One Year Clinical Outcomes after Transcranial Direct Current Stimulation and Virtual Reality for Posttraumatic Stress Disorder

Dear Editor

Posttraumatic stress disorder (PTSD) is a prevalent and disabling psychiatric disorder (1). Unfortunately, current treatment options have modest efficacy and challenges related to accessibility, burden, and side effects (2). Therefore, we and others have been developing non-invasive brain stimulation for PTSD, now with evidence that transcranial magnetic stimulation (TMS) can reduce PTSD symptoms in controlled trials and naturalistic settings (e.g., (3–5) but see also (6)). Of note, TMS likely indirectly modulates the core neurocircuitry of PTSD, which consistently implicates the relationship between the ventromedial prefrontal cortex (VMPFC) and amygdala (7) among other regions. For that reason, we recently demonstrated that VMPFC-targeted transcranial direct current stimulation (tDCS) plus virtual reality (VR) for PTSD could effectively reduce PTSD symptoms (8). This combination was designed to leverage the use of tDCS as a subthreshold perturbation, delivering weak electrical currents to bias neurons to depolarize in a therapeutic exposure-based VR context. In a randomized, sham-controlled trial of six sessions of tDCS+VR over 2–3 weeks, we demonstrated that active stimulation improved self-reported PTSD at one-month (d=−.82, p=.02) and accelerated psychophysiological habituation to VR events (p<.001). Follow up visits at 3-months indicated continued improvements over time in the active group, which included superior improvement in social and occupational function (d=1.2, p=.006) and clinically meaningful (but not statistically significant likely due to attrition) effect sizes in self-reported PTSD reduction (d= −.88, p=.07). These findings indicated increased benefit over time in the active group, yet whether these findings are durable remains unknown.

Here we examined naturalistic 1-year clinical outcomes, hypothesizing that active tDCS+VR would yield continued superior outcomes. The VA Providence IRB approved all procedures; full information of the parent clinical trial can be found in (8; NCT03372460). VA electronic medical records were reviewed to evaluate naturalistic outcomes from the 1-month primary outcome timepoint. Assignment followed randomization from the parent study (n=26 in the active group and n=28 in sham) with balanced groups (see Supplemental Information, Supplemental Table 1 and (8)).

The primary outcome for this study was psychiatric relapse, operationally defined as any suicide attempt, inpatient treatment, PTSD residential program admission, rescue TMS, or psychiatric emergency/urgent care visit for PTSD or suicidal ideation. This definition was identical to our prior examination of long-term outcomes following intermittent theta burst stimulation (iTBS) for PTSD (9). Kaplan-Meyer survival curves and odds ratios were calculated for relapse within one year to compare active to sham. Medication use was evaluated to determine if a psychiatric regimen had increased, decreased, or remained unchanged, examined via chi square tests. Non-psychiatric medications were not included, although medications with clear dual use (e.g., prazosin) were included. Due to the varied changes, statistical evaluation of individual medications or classes was not performed. Lastly, we examined whether reduction in psychophysiology arousal to VR events (i.e., successful habituation) predicted survival, via Spearman correlation. Analyses were performed in SPSS (v20, IBM, Armonk, NY)

In the one-year period, there were no deaths from suicide (or other causes). Overall, n=17 (31%) of patients relapsed in one year. The majority of these (n=13, 76.5%) were in the sham+VR group, and four (23.5%) were in the active group. Survival curves demonstrated superiority of active tDCS+VR: survival time in the active group was 315±22 days versus 231±28 days in the sham+VR group (log rank chi-square=5.59, df=1, p=.019; Figure 1). Odds ratios demonstrated that participants randomized to sham were more likely to relapse (OR= 3.02; 95% CI 1.1–8.1). The average time to relapse was 45 days (SD 11.1) in the active tDCS+VR group and 76.9 days (SD 58.4) in the sham+VR group, which did not significantly differ (t=−1.86, p=0.08). Post hoc exploration did not indicate that any one component of relapse drove observed results. Medication changes favored the active group (chi-square=8.62, df=2, p=.013). In the active group (n=26), 18 (69.2%) exhibited no changes, 10 (26.9%) had reductions, and one (3.8%) had an increase. In the sham group (n=28), 14 (50%) had no changes, four (14.2%) had reduced medications, and 10 (35.7%) had increased pharmacology (Supplemental Table 1). Due to the low 1-year relapse rate in the active group (n=4), only the sham group had sufficient power to examine the relationship between habituation and relapse, and in that group successful habituation correlated positively with longer survival (Spearman’s rho=0.58, p=.04, n=13).

Figure 1.

Survival curve analyses comparing active tDCS+VR versus sham tDCS+VR for PTSD over one year

Figure Legend: Relapse was operationally defined following prior work (9), as any instance of suicide attempt, inpatient treatment or PTSD residential program admission, rescue TMS, or emergency department or urgent care visit for PTSD symptoms or suicidal ideation (log rank chi-square=5.59, df=1, p=.019, favoring active tDCS+VR vs sham).

Key: tDCS, transcranial direct current stimulation; VR, virtual reality.

This 1-year naturalistic study of Veterans with PTSD demonstrated a benefit of active tDCS+VR versus sham+VR. Outcomes were superior across multiple domains, as the active group had a lower rate of psychiatric relapse, and their psychiatric medications were more often reduced. The majority of relapse in the active tDCS group appeared within the first two months, which is consistent with the timeframe of consolidation-based neuroplasticity. Early changes in psychophysiological arousal were also predictive of longer-term outcomes, at least in the sham group.

Of note, the parent study differs from a recent study of tDCS over the dorsolateral prefrontal cortex combined with script-driven exposure for PTSD (10), which did not find separation between active and sham; the primary differences between these studies was our use of tDCS administered simultaneously with VR and targeting the VMPFC. Parenthetically, relapse rates in this study appeared lower than those observed in our comparable follow up of iTBS for PTSD (9), although direct comparisons should be the focus of future study.

Limitations of this work are those inherent to chart reviews. Symptoms were not systematically measured; some participants had rating scales as part of routine measurement-based care, but these were not evenly distributed. Our definition of relapse was based upon prior work yet are indicators of more severe illness and subtle changes may have been missed. It is also possible that other factors, such as substance use, may have factored into the clinical course. Several factors may also reduce generalizability; the sample skewed male and white, and were required to have warzone exposure, although many kinds of trauma (including extensive pre-military trauma) were highly represented (see supplement in [17]).

In closing, these findings provide important, real-world evidence that the brief use of tDCS+VR can provide acute and long-term benefits for patients with PTSD.