsy), the Food and Drug Administration required a prospective study to evaluate whether real-world safety and effectiveness differed from outcomes in the RCT.
Methods
This open-labeled study enrolled adult participants who met the RNS System–approved indication for use. The primary effectiveness end point was median percent change in seizure frequency at 3 years of treatment. Interim safety is presented; the primary safety endpoint analysis will be conducted at 5 years.
Results
Across 32 US epilepsy centers, 324 patients (mean age 37.1, 59.6% female individuals) were implanted and 271 completed 3 years of follow-up. The median percent reduction in seizure frequency at 6 months was 62% and 82% at 3 years (p < 0.0001; Wilcoxon signed-rank test); 41% had a ≥90% reduction in seizure frequency at 3 years, 42.5% of participants had at least 1 seizure-free period of 6 months or more, and 22.0% experienced seizure freedom for 12 months or more. Observed effectiveness was similar across patients with 1 or 2 seizure onsets and across onset locations (mesial temporal, neocortical, or both mesial temporal and neocortical). No serious stimulation-related adverse events were reported. Combining data from all RNS System trials (n = 645), the sudden unexplained death in epilepsy (SUDEP) rate was 2.3/1,000 patient years, which was significantly lower than predefined comparators (p < 0.05; 1-tailed χ2).
Discussion
This prospective real-world study contributes to the body of evidence that adjunctive direct brain-responsive neurostimulation provides significant and sustained reductions in the frequency of focal seizures. Seizure reductions were greater and were achieved faster than in the RCT and long-term treatment trials but were similar to a more recent retrospective multicenter real-world study. As in the preapproval studies, treatment was well-tolerated and safe, and the SUDEP rate was low. The RNS System showed similar safety and improved seizure outcomes in real-world use compared with the RCT. Improvements in efficacy may reflect changes in programming practices. Future research efforts will focus on using the brain data obtained by the device to optimize detection and stimulation paradigms for each patient.
Trial Registration Information
ClinicalTrials.gov, NCT02403843, submitted March 26, 2015.
Classification of Evidence
This study provides Class IV evidence that in adults with refractory focal-onset seizures, direct brain-responsive neurostimulation reduces seizure frequency without serious adverse events up to 3 years.
Introduction
Responsive direct brain neuromodulation is an effective treatment option for many of the approximately 30% of patients with drug-resistant focal epilepsy who are not candidates for resection or ablation procedures.1,23 Only 1 responsive direct brain neurostimulation device is approved by the US Food and Drug Administration (FDA) for any epilepsy indication; the RNS System is indicated as an adjunctive treatment for adults with medically refractory focal-onset seizures arising from 1 or 2 seizure foci or regions. The evidence for safety and effectiveness that led to FDA approval in 2013 was provided by a 2-year feasibility trial (n = 65) and a double-blinded randomized controlled trial23 (n = 191), followed by an open-label 7-year long-term treatment (LTT) trial that enrolled patients from both the feasibility and pivotal studies4 (n = 230).
An open-label, prospective, postapproval study (PAS) was required by the FDA as a condition of the RNS System approval. This study was designed to provide real-world data in patients treated according to the approved indication for use. Results for the primary efficacy end point at 3 years after implant, as well as interim safety results, are presented here. These data represent the largest multicenter, prospective, real-world study on neuromodulation for drug-resistant focal epilepsy to date.
Methods
The RNS System (NeuroPace, Inc., Mountain View, CA) and its functionality have been described in detail elsewhere.4 In brief, a cranially implanted programmable neurostimulator is connected to two 4-contact depth and/or subdural cortical strip leads. Physicians select the location of the leads according to the region of seizure onset; after implant, the physician programs the device to detect patient-specific epileptiform activity based on the device-recorded data (Figure 1). Intracranial EEG activity is continuously sensed, and when the predetermined detection patterns occur, stimulus pulses are provided. Detection and stimulation parameters are adjusted for each patient as needed for seizure reduction.
Figure 1. Schematic Representation of the RNS System.
The top panel illustrates the placement of the neurostimulator, a cortical strip lead, and a depth lead. The bottom panels show the physician interface with the patient data management system. The physician can visualize continuous detection trends over time to review captured intracranial EEG data and change programming settings using this interface. Copyright © 2025 NeuroPace, Inc.
The PAS enrolled participants who met the FDA-approved indication for use: aged 18 years or older, had focal seizures, had undergone diagnostic testing that localized no more than 2 epileptogenic foci, were refractory to 2 or more antiseizure medications (ASMs), and had frequent and disabling seizures (focal preserved consciousness [FPC] motor, focal impaired consciousness [FIC], and/or secondarily generalized seizures) (this terminology has been updated in line with the 2025 International League Against Epilepsy Guidelines5). The study was conducted at US clinical epilepsy centers by qualified neurologists, epileptologists, and neurosurgeons trained on the RNS System.
Participants provided a retrospective 3-month preimplant seizure rate and then maintained a seizure diary from implant throughout study participation. Follow-up visits were designated at 2 weeks; 1, 3, 6, 9, 12, and 16 months; and then every 4 months thereafter through 60 months. Office visits included a neurologic examination, medication and diary review, and review of any detection/stimulation programming changes.
Adverse events (AEs), as defined by FDA, were collected from the time of enrollment and were categorized by the clinical site principal investigator by severity (mild or serious) and relatedness to the device (device related, of uncertain device relation, or not device related). All enrolling sites had monitoring visits conducted to assess data for accuracy and completeness per the study monitoring and data management plans. AEs were assigned codes by the NeuroPace Chief Medical Officer according to the Medical Dictionary for Regulatory Activities. An independent data monitoring safety committee reviewed all AEs to ensure ongoing safety to study participants and provided a recommendation after each meeting regarding any necessary modifications and for continuation/discontinuation of the study. An independent sudden unexplained death in epilepsy (SUDEP) analysis committee adjudicated SUDEP events.
Participants could withdraw consent at any time in the study, or the physician could decide that continued participation was not in the patient's best interest. Patients were required to be withdrawn if they enrolled in another therapeutic investigational drug or device study, received another device that provides electrical energy to the brain, were noncompliant with protocol requirements or elected to stop treatment, or did not replace a device at end of service.
