Abstracts

Rationale: Low frequency repetitive transcranial magnetic stimulation (rTMS) is emerging as a novel noninvasive therapeutic tool in epilepsy. In recent years, several open label trials have shown an encouraging reduction in seizure frequency in patients with epilepsy. However, the data from controlled trials are mixed with respect to antiepileptic rTMS efficacy, and would benefit from replication in further carefully-controlled trials. Towards this goal, an important experimental design accept is the choice of a suitable control intervention. Sham rTMS is possible with specially designed stimulation coils, but complete blinding of patients is difficult, particularly when multiple sessions of stimulation are applied. To facilitate future controlled rTMS clinical trials, we performed the present metanalysis to estimate the placebo effect of rTMS. Methods: Individual subject outcome data from three published placebo-controlled rTMS trials (Theodore et al., 2002, Fregni et al., 2006 and Cantello et al., 2007) were systematically analyzed to calculate median reduction in seizure frequency. The data were then pooled to estimate an overall median change in seizure frequency at follow-up intervals of 2, 4 and 8 weeks after placebo rTMS treatment. Results: Three mehods for sham rTMS were employed: (1) coil positioning orthogonal to the scalp (Theodore et al., 2002), (2) a spring-loaded sham coil (Fregni et al., 2006), and (3) a double active-sham coil (Cantello et al., 2007). Although three individual methods for sham rTMS were employed, the placebo response was overall similar across studies (Kruskal-Wallis p > 0.2). Overall median change in seizure frequency was 2% at 2 weeks (n = 63; 95% CI -4.3 to 19.3) 0% at 4 weeks (n=63; 95% CI -12.5 to 10.4) and 0% at 8 weeks 8 (n = 20; 95% CI -30.2 to 13.5). Conclusions: We find a small consistent rTMS placebo effect as indexed by median seizure frequency is identified at 2, 4 and 8 weeks following stimulation. This approximation may be factored into future rTMS study design, particularly when calculation of sample size is required. Our observation that seizure frequency change in similar across distinct placebo rTMS methods suggests that the placebo effect on seizure frequency may not depend on the placebo coil design, but may instead relate to other elements of the rTMS procedure. Interestingly, placebo effects associated with rTMS may be dependent on the condition being studied as a recent meta-analysis showed that placebo rTMS in major depression has a large effect size (Brunoni, Fregni, 2009).