Portable Transcranial Electrical Stimulation and Internet-Based Behavioral Therapy for Major Depression Study

Major depressive disorder (MDD) is a prevalent, debilitating and chronic mental disorder, characterized by frequent recurrences and resistance to first-line treatments: antidepressants and onsite cognitive-behavioral therapy (CBT) are associated with discontinuation due to side-effects, and logistical burdens, respectively.

Non-invasive brain stimulation techniques (NIBS), such as repetitive transcranial magnetic stimulation (rTMS) and transcranial electrical stimulation (tES) (of which the most widely studied format is transcranial direct current stimulation - tDCS), are considered safe and tolerable interventions for MDD. While rTMS is already approved by the FDA for MDD, it involves daily visits to external facilities and is costly.

tES, on the other hand, albeit having failed to prove non-inferiority to a conventional antidepressant (escitalopram 20mg/day) in a previous trial, is less expensive than rTMS, and portable. Therefore, tES could be suitable for home-use under remote supervision by trained clinical staff, especially as an add-on intervention for MDD. In this scenario, recent pilot trials have suggested a possible combined effect of tES with psychological interventions, based on a mechanism known as "functional targeting", that is, using two or more interventions to engage the same brain area of interest.

The PSYLECT study is a parallel, 3-arm, randomized, double-blind, sham-controlled, single-center clinical trial, with digital features, devised to assess the efficacy, tolerability, safety and usability of: (1) active portable tES (ptES) + active internet-based behavioral therapy (iBT) ("double active"), (2) active ptES + sham iBT ("ptES-only"), and (3) sham ptES + sham iBT ("double-sham"), for the treatment of MDD, in adults, during 6 weeks.

The investigators will enroll 210 adult patients (70 per arm) diagnosed with MDD per DSM-5 criteria, regardless of being in treatment with conventional antidepressants, to receive either: "double-active", "ptES-only" or "double-sham", during 6 weeks.

The use of sham (placebo) is necessary in order to better estimate the combined active treatment effect, but patients will be invited to receive the active treatments at endpoint, in an open-label 6-week, crossover phase, if they were previously randomized to receive "double-sham", and did not display response at endpoint (defined ≥ 50% reduction in depressive symptoms). Furthermore, patients who were randomized to "double-active" or "ptES-only" and display response at endpoint, will be invited to participate in a maximum 6-month open-label follow-up phase (including those patients that respond at the end of the cross-over period). However, participants who were randomized to "ptES-only", but do not achieve response, will not be eligible for open-label follow-up, as it is unlikely they would profit from a continuation.

The study has received approval by the Ethics Committee of the University of Sao Paulo, in accordance with the Declaration of Helsinki guidelines. Prior to enrollment, participants and researchers will receive and jointly sign an informed consent document.

In this study, 3 co-primary hypotheses are presented: (1) changes in depression scores in the Hamilton Depression Rating Scale 17-item version (HDRS-17), from baseline to endpoint, will be larger in the "double active" compared to the "ptES-only" arm; (2) changes in depression scores (HDRS-17), from baseline to endpoint, will be larger in the "double active" compared to the "double-sham" arm; and (3) changes in depression scores (HDRS-17), from baseline to endpoint, will be larger in the "ptES-only" as compared to the "double-sham" arm.

The secondary hypotheses are that: (1) changes in depression scores will be larger in "double-active" compared to "ptES-only", in "double-active" compared to "double-sham", and "ptES-only" compared to "double-sham" arms, using additional depression-rating scales; (2) response (defined as ≥ 50% reduction in HDRS-17) and remission (defined as HDRS-17 ≤ 7) will be larger in "double active" versus "ptES-only", "double-active" versus "double-sham", and "ptES-only" versus "double sham" arms; (3) reduction in anxiety symptom scores will be greater in "double-active" compared "ptES-only", "double-active" compared to "double-sham", and "ptES-only" compared to "double-sham"; (4) the clinical usability of the "double active" protocol will be regarded as (very) easy by ≥ 80% of recipients, according to a Likert scale; (5) all three protocols will be equally safe and tolerable, according to the tDCS Adverse Events Survey.

Prior to performing online home-based sessions, participants will receive proper onsite training with the ptES device and iBT app, and from then on, will have remote access to the investigators through a phone line and email for questions and management of adverse events. In this case, patients can be redirected to the research institution's clinical facilities (including its emergency departments), in case of clinical and/or psychiatric adverse events which are deemed to need onsite evaluation and follow-up. The aforementioned clinical facilities are situated at the Clinics and University Hospitals of the University of Sao Paulo, Brazil, which are, respectively, tertiary and secondary care centers, and reference public hospitals in the Sao Paulo city area.