Sample size calculation for a regression model, with a log link and including a term for linear slope in the dependent variable, indicated that 240 patients would provide more than 80% power for the ability to detect a possible slope equivalent to a 20% cumulative increase in seizures over the period from the preimplant baseline to 30–36 months after implant. A dropout rate of up to 25% was used to ensure an adequate sample size.
Interim safety outcomes are presented as types and rates of AEs; final safety results will be available when the final participant reaches 5 years of follow-up after implant. Differences between serious AE (SAE) rates in the subpopulations were evaluated using the pairwise z-test for proportions.
The prespecified primary effectiveness end point was the median percentage reduction in disabling seizures at 30–36 months after implant compared with the preimplant baseline. A secondary safety endpoint was to assess whether there was an increase in seizure rates from 6 months after implant through 36 months after implant. This analysis fit seizure data into a generalized estimating equation (GEE) model. The dependent variable was the 6-month period, and the independent variable was the percent change from baseline in each 6-month period. A compound symmetric correlation structure was used. A straight line was fitted to the data to evaluate the presence of a linear trend over time. This analysis was conducted using all available data. Participants who did not provide seizure frequency data for the 3 months preceding implant and/or for the 6-month interval were not included in the analysis. For participants with available seizure data during any portion of the preimplant or 6-month interval, those data were used in the calculation of the overall mean for each respective period.
Safety outcomes based on the experience of the implanting physician and center were also evaluated. The adjusted SAE rate during the first year after implant (implant through day 360) was compared against the total number of patients who were ever implanted with an RNS System at that site using a negative binomial analysis.
Although the study was not specifically powered to detect differences between subgroups, exploratory analyses were implemented to test for observable differences between subpopulations of interest. The Kruskal-Wallis test (α = 0.05) was used to asses differences in outcomes based on clinical features such as age at enrollment (by median split), age at onset (by median split), previous surgery for epilepsy, previous intracranial monitoring, previous treatment with vagus nerve stimulation (VNS), abnormality on brain MRI, number of seizure-onset zones (1/2), and seizure-onset location (mesial temporal [MTL]/neocortical/both).
ASMs could be adjusted as medically necessary during the study; medication changes were tracked and the effect of changing ASM on seizure frequency was evaluated. A significant change in ASM was defined as a ≥25% difference in the dose, the addition of a medication, the discontinuation of a medication, or any combination of these. A qualifying increase combined with a qualifying decrease in 1 or more ASM was considered “mixed.” The reduction in clinical seizure frequency during the past 6 months of follow-up was compared between the 4 patient groups (increase, decrease, mixed, or no change) using a Kruskal-Wallis omnibus test (α = 0.05).
Standard Protocol Approvals, Registrations, and Patient Consents
All study protocol versions were approved by the FDA and applicable institutional review boards, in accordance with all applicable United States Code of Federal Regulations governing the protection of human participants. All participating clinical investigators provided financial disclosure forms at the start of the study, indicating any financial or competing interests. All patients or their legal representatives gave written informed consent after being provided with all trial information by a trained member of the study staff at the clinical site.
The PAS study was registered on ClinicalTrials.gov (NCT02403843) in March 2015 and includes a listing of all participating clinical sites and will list all final study results, as available.
Data Availability
The data analyzed in this study is subject to the following licenses/restrictions: Data in this publication may be made available to researchers for academic use. Requests sent to research-requests@neuropace.com will be reviewed in accordance with NeuroPace data sharing policy.
Results
A total of 343 participants enrolled across 32 epilepsy centers in the United States, and 324 participants were implanted with the NeuroPace RNS Neurostimulator and Leads (lead locations are provided in eTable 1). The first enrollment occurred in June 2015 and the final enrollment on December 30, 2020. The final patient reached 3-year postimplant follow-up on November 28, 2023, providing 1,381 patient-implant years and 1,301 years of brain-responsive neurostimulation for the 3-year analysis. Participant accountability is illustrated in Figure 2. The most common reasons for withdrawal from the study were to pursue other treatments, because the participant moved away from the clinical site's geographic area, or because of physician preference due to poor participant compliance.
Figure 2. Participant Accountability.
Of the 343 participants enrolled, 324 were implanted, 307 completed 1 year, 289 completed 2 years, and 271 completed 3 years of follow-up. These numbers represent the intent-to-treat cohort at their respective time points.
Demographic and baseline characteristics for all implanted patients are provided in eTable 2.
The mean age at enrollment was 37 years (range 18–76), and the mean duration of epilepsy was 17.8 years (range 2–62). The median preimplant seizure frequency was 6 seizures per month. Fifty-five percent of participants had seizures arising from the MTL (75.8% of this subset had bilateral MTL onsets), 34% from the neocortex, and 11% from both the neocortex and MTL. Seizure types at baseline in the full cohort were generalized tonic clonic (87.7%), FIC (87.7%), and FPC motor (16.0%) and nonmotor (62.7%). Previous epilepsy-related procedures included intracranial monitoring for seizure localization (67% for all implanted), epilepsy surgery (resection or laser interstitial thermal therapy (LITT); 19.4% for all implanted), and VNS (9.3% for all implanted patients).
Effectiveness
Seizure Reduction
The median percentage change from baseline in seizure frequency by 6-month periods beginning from time of implant to 36 months is given in Table 1. There was a statistically significant (Wilcoxon signed-rank test) reduction in the frequency of seizures compared with baseline at each of the 6-month intervals. In the first 6 months of responsive stimulation treatment, there was a 62.5% median reduction in seizures (p = 3.3 × 10−17). Seizure reduction improved over time, reaching 82% at year 3 (p = 2.2 × 10−22). Similar results in the intent-to-treat population (full cohort) were observed using a constant cohort (patients in trial for entire duration with sufficient seizure diary data to be included at 1 or more time points), complete cohort (patients in trial for entire duration with seizure diary data at the 30–36 month time point), and last observation carried forward (LOCF, eTable 3).
Table 1.
