Abstracts

Rationale: The default frequency/burst stimulation paradigm (DFB: 200 Hz/100 ms) in Responsive Neurostimulation System (RNS, NeuroPace Inc.) patients is sometimes not effective in terminating the ictal discharges, necessitating alternative stimulation parameters.

Methods: We retrospectively investigated the utility of low frequency/long burst (LFLB: ≤50 Hz/≥200 ms) stimulations in consecutive patients implanted with RNS who were switched to LFLB due to suboptimal response to the initial DFB paradigm. We evaluated the ECoGs stored on the PDMS as "long episode," "long episode saturation," "saturation," and "magnet." Among these epochs, we selected those that showed evolving or persistent epileptiform discharges (EPED) that were not terminated by 5 dual-burst therapies delivered per the standard initial protocol. We used EPED as surrogate markers for clinical or electrographic seizures. We defined two most recent 28-day evaluation periods, one for DFB and another for LFLB, separated by at least 14 days from any programming changes. For these periods, we manually counted the number of EPED. We compared the 2 periods with respect to changes in EPED and charge density (CD), as the primary and secondary outcome measures respectively, using paired t test. Patient-reported seizures were not assessed.

Results: Out of 37 patients in our practice, 7 patients (4 males, 3 females; age at implant 19-43 years) were switched from DFB to LFLB (Table 1). All had interictal or ictal high frequency oscillations (HFOs: ≥70 Hz) on intracranial recordings. 5 had gamma activity in the implanted RNS leads, placed uni-/bilaterally in various brain regions. Median follow-up period was 36 months (range, 24-59) for DFB and 13 months (range, 2-57) for LFLB (p=0.194) (Table 2). DFB stimulations were at 200 Hz, 100-210 ms; LFLB stimulations were at 1.5-50 Hz, 200-5000 ms. Number of EPED was significantly lower during the LFLB than DFB evaluation period (mean, 5.4 vs. 13.1, p=0.009, 95% confidence interval 2.7 to 12.7). CD was higher with LFLB than DFB (mean, 3.5 vs. 3.2 μC/cm², p=0.068). There was no statistically significant correlation between EPED and CD (rho=0.17, p=0.72) or EPED and follow-up length (rho=0.29, p=0.53) during LFLB stimulations. Pulse width remained constant at 160 μs during both time periods. No adverse effects were seen after switching to the LFLB paradigm.

Conclusions: This small preliminary study of RNS, assessing objectively recorded data, shows that LFLB (≤50 Hz/≥200 ms) stimulation can be an effective and beneficial alternative paradigm for patients who fail to respond optimally to the recommended default 200 Hz/100 ms stimulation; this benefit seems to occur regardless of changes in CD. It is possible that HFO is a predictor of success with the LFLB paradigm. However, the effect of changes in antiepileptic drugs (AEDs) was not specifically assessed, which is a limitation of our study. Larger, prospective studies that control for changes in AEDs are needed to further understand the benefits of LFLB stimulation in patients implanted with RNS.

Funding: None