PDF: RNS Pivotal Trial
(2011)Objective
To evaluate the safety and efficacy of responsive neurostimulation (RNS System) versus sham stimulation for medically intractable partial epilepsy
Study Summary
• Progressive long-term benefit: median 44% reduction at 1 year, 53% at 2 years; QOLIE-89 improved (p<0.02); surgical risks: ~18% serious procedure-related AEs, ~5% intracranial hemorrhage, ~12% implant-site infection
Intervention
NeuroPace RNS System responsive cortical stimulation vs sham (implanted, not activated)
Inclusion Criteria
Age 18-70, medically intractable focal epilepsy, >=3 disabling seizures/month, failed >=2 AEDs, 1-2 identified foci, not a surgical candidate or declined surgery
Study Design
Arms: Active Responsive Stimulation vs Sham Stimulation
Patients per Arm: Active: 97, Sham: 94
Outcome
• Responder rate (>=50%): 29% vs 27% (not significant during blinded period)
• Long-term open-label: median 44% reduction at 1 year, 53% at 2 years; QOLIE-89 improved (p<0.02); no cognitive deterioration
• AEs: serious procedure-related 18%, implant infection 12%, intracranial hemorrhage 4.7%
Bottom Line
RNS significantly reduced seizures by 37.9% vs 17.3% sham over 12 weeks (P=0.012), with improving response through 2 years (46% responder rate, 7.1% seizure-free at last 3-month period). No adverse cognitive or mood effects. Serious hemorrhage 2.1%, serious infection 4.2%. First Class I evidence for responsive (closed-loop) cortical stimulation. 191 patients, 32 US centers.
Major Points
- Primary endpoint met: seizure reduction 37.9% vs 17.3% sham (P=0.012). Effect grew each month: Month 3 -41.5% vs -9.4% (P=0.008).
- Responder rate improved over time: 29% at 12-week BEP → 43% at 1 year → 46% at 2 years. 7.1% seizure-free at last 3-month period.
- Sham crossover confirmed stimulation effect: when sham group received active stimulation in OLP, seizures reduced (P=0.04).
- No adverse cognitive effects — improvements in verbal, visuospatial, and memory at 1-2 years. QOLIE-89 improved significantly.
- Safety: serious hemorrhage 2.1%, serious infection 4.2% (4 explants). 4 SUDEP deaths over 340 patient-years (within expected range).
- Subgroup consistency: benefit regardless of mesial temporal vs neocortical, 1 vs 2 foci, prior VNS or surgery.
- Closed-loop mechanism: continuous ECoG sensing → detection of abnormal activity → responsive stimulation to seizure focus.
- 191 patients, 32 US centers. Double-blind, sham-controlled (12-week BEP) then open-label.
- 49.7% had prior depression; no increase in depression or suicidality with RNS — important safety distinction from some AEDs.
- First FDA-approved responsive neurostimulation device for epilepsy. Led to NeuroPace RNS System approval.
Study Design
- Study Type
- Multicenter, double-blind, randomized, sham-stimulation controlled trial
- Randomization
- Yes
- Blinding
- Double-blind. Blinded physician gathered outcomes; separate unblinded physician managed device. Adaptive 1:1 randomization.
- Sample Size
- 191
- Follow-up
- 12-week blinded period + 84-week open-label (total ~2 years). Data cutoff June 2010.
- Centers
- 32
- Countries
- United States
Primary Outcome
Definition: Reduction in mean seizure frequency during 12-week BEP vs preimplant
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| -17.3% (95% CI -29.9 to -2.3%) | -37.9% (95% CI -46.7 to -27.7%) | - (GEE P=0.012) | 0.012 |
Limitations & Criticisms
- Industry-sponsored (NeuroPace). Lead author is NeuroPace employee with stock options.
- Short 12-week blinded period. Long-term data are open-label/uncontrolled.
- Implant effect: both groups had early post-implant reduction (confounded Month 1, P=0.279).
- Moderate effect size: only 2.1% seizure-free during BEP.
- Responder rates similar during BEP (29% vs 27%) — difference driven by degree of reduction.
- No active comparator (VNS or DBS).
- Stimulation parameters not specified in publication.
- High baseline psychiatric comorbidity (50% depression) limits generalizability.
Citation
Neurology. 2011;77(13):1295-1304.