Primary Effectiveness End Point by Time From Implant
| Trial periods | Full cohort | MTL | Neocortical | MTL + neocortical | ||||
| N | Median (upper and lower quartiles), % | N | Median (upper and lower quartiles), % | N | Median (upper and lower quartiles), % | N | Median (upper and lower quartiles), % | |
| Months 0–6 | 314 | −62.5 (−91.6, −11.8) | 172 | −60.9 (−91.6, −15.8) | 106 | −66.0 (−91.6, 0.0) | 36 | −59.0 (−91.4, −22.2) |
| Months 6–12 | 292 | −62.3 (−93.3, −14.6) | 159 | −61.1 (−91.7, −14.3) | 99 | −64.4 (−94.7, −12.1) | 34 | −56.3 (−97.6, −18.8) |
| Months 12–18 | 282 | −68.2 (−97.4, −15.5) | 153 | −66.9 (−95.1, −16.7) | 97 | −62.5 (−97.4, −22.4) | 32 | −84.5 (−100.0, −6.4) |
| Months 18–24 | 273 | −69.4 (−98.4, −19.4) | 148 | −64.9 (−97.7, −14.1) | 93 | −69.0 (−98.3, −25.0) | 32 | −85.6 (−100.0, −42.1) |
| Months 24–30 | 260 | −75.0 (−99.2, −25.8) | 137 | −75.0 (−97.2, −27.3) | 90 | −75.0 (−97.7, −28.6) | 33 | −81.3 (−100.0, −19.6) |
| Months 30–36 | 255 | −82.0 (−100.0, −29.2) | 137 | −73.5 (−97.8, −25.0) | 87 | −90.0 (−100.0, −47.8) | 31 | −89.6 (−100.0, −15.6) |
Abbreviations: ITT = intent-to-treat; LOCF = last observation carried forward; MTL = mesial temporal.
Changes in seizure frequency after implant compared with preimplant baseline are shown by 6-month intervals; negative values represent a reduction in seizure frequency. The first and third quartiles are shown in parentheses. N-values represent the number of participants with seizure diary data in the corresponding time intervals. Values represent the outcomes for the ITT cohort; results for the constant cohort, complete cohort, and LOCF can be found in the eTable 4, eTable 5, and eTable 3, respectively.
Figure 3A presents the median percent reduction over time for all participants (full cohort), participants with MTL leads only, participants with neocortical leads only, and participants with a neocortical lead and an MTL lead. The magnitude of reduction of disabling seizures was similar in participants with MTL seizure onsets (unilateral or bilateral), neocortical onsets, or both MTL and neocortical onsets (p = 0.168; Kruskal-Wallis).
Figure 3. Seizure Outcomes for All Participants.
(A) Median percent reduction in seizure frequency by 6-month periods by onset zone. The median percent reduction from baseline is shown for all participants with data during the specified time period. A negative number represents a reduction in seizure frequency. The different lines represent the complete cohort (solid black), participants with MTL leads only (solid blue line), participants with neocortical leads only (dashed), and participants with both a neocortical lead and an MTL lead (dotted). The seizure reduction from baseline increased for all cohorts over time, regardless of lead location. The proportions of participants in each cohort at each time point and CIs are given in Table 1. (B) The percent reduction in seizure frequency compared with baseline is shown for all participants with at least 91 days of seizure diary data in the analysis window. The analysis window includes the 6 months before the 36-month appointment or the 6 months before the last follow-up (last observation carried forward). Values less than zero indicate an increase in seizure frequency; values greater than zero indicate a decrease in seizure frequency. The horizontal lines indicate an increase of ≥50% (yellow), a decrease of ≥50% (blue), a decrease of ≥90% (purple), and seizure freedom (green), respectively. MTL = mesial temporal.
Figure 3B presents the distribution of individual responses for the most recent 6-month period for which there was at least 91 days of seizure diary data (LOCF). The responder rate (≥50% reduction in seizures) was 66%, 41% had a 90% or greater reduction in their seizures, and 24% were seizure free in the past 6 months of follow-up.
For participants who had sufficient data to perform the analysis at each duration, 42.5% (n = 114) had at least 1 seizure-free period of 6 months or more, 22.0% (n = 59) had at least one 12-month or longer period without seizures, and 16.0% (n = 43) were seizure free for 18 months or longer at some point during their participation in the study.
Seizure Reductions and Clinical Characteristics
There was no observed difference in median percent reduction in seizure frequency with respect to the participants' clinical characteristics (Kruskal-Wallis). Improvements in the median percent reduction in seizure frequency were similar for participants with and without previous epilepsy surgery (p = 0.54), previous VNS treatment (p = 0.13), or intracranial monitoring (p = 0.30). Similarly, the reduction in seizure frequency was not observably influenced by age at enrollment (p = 0.96), duration of epilepsy (p = 0.53), the presence or absence of any brain abnormality on imaging (p = 0.90), or whether there were 1 or 2 seizure foci (p = 0.31).
Seizure Reductions and ASMs
Participants' ASMs were often adjusted over the follow-up period, as permitted in the protocol. At last follow-up, no statistically significant (Kruskal-Wallis) differences were observed in the efficacy endpoints between those who had an addition or increase in ASM (14%), those who had a decrease (13%), those who had both an increase and a decrease (56%), and those who had no change (16%) (p > 0.05) (Table 2). Outcomes for patients taking specific ASMs are presented in eFigure 1.
Table 2.
Changes in Seizure Frequency and Response Rate Categorized by Changes to ASMs
| Changes in ASMs | No. of patients | Median change (%) | 0–25th Percentile (%) | 75–100th Percentile (%) | Responder rate (%) | No. of responders |
| No change | 51 | −82 | −52 | −100 | 76 | 39 |
| Increase | 45 | −67 | −16 | −98 | 60 | 27 |
| Mixed | 176 | −76 | −25 | −98 | 64 | 112 |
| Decrease | 42 | −95 | −50 | −100 | 74 | 31 |
Abbreviation: ASM = antiseizure medication.
All values represent the last observation carried forward. There were no significant differences between groups (Kruskal-Wallis test, p = 0.0686).
