Bilateral cerebellar high-frequency rTMS for refractory epilepsy: a randomized, single-blind, sham-controlled study

Background For patients with drug-resistant epilepsy who are not suitable for surgery, transcranial magnetic stimulation (TMS) is an important non-invasive neuromodulation therapy. The cerebellar cortex can inhibit cortical excitability via the cerebello-thalamo-cortical pathway. Here, we explore whether high-frequency (20 Hz) cerebellar repetitive TMS (rTMS) reduces seizure frequency and interictal epileptiform discharges (IEDs) in refractory epilepsy. Methods Eighty-seven patients with refractory epilepsy were included in the intention-to-treat (ITT) population ( n = 87) and then randomized to active rTMS ( n = 44) or sham stimulation ( n = 43). One patient in treatment group had no post-baseline efficacy data; efficacy analyses were therefore performed in the modified intention-to-treat (mITT) population ( n = 86). The active group received two cycles of 20 Hz rTMS targeting the cerebellum. Seizure frequency and IEDs were assessed before and after treatment. Results Poisson generalized linear mixed models revealed a significant group×time interaction for seizure frequency [ F (2,236) = 162.65, p < 0.001]. The active group had significantly lower seizure frequencies at months 1, 2, and 3 (all p < 0.001), with the largest difference at month 2. Percentage reductions of seizure frequency in the active group were 64.6%, 73.2%, and 13.3%, respectively, compared with 3.9%, 0.6%, and −8.4% in the control group (all P < 0.001). Responder rates (≥50% reduction) in the active group were 90.7%, 95.3%, and 0% vs. 0% in controls ( p < 0.001 for months 1–2). Within-group analysis showed significant IED reduction after active rTMS ( p < 0.001) but no change in controls. Moreover, a significant group × baseline interaction ( p < 0.001) indicated greater suppression in patients with higher baseline IEDs. Exploratory subgroup analysis further confirmed this. All adverse events were mild, with no significant difference in the incidence between the two groups. Conclusion Active cerebellar rTMS was associated with reduced seizure frequency, improved responder rates, and suppressed IEDs in refractory epilepsy with no obvious discomfort. The optimal efficacy duration was approximately 2 months, with greater IED suppression in patients with higher baseline levels.