T h e NEWENGL AND JOUR NAL o f MEDICINE
ORIGINAL ARTICLE
Trial of Electrical Direct-Current Therapy
versus Escitalopram for Depression
A.R. Brunoni, A.H. Moffa, B. Sampaio Junior, L. Borrione, M.L. Moreno, R.A. Fernandes, B.P. Veronezi, B.S. Nogueira, L.V.M. Aparicio, L.B. Razza, R. Chamorro, L.C. Tort, R. Fraguas, P.A. Lotufo, W.F. Gattaz, F. Fregni, and I.M. Benseñor, for the ELECT-TDCS Investigators*
A BS T R AC T
BACKGROUND
We compared transcranial direct-current stimulation (tDCS) with a selective sero-tonin-reuptake inhibitor for the treatment of depression.
METHODS
In a single-center, double-blind, noninferiority trial involving adults with unipolar depression, we randomly assigned patients to receive tDCS plus oral placebo, sham tDCS plus escitalopram, or sham tDCS plus oral placebo. The tDCS was adminis-tered in 30-minute, 2-mA prefrontal stimulation sessions for 15 consecutive week-days, followed by 7 weekly treatments. Escitalopram was given at a dose of 10 mg per day for 3 weeks and 20 mg per day thereafter. The primary outcome measure was the change in the 17-item Hamilton Depression Rating Scale (HDRS-17) score (range, 0 to 52, with higher scores indicating more depression). Noninferiority of tDCS versus escitalopram was defined by a lower boundary of the confidence in-terval for the difference in the decreased score that was at least 50% of the differ-ence in the scores with placebo versus escitalopram.
RESULTS
A total of 245 patients underwent randomization, with 91 being assigned to esci-talopram, 94 to tDCS, and 60 to placebo. In the intention-to-treat analysis, the mean (±SD) decrease in the score from baseline was 11.3±6.5 points in the escitalopram group, 9.0±7.1 points in the tDCS group, and 5.8±7.9 points in the placebo group. The lower boundary of the confidence interval for the difference in the decrease for tDCS versus escitalopram (difference, −2.3 points; 95% confidence interval [CI], −4.3 to −0.4; P = 0.69) was lower than the noninferiority margin of −2.75 (50% of placebo minus escitalopram), so noninferiority could not be claimed. Escitalo-pram and tDCS were both superior to placebo (difference vs. placebo, 5.5 points [95% CI, 3.1 to 7.8; P<0.001] and 3.2 points [95% CI, 0.7 to 5.5; P = 0.01], respec-tively). Patients receiving tDCS had higher rates of skin redness, tinnitus, and nervousness than did those in the other two groups, and new-onset mania devel-oped in 2 patients in the tDCS group. Patients receiving escitalopram had more frequent sleepiness and obstipation than did those in the other two groups.
CONCLUSIONS
In a single-center trial, tDCS for the treatment of depression did not show non-inferiority to escitalopram over a 10-week period and was associated with more adverse events. (Funded by Fundação de Amparo à Pesquisa do Estado de São Paulo and others; ELECT-TDCS ClinicalTrials.gov number, NCT01894815.)
The authors’ full names, academic de-grees, and affiliations are listed in the Appendix. Address reprint requests to Dr. Brunoni at the Interdisciplinary Cen-ter for Applied Neuromodulation and Service of Interdisciplinary Neuromodu-lation, LIM-27, University of São Paulo, Av. Prof. Lineu Prestes, 2565, 3o andar, CEP 05508-000, São Paulo, Brazil, or at brunoni@usp.br.
* A complete list of the investigators in the Escitalopram versus Electrical Cur-rent Therapy for Treating Depression Clinical Study (ELECT-TDCS) is provided in the Supplementary Appendix, avail-able at NEJM.org.
N Engl J Med 2017;376:2523-33.
DOI: 10.1056/NEJMoa1612999
Copyright © 2017 Massachusetts Medical Society.
N ENGL J MED 376;26 NEJM.ORG JUNE 29, 2017 | 2523 |
The New England Journal of Medicine
Downloaded from nejm.org on June 28, 2017. For personal use only. No other uses without permission. Copyright © 2017 Massachusetts Medical Society. All rights reserved.
T h e NEWENGL A ND JOUR NA L o f MEDICINE
MAJOR DEPRESSIVE DISORDER IS A HIGH-ly prevalent condition.1 There is interest in the effectiveness and safety of new
and nonpharmacologic treatments for depression. In 2009, transcranial magnetic stimulation was approved by the Food and Drug Administration for the treatment of major depressive disorder.2 The procedure has had mixed results in various trials,3 is associated with a small risk of seizure,4 and is costly.