Neurostimulator Settings
During the 3-year follow-up period, participants had a median of 867 detections per day (mean = 1,077), leading to a median of 878 stimulations delivered (mean = 1,066), and a median of 217.32 seconds of stimulation per day (range 0.64–9,340.5 seconds). The most common stimulation settings were 200-Hz burst frequency, 160-μs pulse width, and 100-msec burst duration. These are consistent with a suggested therapy protocol based on experience from the preapproval trials (RNS System Suggested Therapy Protocol). The Therapy Protocol also suggests an initial charge density of 0.5 μC/cm2 with a stepwise increase over time, according to tolerability and the patient's clinical response. The most common charge density over the 3 years was 2.0 μC/cm2. Differences in frequency and charge density programming for participants in the LTT study (includes the feasibility and pivotal study cohorts), and this PAS, are presented in Figure 4.
Figure 4. Programming Differences Between RNS System Focal Epilepsy Clinical Trials.
(A) Average charge density in RNS System focal epilepsy clinical trial participants over time. Bars represent the mean charge density for all patients with data in the specified time range for participants in the LTT trials (orange) and the PAS trial (blue); error bars correspond to standard error. The charge density is a derived value and is a function of the current amplitude and the stimulation pathway configuration. (B) Distribution of programmed frequencies for participants in the LTT and PAS trials. Values represent the percentage of all time spent at each frequency normalized to the total amount of time in the trial across all patients enrolled in the respective studies. LTT = long-term treatment; PAS = postapproval study.
Safety
There were no SAEs reported to be related to the chronic use of the RNS System. There were 22 SAEs in 22 participants that were reported as related to the surgical procedure; all were designated by the site principal investigator as anticipated and all resolved within the follow-up period. The surgery-related SAEs were implant site infection (n = 7; 2.2%), intracranial hemorrhage (n = 6; 1.8%), implant site erosion/dehiscence (n = 2; 0.6%), CSF leakage (n = 2; 0.6%), and 1 event each (0.3%) of convulsive status epilepticus, facial droop, airway edema, postoperative pain, and pulsatile tinnitus. The percentage of participants reporting an SAE each year was similar in each of the onset zone subpopulations (MTL vs neocortex, p = 0.38; MTL vs MTL + neocortex, p = 0.54; neocortex vs MTL + neocortex, p = 0.97).
The GEE analysis showed no evidence of seizure worsening. The estimate of the slope of linear change was negative with a p value of 0.0015, showing a statistically significant decrease in seizures over time. Similarly, the estimates of the slope of linear change were negative for all subpopulations, suggesting trends toward a decrease in seizures over time.
SAEs related to intracranial hemorrhage totaled 12 events in 12 participants (3.7%); 7 of the 12 events (58.3%) occurred within the 84-day postoperative period (6 of these were reported as procedure related). These were intracranial hemorrhage (n = 2), subdural hematoma (n = 2), cerebral hemorrhage (n = 1), epidural hemorrhage (n = 1), and extradural hematoma (n = 1). Intracranial hemorrhages occurring outside the postoperative period (n = 5) were intracranial (n = 2), subdural (n = 2), and cerebral (1). Two of these events were due to a fall associated with a seizure. All hemorrhage events resolved, except for 1 participant with a mild expressive aphasia at study exit.
Implant-site infections were reported in 9 participants (9/324, 2.8%), a median of 89 days after implant (mean 278 days, range 22–830 days), and all were considered resolved during the follow-up period without ongoing sequelae. The neurostimulator was explanted in 8 participants, and 2 elected to have a reimplantation.
There were no psychiatric-related SAEs reported as related to the implant procedure or the device. Twenty-three psychiatric-related SAEs that were not procedure or device-related were reported in 17 participants (17/324, 5.2%). The most frequently reported events were suicidal ideation (n = 4) and postictal psychosis (n = 4). Three events were related to attempted suicide: there were no suicides. Other SAEs in this category were psychogenic seizure (n = 3), mental disorder (n = 3), psychotic disorder (n = 1), depression (n = 1), conversion disorder (n = 1), agitation (n = 1), adjustment disorder with depressed mood (n = 1), and alcoholism (n = 1). All participants with suicidality-related SAEs (n = 7) had a preexisting history of depression; 5 had a preexisting history of suicidality or suicide attempts.
A comparison of the 10% of patients who had a ≥50% increase in seizures at the last observation with patients who either improved or for whom there was no change in seizure frequency at the last observation revealed similar demographic characteristics; the distribution of features suggests that these patients may have had a longer lasting and/or more severe disease, but no single characteristic was associated with seizure worsening. There were no AEs related to new or severe disabling seizure types, and no participants reported an increase in seizures as a reason for withdrawal from the study.
SUDEP
Two deaths were adjudicated as probable or definite SUDEP events in the 3-year period. To estimate the true rate of SUDEP after RNS System implantation, all 645 participants in the RNS System premarket and PAS clinical trials were included in an analysis, together representing more than 3,000 patient years and providing a SUDEP rate of 2.3/1,000 patient years. This rate is statistically significantly less than the expected rate of 6.9/1,000 patient years for patients allocated to the placebo arm of ASM trials.6
Safety by Physician Experience With the RNS System
Safety outcomes did not differ based on neurosurgeon experience or center experience (negative binomial test). There were no differences in SAE rates based on the experience of these NeuroPace qualified and trained surgeons (p < 0.51) and no difference in the rate of SAEs overall based on epilepsy center (p < 0.63).
Classification of Evidence
This study provides Class IV evidence that in adults with refractory focal-onset seizures, direct brain-responsive neurostimulation reduces seizure frequency without SAEs up to 3 years.
Discussion
This postapproval study supports the safety and effectiveness of the RNS System as used in the real world for adults with drug-resistant focal-onset seizures who meet the current indication for use. Three hundred twenty-four adults were treated with the RNS System in this PAS, and 271 provided data on effectiveness for more than 3 years of treatment. The effectiveness of treatment exceeded the experience in the Pivotal2 and LTT4 trials. Surgical and clinical outcomes did not vary by surgeon or site experience, suggesting that current qualification and training requirements are sufficient to support safe and effective use.
