Informing Further Research in the Use of Brain Stimulation in Psychiatric Disorders: Response to Syed and Smith

To the Editor: We thank Drs. Syed and Smith for their interest in our study (1). Our primary outcome measure was clinician-reported symptoms of posttraumatic stress disorder (PTSD) at 2 weeks, which showed no effect of active intermittent theta-burst stimulation (iTBS), a novel form of transcranial magnetic stimulation (TMS), as stated in our article and presented in Table 1. Here, we want to highlight the importance of effect sizes, which appear throughout the article to permit readers to judge the potential clinical relevance of our findings, as consistent with current opinion in the field (2) and correctly emphasized by Dr. Smith in a prior letter to the editor (3). To this end, we found clinically meaningful effect sizes in social and occupational function as well as self-reported PTSD and depressive symptoms at 2 weeks, with further meaningful effects across later observations. We did not employ correction for multiple comparisons because analyses were preplanned, and correction would not change observed effect sizes.

The letter brings up an underlying question we would like to address: As we designed the study, did we know how many weeks of stimulation would be sufficient to observe clinical effects? Only with the benefit of hindsight. When we designed the study, the most relevant evidence was work using 2 weeks of sham-controlled TBS (4). As such, when designing this first study of iTBS for PTSD, we felt that putting our primary outcome at 2 weeks was appropriately conservative. We then incorporated further stimulation exposure and neuroimaging to inform future research, with additional contributions from data on 1-year clinical and imaging outcomes (5).

Regarding the query about whether the majority of effect from active iTBS was observed at 1 week during the double-blind phase: the mean baseline scores on the PTSD Checklist for DSM-5 (PCL) were 49 (SD=9) in the active group and 50 (SD=11) in the sham group. Scores in the active group were 36 (SD=14) at week 1 and 36 (SD=14) at week 2. This was in contrast to the sham group PCL scores of 42 (SD=16) and 39 (SD=17) at weeks 1 and 2, respectively. In our opinion, further testing of these results would represent an unjustified stretching of the data.

We appreciate these readers’ interest in this first study of iTBS in PTSD that incorporated a sham-controlled design, dose-finding elements, and functional neuroimaging. In our article, we stated that while our results require replication, they were promising, and that further research needs to be done in this area. We stand by that statement.