The ptES devices used in this trial are manufactured by Flow Neuroscience (Malmö, Sweden), and consist of a one-size-fits-all headset with circular electrodes, with the anode positioned over the left prefrontal cortex and the cathode over the right prefrontal cortex. The device comes with a set of 21 sealed, saline-humidified disposable sponges, an electrical charger, and has pre-set and fixed stimulation parameters. This device has been previously tested by the investigators in a computer-modeling procedure, with simulations displaying adequate electrical field strength and distribution in prefrontal brain areas of interest (ie, dorsolateral prefrontal cortex). Furthermore, the Flow device has been approved in the United Kingdom, European Union and Brazil (National Agency of Sanitary Vigilance, ANVISA), for home-use in patients with MDD.

The investigators' team will consist of 2 separate groups: (1) unblinded "session supervisors", who will offer an initial onsite training session to participants and will be available remotely for supervised guidance for the following online home-based sessions; (2) blinded clinical evaluators, who will evaluate participants online weekly until study endpoint, while being able to refer patients to onsite evaluation, as deemed necessary.

For data collection, the investigators will use REDCap (Research Electronic Data Capture), created at Vanderbilt University (Tennessee, USA) with NIH support. REDCap is a free-source software, that works in offline and online modes and runs on tablets, smartphones and computers, and was designed to build and manage online surveys and databases, replacing paper notebooks. It also has tools that allow dynamic data management and data quality examination. REDCap can be installed in a variety of environments for compliance with such standards as HIPAA, 21 CFR Part 11, FISMA (low, moderate, high), Brazlian regulations (LGPD) and international standards.

Periodically, the data will be exported and saved offline in password-protected external HDs and flash drives. REDCap access will be restricted using individual usernames and passwords. No user will have permission to modify or export the data (except for generating security copies) during the trial. If data edition is necessary (e.g., incorrect input), a new data version will be generated. All changes in the data will be reported.

For data management and quality control, the investigators hired an information technology (IT) specialist for technical support in the project. Data will be coded according to a previously developed data dictionary. Quality of data collection will be monitored by random data quality checks for consistency (e.g., depression scores compatible in different scales) and completeness (absence or few cases of missing data). The investigators will also have access to an online dashboard, that remotely monitors participant adherence to study protocol.

If unblinding is necessary, the PI holds special codes for the procedure, while having to specify the reason for unblinding, which are: an increase ≥ 25% from baseline scores on the Hamilton Depression Rating Scale-HDRS (17-item version) on 2 consecutive clinical evaluations, development of suicidal ideation, serious clinical and psychiatric adverse events, development of hypo(mania) and/or psychotic symptoms, low adherence (missing 2 clinical evaluations or failing to comply with at least 75% of tDCS sessions at home) or patient consent withdrawal.

For the sample size calculations, baseline depression scores and standard deviation (SD) of 25 (±5) on the HDRS-17 were used, distributed equally between groups, based on previous studies. Furthermore, the investigators consider that: (1) placebo effects in the double-sham arm will impact baseline depression scores equal to 1 effect size (ES) in SD units (ES = difference in mean change divided by SD); (2) the "ptES-only" arm will have the same placebo effects present in the double-sham arm (ES = 1), plus a treatment response of ES = 0.4; and, (3) the "double-active" arm will have placebo effects (ES = 1) plus a larger response of ES = 0.8. Therefore, probable endpoint scores of 20, 18 and 16 were used, for "double-sham", "ptES-only" and "double-active", respectively. The effect sizes were also chosen based on the results from previous trials, and considering that tDCS and online behavioral therapies have small to moderate effect sizes.

Sample size per arm was calculated based on the smallest detectable group difference (ES tDCS vs. combination = -0.4, ES sham vs. tDCS = -0.4). Significance levels were Bonferroni corrected for 3-way pairwise comparisons (α = 0.05/3) to control family-wise error rate, while still allowing for more anti-conservative adjustment in the final analysis. Dropouts were assumed to be monotonically increasing (Weibull) and equally distributed between treatment arms. The investigators obtained a total n = 210 participants, assuming a 10% attrition rate (therefore, 70 persons per arm).

Results will be considered statistically significant if a two-sided p<0.0167 (Bonferroni-corrected) is obtained. For the analysis of the co-primary outcomes, a linear mixed-effect model (LLM) will be employed, using a first order regressive covariance structure, which includes all observed variables without the need of imputing missing data. The dependent variable is score change on the HDRS-17. Independent variables are time (all observations until week 6), and group ("double-active", "ptES-only" and "double-sham"). The investigators will test the statistical significance between the pairwise comparisons per our primary hypotheses. The investigators will employ an intention-to-treat (ITT) approach. For the analysis of the secondary outcomes, linear mixed-effects models will be employed analogically to the analysis for the primary outcomes. Binary outcomes (response and remission) will be modeled using mixed logistic regression at each time-point. Improvement in other depressive domains will be evaluated using the same linear hierarchical models described.