Treatment with the RNS System produced a 62% median reduction in seizure frequency at 6 months with progressive improvement over time. At 30–36 months after neurostimulator and leads implant, the observed median reduction in seizure frequency was 82%, with more than one-third of participants reporting a 90% or greater reduction in seizures. Many participants experienced extended periods of seizure freedom. Over the course of 3 years, 42% of participants experienced at least 1 seizure-free period that lasted 6 months or more, and 22% had at least 1 period of 12 months or longer without seizures. These results are especially meaningful when considering that, on average, these participants had a nearly 20-year history of epilepsy, had more than 5 disabling seizures a month at baseline, had failed multiple ASMs, and often underwent surgical and VNS treatments.
Participant characteristics that differed between this study and the previous LTT trial44 were that fewer had already been treated with an epilepsy surgery (19.4% vs 34%) or VNS therapy (7.7% vs 32%). This could indicate that brain-responsive neurostimulation is being considered earlier in the course of patient therapy or that the RNS System is being offered to patients who are not considered to be good candidates for epilepsy surgery or VNS and would not have otherwise been offered another therapy. This suggests that treatment options for patients with drug-resistant focal epilepsy have expanded in practice.
Although the study was not powered for subgroup comparisons, differences in reductions in seizure frequency between participants with and without previous resective surgery, VNS, or intracranial monitoring were not observed. Similarly, there were no observed differences in seizure reduction based on participant characteristics, including seizure-onset location (MTL or neocortical), number of foci (1 or 2), or identification of a lesion on brain MRI.
The observed seizure reduction of 82% at the 3-year time point in this study was notably better than that observed in the LTT trial, in which the median percent reduction in seizure frequency was 58%. This is unlikely to be related to patient selection, because the clinical characteristics of participants in this study were similar to those of the LTT trial participants. It is more likely that these results reflect changes in the application of responsive neurostimulation based on experience gained in the Pivotal and LTT trials (RNS System Suggested Therapy Protocol; Figure 4). One notable change in programming was the recommendation to begin with a lower initial charge density, followed by slower stepwise increases in charge density as clinically indicated. Similar patterns in charge density programming were observed in a real-world retrospective study7 in 150 patients across 8 centers that reported similar seizure reductions as in this prospective PAS. An additional observed difference in device programming in the PAS was that more patients were programmed to 200-Hz stimulation frequency, rather than 100 Hz, which was more frequently used in the LTT trial. This also reflects the current recommendation to begin stimulation at 200 Hz. Future investigations will use empirical clinical data from this trial, as well as the intracranial EEG data obtained by the device, to further understand how device programming is related to clinical outcomes and to continue to refine therapy strategies.
Consistent with other studies in neuromodulation,1,7-9 the clinical response to brain-responsive neurostimulation improved over time. This raises the possibility that there could be longer term neuroplastic effects of neurostimulation that result in continued improvement in outcomes. Further research is required to better understand this phenomenon.
The primary safety endpoint and the full safety analysis are performed at 5 years. These interim safety results support that responsive neurostimulation is well tolerated and safe over time. AEs related to the implanted device, including infection, were anticipated, and the rates were not higher than that reported in the LTT trial, with implantation of intracranial electrodes to localize the seizure focus,10-12 with resective epilepsy surgery,10,13,14 or with DBS devices for treatment of movement disorders15 or for epilepsy.4,9,16
Because SUDEP is a rare event, the SUDEP rate after RNS System implantation was estimated using data from all RNS System prospective clinical trials, including this study. The estimated rate of 2.3/1,000 patient years is significantly less than the rate of 6.9/1,000 patient years reported for patients randomized to the placebo arm of ASM studies.6 This 3-fold reduction in SUDEP rate is a clinically meaningful benefit.
Psychiatric SAEs, including suicidality and suicidal ideation, were infrequent and not device related. There were no SAEs related to memory impairment. Although neuropsychiatric outcomes were not collected as part of this study, previous studies showed no overall adverse changes in mood or suicidality17 and an improvement in naming and verbal learning after therapy with the RNS System.18
A standard qualification process developed by the device manufacturer in consultation with the FDA ensures that neurosurgeons implanting the neurostimulator and leads and neurologists managing treatment with the device have appropriate professional and device-specific training. For these qualified PAS physicians, previous experience in treating patients with the RNS System did not affect patient safety or likelihood of benefit. There was no difference in safety at the 3-year endpoint based on the experience of NeuroPace qualified implanting surgeons or comprehensive epilepsy centers.
These results are from an open-label, real-world study and may be influenced by selection bias, a retrospective baseline, expectation bias, a prolonged placebo response, or regression to the mean. However, data were collected over several years and seizure frequency improved in all seizure-onset groups. Another consideration is that the trial was conducted during the coronavirus disease 2019 pandemic. During this time, patient appointments were canceled, some patients were lost to follow-up, and some data were missing. However, all trial data were monitored against source documents, assuring that data were as complete as were possible. Ultimately, these potential limitations and biases are aspects of real-world use, which was the primary objective of the study.
This FDA required and reviewed PAS represents a large prospective study of brain neuromodulation in patients with drug-resistant focal epilepsy. The median percent reduction in seizures was greater and achieved more rapidly compared with earlier randomized controlled (Pivotal) and prospective open-label long-term follow-up (LTT) trials. The safety of the surgical procedure and the implanted device was consistent with these earlier RNS System trials and compared favorably with other brain stimulation devices used for treatment of movement disorders15 and epilepsy.9 Patients with leads in the MTL only, the neocortex only, and both MTL and the neocortex all showed significant reduction in median seizure frequency that improved over time, as did patients with 1 or 2 onset zones. This speaks to the flexibility of the RNS System which can treat focal seizures arising from all of these brain areas.
These results contribute to the body of evidence that adjunctive treatment with brain-responsive neurostimulation at the site(s) or region(s) of seizure onset can provide substantial benefit to patients with drug-resistant focal-onset epilepsy. Future research will seek to identify detection and stimulation paradigms that optimize therapy by combining the clinical trial data with machine learning and other Artificial artificial Intelligence (AI) analyses of the device-obtained intracranial EEG data. Given the numerous detection and stimulation options provided by the device, it is likely that these investigations will be fruitful and will further drive the benefit of responsive neurostimulation in patients with drug-resistant epilepsy.
Acknowledgment
The authors thank the patients and families for participating in the RNS System trial. They also thank the members of the Data Safety Monitoring Board, who were responsible for independently monitoring the safety of interventions by reviewing data during the RNS System trials, and members of the SUDEP Analysis Committee, who were responsible for reviewing data regarding any deaths that occur for participants participating in the RNS System trials. Data Safety Monitoring Board: David Spencer, MD (Chair), Robert Goodman, MD, PhD, Steven Karceski, MD, Tami Crabtree, PhD. Previous Data Safety Monitoring Board: Roger Porter, MD (chair), Gary Mathern, MD, PhD, John Pellock, MD (in memoriam), Joan Conry, MD, Lorene Nelson, PhD, and Dan Bloch, PhD. In addition, the authors thank the following individuals for their contributions: Celeste Gonzalez (Clinical study coordinator, AECC); Rachel VandeGuchte, RN (Clinical study coordinator, Corewell Health); Easton Lumsden (Clinical study coordinator, Corewell Health), Rachel Fabris (Sub-Investigator, Corewell Health), Shan Abbas (Sub-Investigator, Corewell Health), Jodi Kortman (Sub-Investigator, Corewell Health); Krishna Saini (Clinical study coordinator, Dell Seton Medical Center at the University of Texas); Paulina Henriquez Rojas, MD (Clinical study coordinator, Yale); Tyler Gray (Clinical study coordinator, Yale); Tuan Bui (Clinical study coordinator, Yale); Bonnie Chen (Clinical research coordinator, Yale); Joan Nye (Clinical research coordinator, Yale); Kelly Benson-Atwood (Clinical research coordinator, Yale); Jordan Seliger (Clinical study coordinator, Stanford Hospital and Clinics); Christine Lin (Clinical study coordinator, Stanford Hospital and Clinics); Ruba Shaik (Clinical study coordinator, Stanford Hospital and Clinics); Brianna Heath (Clinical study coordinator, Stanford Hospital and Clinics); Adele Viviani (Clinical study coordinator, Stanford Hospital and Clinics); Marina Azevedo (Clinical study coordinator, University of South Florida); Particia Orozco (Clinical study coordinator, University of North Carolina Hospitals); Jessica Ferrall (Clinical study coordinator, University of North Carolina Hospitals); Kathleen Hernando (Clinical study coordinator, University of Alabama Birmingham); Samantha Fry (Clinical study coordinator, University of Alabama Birmingham); Anna Moyana (Clinical study coordinator, University of Alabama Birmingham); Kaitlyn McCormick-Kane (Clinical study coordinator, University of Alabama Birmingham); Blake Newman (Neurologist, University of Utah); Amir Arain (Neurologist, University of Utah); Emerald Wan (Clinical research coordinator, UCSF) and Meera Rao (Clinical Research Coordinator, UCSF). The statistical analysis was performed by independent contractors Tami Crabtree and Felice Sun. The authors acknowledge the contributions of NeuroPace team members, Felicia Elefant, Tricia Cunningham, Keila Benjamin Sequeira, Roger Duguid, Julie Park, Victoria Nagle, Emy Chow-Greiner, Jody Ellman, Caryl Tongco, and Robert Blair.
Glossary
- AE
adverse event
- ASM
antiseizure medication
- GEE
generalized estimating equation
- FDA
Food and Drug Administration
- FIC
focal impaired consciousness
- FPC
focal preserved consciousness
- LOCF
last observation carried forward
- LTT
long-term treatment
- MTL
mesial temporal
- PAS
postapproval study
- SAE
serious AE
- SUDEP
sudden unexplained death in epilepsy
- VNS
vagus nerve stimulation
Appendix. Coinvestigators
| Name | Location | Role | Contribution |
| Kort Elisevich | Corewell Health (Butterworth) | Neurosurgeon | Coinvestigator and performed early implantations |
| Emory Peng | Massachusetts General Hospital | Study coordinator | Study Coordination |
| Rajashi Mazumder | Ronald Reagan UCLA Medical Center | Site investigator | CRF |
| Tara Jennings | University of Pennsylvania | Site investigator | Trial visits |
| Melissa Johnston Esparza | University of Pennsylvania | Project manager | Managed personnel |
| Jacob Pellinen | University of Colorado Hospital | Site investigator | Study support and contributing participants |
| Nitish Harid | University of Colorado Hospital | Site investigator | Study support and contributing participants |
| Pue Farooque | Yale New Haven Hospital | Site investigator | Conducted study visits |
| Aline Herlopian | Yale New Haven Hospital | Site investigator | Conducted study visits |
| Lawrence Hirsch | Yale New Haven Hospital | Site investigator | Conducted study visits |
| Adithya Sivaraju | Yale New Haven Hospital | Site investigator | Conducted study visits |
| Eyiyemisi Damisah | Yale New Haven Hospital | Neurosurgeon | Performed implantations |
| Jason Gerrard | Yale New Haven Hospital | Neurosurgeon | Performed implantations |
| Angela Wabulya | University of North Carolina Hospitals | Site investigator | Patient recruitment, study support |
| Kristen Riley | University of Alabama Birmingham | Neurosurgeon | Surgery |
| J. Nicole Bentley | University of Alabama Birmingham | Neurosurgeon | Surgery |
| Sandipan Pati | University of Alabama Birmingham | Neurologist | Patient care |
| Jessica Johnson | Memorial Hermann Healthcare | Study coordinator | Study coordination and data management |
| Kaitlyn Kirchoffer | Ronald Reagan UCLA Medical Center | Study coordinator | Aiding in study |
| Yvan Bamps | Emory University | Study coordinator | Study support |
Contributor Information
Collaborators: Kort Elisevich, Emory Peng, Rajashi Mazumder, Tara Jennings, Melissa Johnston Esparza, Jacob Pellinen, Nitish Harid, Pue Farooque, Aline Herlopian, Lawrence Hirsch, Adithya Sivaraju, Eyiyemisi Damisah, Jason Gerrard, Angela Wabulya, Kristen Riley, J. Nicole Bentley, Sandipan Pati, Jessica Johnson, Kaitlyn Kirchoffer, and Yvan Bamps
Author Contributions
D. Eliashiv: major role in the acquisition of data. V.R. Rao: major role in the acquisition of data. B.C. Jobst: major role in the acquisition of data. J.P. Szaflarski: major role in the acquisition of data. J.D. Rolston: major role in the acquisition of data. L.C. Kaye: major role in the acquisition of data. T.M. Ganguly: major role in the acquisition of data. K. Bullinger: major role in the acquisition of data. P.C. Dugan: major role in the acquisition of data. D.E. Burdette: major role in the acquisition of data. A.Y. Peters: major role in the acquisition of data. A. Sheikh: major role in the acquisition of data. K.F. Haas: major role in the acquisition of data. D.R. Nair: major role in the acquisition of data. L. Mnatsakanyan: major role in the acquisition of data. I.H. Quraishi: major role in the acquisition of data. M.K. Bensalem-Owen: major role in the acquisition of data. M.J. Doherty: major role in the acquisition of data. B. Razavi: major role in the acquisition of data. T.L. Fisher: major role in the acquisition of data. C. Skidmore: major role in the acquisition of data. P.N. Modur: major role in the acquisition of data. T.M. Constantino: major role in the acquisition of data. V. Salanova: major role in the acquisition of data. A.J. Cole: major role in the acquisition of data. O. Taraschenko: major role in the acquisition of data. A. Rivera-Cruz: major role in the acquisition of data. J.W. Wheless: major role in the acquisition of data. N. Tandon: major role in the acquisition of data. A. Balabanov: major role in the acquisition of data. S. Aboumatar: major role in the acquisition of data. I. Fried: major role in the acquisition of data. C. Drees: major role in the acquisition of data. H.W. Shin: major role in the acquisition of data. Z. Jaisani: major role in the acquisition of data. S.E. MacIver: major role in the acquisition of data. S.E. Patra: major role in the acquisition of data. E.F. Chang: major role in the acquisition of data. J.T. Willie: major role in the acquisition of data. R. Gwinn: major role in the acquisition of data. T. Stoub: major role in the acquisition of data. J.M. Stern: major role in the acquisition of data. T. Crabtree: analysis or interpretation of data. C.G. Seale: drafting/revision of the manuscript for content, including medical writing for content; study concept or design; analysis or interpretation of data. S.C. McFadden: drafting/revision of the manuscript for content, including medical writing for content. J.F. Norman: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. L. Johnson: drafting/revision of the manuscript for content, including medical writing for content; analysis or interpretation of data. M.J. Morrell: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.
Study Funding
Funding for this study was provided by NeuroPace, Inc. (research-requests@neuropace.com).
Disclosure
NeuroPace Inc. provided funding for the clinical study and is responsible for the study design. D. Eliashiv serves as a consultant and principal investigator for NeuroPace and Medtronic. V.R. Rao served on the Medical Advisory Board for NeuroPace. B.C. Jobst reports disclosures affiliated with the CDC, DoD, Neuropace, NIH, Harvard Pilgrim, Diamond Foundation, NSF, Emory MEW, HOBSCOTCH Institute, Yale START, and Frank Louis and Ruth Professorship. J.P. Szaflarski has received funding from NIH, NSF, DoD, Shor Foundation for Epilepsy Research, UCB Biosciences, NeuroPace Inc., LivaNova Inc., and State of AL; has had consulting/advisory board roles for PureTech Health, Biopharmaceutical Research Company, LivaNova Inc., UCB Biosciences, AdCel Pharma, and iFovea Inc.; is editor-in-chief of Epilepsy & Behavior Reports (paid); is an editorial board member for Epilepsy & Behavior, Journal of Epileptology (associate editor), Journal of Medical Science, and Folia Medica Copernicana; and has served on the Alabama State Medical Cannabis Study Commission (nominated by Gov. Ivey) and on the Alabama Medical Cannabis Commission (nominated by Dr. Harris, State Health Officer). J.D. Rolston receives consulting fees from NeuroPace, Medtronic, Corlieve, Adraxe, and Turing Medical. L.C. Kaye and T.M. Ganguly report no disclosures relevant to the manuscript. K. Bullinger reports receiving grants from the CDC and NIH; a research contract with NeuroPace; and receiving royalties from Wolters Kluwer and Elsevier. P.C. Dugan reports no disclosures relevant to the manuscript. D.E. Burdette conducts contract research and gives promotional talks for NeuroPace. A.Y. Peters, A. Sheikh, K.F. Haas, D.R. Nair, and L. Mnatsakanyan report no disclosures relevant to the manuscript. I.H. Quraishi receives research support from NeuroPace, Lundbeck, Rapport Therapeutics, and the C. G. Swebilius Trust. M.K. Bensalem-Owen serves as a principal investigator for NeuroPace, Xenon, and UCB. M. Doherty reports no disclosures relevant to the manuscript. B. Razavi received research support from NeuroPace, related to this manuscript. T.L. Fisher serves on a Speakers Bureau for Neurelis Inc. C. Skidmore, P.M. Modur, T.M. Constantino, V. Salanova, A.J. Cole, O. Taraschenko, and A. Rivera-Cruz report no disclosures relevant to the manuscript. J.W. Wheless receives funding from the Shainberg Foundation, Jazz, UCB, Neurelis, Marinus, Stoke, SKLSI, TSC Alliance, NeuroEvent Labs, Praxis, Azurity, Biocodex, and LivaNova; is a consultant for Jazz, Neurelis, Takeda, Prais, SJCRH - PTNI, and Azurity; and serves on a Speaker's Bureau for Jazz, UCB, Neurelis, Marinus, SKLSI, Biocodex, and LivaNova. N. Tandon and A. Balabanov report no disclosures relevant to the manuscript. S. Aboumatar serves as a consultant and advisor for UCB Pharma and as a consultant, advisor, and speaker for SK Life Science Inc. I. Fried, C. Drees, and H.W. Shin report no disclosures relevant to the manuscript. Z. Jaisani serves as a coinvestigator on multiple clinical studies for NeuroPace. S.E. MacIver reports no disclosures relevant to the manuscript. S.E. Patra conducted contracted research and promotional talks for NeuroPace. E.F. Chang reports no disclosures relevant to the manuscript. J.T. Willie has a research contract with NeuroPace; has a research contract with and serves as a lecture honoraria and consultant for Medtronic; has a research contract with Abbott; and serves as a consultant for AiM Medical Robotics, Fortec Medical, and Turing Medical. R. Gwinn and T. Stoub report no disclosures relevant to the manuscript. J.M. Stern is a consultant for Xenon, Neurelis, Jazz, UCB, LivaNova, SK Life Sciences, and Ceribell. T. Crabtree is a paid NeuroPace Biostatistician Consultant. C.G. Seale, S.C. McFadden, J.F. Norman, L. Johnson, and M.J. Morrell certify that they are an employee of NeuroPace and have equity ownership/stock options with NeuroPace. Go to Neurology.org/N for full disclosures.
References
- 1.Bergey GK, Morrell MJ, Mizrahi EM, et al. Long-term treatment with responsive brain stimulation in adults with refractory partial seizures. Neurology. 2015;84(8):810-817. doi: 10.1212/WNL.0000000000001280 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Heck CN, King-Stephens D, Massey AD, et al. Two-year seizure reduction in adults with medically intractable partial onset epilepsy treated with responsive neurostimulation: final results of the RNS System Pivotal trial. Epilepsia. 2014;55(3):432-441. doi: 10.1111/epi.12534 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Morrell MJ; RNS System in Epilepsy Study Group. Responsive cortical stimulation for the treatment of medically intractable partial epilepsy. Neurology. 2011;77(13):1295-1304. doi: 10.1212/WNL.0b013e3182302056 [DOI] [PubMed] [Google Scholar]
- 4.Nair DR, Laxer KD, Weber PB, et al. Nine-year prospective efficacy and safety of brain-responsive neurostimulation for focal epilepsy. Neurology. 2020;95(9):e1244-e1256. doi: 10.1212/WNL.0000000000010154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Beniczky S, Trinka E, Wirrell E, et al. Updated classification of epileptic seizures: position paper of the International League Against Epilepsy. Epilepsia. 2025;66(6):1804-1823. doi: 10.1111/epi.18338 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ryvlin P, Cucherat M, Rheims S. Risk of sudden unexpected death in epilepsy in patients given adjunctive antiepileptic treatment for refractory seizures: a meta-analysis of placebo-controlled randomised trials. Lancet Neurol. 2011;10(11):961-968. doi: 10.1016/S1474-4422(11)70193-4 [DOI] [PubMed] [Google Scholar]
- 7.Razavi B, Rao VR, Lin C, et al. Real-world experience with direct brain-responsive neurostimulation for focal onset seizures. Epilepsia. 2020;61(8):1749-1757. doi: 10.1111/epi.16593 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kaufmann E, Peltola J, Colon AJ, et al. Long-term evaluation of anterior thalamic deep brain stimulation for epilepsy in the European MORE registry. Epilepsia. 2024;65(8):2438-2458. doi: 10.1111/epi.18003 [DOI] [PubMed] [Google Scholar]
- 9.Salanova V, Witt T, Worth R, et al. Long-term efficacy and safety of thalamic stimulation for drug-resistant partial epilepsy. Neurology. 2015;84(10):1017-1025. doi: 10.1212/WNL.0000000000001334 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Behrens E, Schramm J, Zentner J, König R. Surgical and neurological complications in a series of 708 epilepsy surgery procedures. Neurosurgery. 1997;41(1):1-10. doi: 10.1097/00006123-199707000-00004 [DOI] [PubMed] [Google Scholar]
- 11.Silberbusch MA, Rothman MI, Bergey GK, Zoarski GH, Zagardo MT. Subdural grid implantation for intracranial EEG recording: CT and MR appearance. AJNR Am J Neuroradiol. 1998;19(6):1089-1093. [PMC free article] [PubMed] [Google Scholar]
- 12.Wong CH, Birkett J, Byth K, et al. Risk factors for complications during intracranial electrode recording in presurgical evaluation of drug resistant partial epilepsy. Acta Neurochir (Wien). 2009;151(1):37-50. doi: 10.1007/s00701-008-0171-7 [DOI] [PubMed] [Google Scholar]
- 13.Engel J, Wiebe S, French J, et al. Practice parameter: temporal lobe and localized neocortical resections for epilepsy. Neurology. 2003;60(4):538-547. doi: 10.1212/01.wnl.0000055086.35806.2d [DOI] [PubMed] [Google Scholar]
- 14.Wiebe S, Blume WT, Girvin JP, Eliasziw M; Effectiveness and Efficiency of Surgery for Temporal Lobe Epilepsy Study Group. A randomized, controlled trial of surgery for temporal-lobe epilepsy. N Engl J Med. 2001;345(5):311-318. doi: 10.1056/NEJM200108023450501 [DOI] [PubMed] [Google Scholar]
- 15.Weaver FM, Follett K, Stern M, et al. Bilateral deep brain stimulation vs best medical therapy for patients with advanced Parkinson disease: a randomized controlled trial. JAMA. 2009;301(1):63-73. doi: 10.1001/jama.2008.929 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Fisher R, Salanova V, Witt T, et al. Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory epilepsy. Epilepsia. 2010;51(5):899-908. doi: 10.1111/j.1528-1167.2010.02536.x [DOI] [PubMed] [Google Scholar]
- 17.Meador KJ, Kapur R, Loring DW, Kanner AM, Morrell MJ; RNS System Pivotal Trial Investigators. Quality of life and mood in patients with medically intractable epilepsy treated with targeted responsive neurostimulation. Epilepsy Behav. 2015;45:242-247. doi: 10.1016/j.yebeh.2015.01.012 [DOI] [PubMed] [Google Scholar]
- 18.Loring DW, Kapur R, Meador KJ, Morrell MJ. Differential neuropsychological outcomes following targeted responsive neurostimulation for partial-onset epilepsy. Epilepsia. 2015;56(11):1836-1844. doi: 10.1111/epi.13191 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data analyzed in this study is subject to the following licenses/restrictions: Data in this publication may be made available to researchers for academic use. Requests sent to research-requests@neuropace.com will be reviewed in accordance with NeuroPace data sharing policy.