Transcranial direct-current stimulation (tDCS) is a noninvasive brain-stimulation technique that is less costly than transcranial magnetic stimula-tion and has not been associated with seizures.5 In this procedure, weak, direct current is applied through electrodes that are placed on the scalp to induce alterations in cortical activity and excit-ability.6 In patients with major depressive disor-der, tDCS-induced currents are applied to the dorsolateral prefrontal cortex, which is consid-ered to be a target for mood regulation.7
We previously conducted a trial involving pa-tients with major depressive disorder,8 which showed the superiority of tDCS plus sertraline over tDCS only, sertraline only, and placebo. However, that trial was not designed to compare tDCS with pharmacotherapy directly. Further-more, small, placebo-controlled clinical trials testing the efficacy of tDCS in patients with major depressive disorder have shown inconsis-tent results.9
On the basis of the safety of tDCS that has been observed in previous studies, and given the possibility that tDCS would be more acceptable to patients than antidepressants, we conducted a noninferiority trial, the Escitalopram versus Elec-trical Current Therapy for Treating Depression Clinical Study (ELECT-TDCS), to compare the efficacy of tDCS with that of the selective sero-tonin-reuptake inhibitor escitalopram in patients with major depressive disorder. The null hypoth-esis was that the decrease in the score (indicat-ing less depression) on a conventional rating scale of depression in the tDCS group would be 50% or less of the difference in the decreased scores between the escitalopram group and the placebo group; that is, noninferiority would be established if tDCS, as compared with escitalo-pram, was associated with at least 50% of the superiority of escitalopram over placebo. As sec-ondary aims, we investigated biomarkers that are
associated with clinical depression outcomes. We also investigated the effects of tDCS versus placebo and tDCS versus escitalopram over time and assessed the adverse effects and safety of both tDCS and escitalopram.
ME THODS
TRIAL DESIGN
The trial design has been published previously.10 The trial was conducted at the University Hospi-tal and Department and Institute of Psychiatry, University of São Paulo, with a recruitment pe-riod from October 2013 through July 2016. The trial was approved by the local ethics committee. All the patients provided written informed con-sent. All the authors affirm that the trial was conducted, and all analyses were performed, per the original protocol, which is available with the full text of this article at NEJM.org. The authors vouch for the accuracy and completeness of the data and analyses reported.
Soterix Medical provided five tDCS devices (1×1 tDCS-CT) free of charge, and Libbs provided escitalopram oxalate (Reconter, 10-mg pills) free of charge. These companies were not involved in any aspect of the conduct, analysis, or reporting of the trial. The funder, Fundação de Amparo à Pesquisa do Estado de São Paulo, had no role in any aspect of the trial.
In this noninferiority, parallel, placebo-con-trolled trial, patients were randomly assigned in a 2:3:3 ratio, with the use of a permuted-block design, according to a computer-generated list, to receive one of three regimens: sham tDCS plus placebo (placebo group), sham tDCS plus escital opram (escitalopram group), and active tDCS plus placebo (tDCS group).
PATIENTS
We included patients 18 to 75 years of age who had unipolar depression that had been diagnosed according to Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5), criteria and confirmed by psychiatrists by means of the Mini-International Neuropsychiatric Interview (MINI).11 Patients had to have a score of 17 points or more on the 17-item Hamilton Depression Rating Scale (HDRS-17; scores range from 0 to 52, with higher scores indicating more depression; a score of 24 or more indicates severe depression; minimal
2524N ENGL J MED 376;26 NEJM.ORG JUNE 29, 2017
The New England Journal of Medicine
Downloaded from nejm.org on June 28, 2017. For personal use only. No other uses without permission. Copyright © 2017 Massachusetts Medical Society. All rights reserved.
ELECTRICAL DIRECT-CURRENT THERAPY VS. ESCITALOPRAM
clinically significant difference, 3 points) as well as a low risk of suicide (evaluated with the use of the MINI). The exclusion criteria were bipolar disorder, substance abuse or dependence, demen-tia, personality disorder, brain injury, pregnancy, specific contraindications to tDCS (e.g., cranial plates), current or previous escitalopram use, and previous or concomitant participation in other trials of tDCS. Patients who presented with an anxiety disorder (generalized anxiety disorder, specific phobia, panic disorder, or social anxiety disorder) as a coexisting condition were not ex-cluded.
Participants were recruited by means of ad-vertisements and physician referrals and were prescreened by means of telephone and e-mail. Persons who met the inclusion criteria underwent on-site screening. Before the onset of the trial, the patients either were not using antidepres-sants or underwent a drug washout and re-mained free of antidepressant medications for five or more drug half-lives. Benzodiazepines were allowed but were tapered to a maximum dose of 20 mg per day of a diazepam equivalent; the dose remained stable during the trial.
INTERVENTIONS
Anode and cathode electrodes were placed over the left and right dorsolateral prefrontal cortexes, respectively, with the use of the Omni-Lateral-Electrode system.12 In a total of 22 sessions that lasted 30 minutes per day, 2 mA of direct-current stimulation were administered in each session. The first 15 sessions took place daily, except for weekends, and the remaining 7 sessions took place once a week, until week 10. The tDCS pro-tocol that was used in this trial had more ses-sions than in earlier trials because recent studies have suggested that more sessions could pro-duce greater clinical effects.13,14
Trained nurses administered the tDCS regi-men. The same protocol was used for active and sham tDCS, but the current was turned off auto-matically after 30 seconds in patients receiving sham tDCS by devices that were programmed to deliver active or sham stimulation according to the randomized code.
Patients received 10 mg per day of escitalo-pram (or matching placebo) for the first 3 weeks and 20 mg per day thereafter. The School of Pharmaceutical Sciences of the University of São
Paulo produced the placebo pills. They had the same size, color, appearance, and taste as the escitalopram pills and were stored in identical bottles. Escitalopram was chosen as representative of a first-line therapy for major depressive dis-order,15 with the maximally effective dose (20 mg per day) administered close to the initial dose (10 mg per day), which thus allowed the maxi-mum dose to be reached within weeks after the initiation of the intervention if necessary.
To assess the integrity of trial-group blinding, patients were asked to guess which intervention they had received and to rate the confidence in their prediction. Adherence to the escitalopram and placebo regimens was determined by means of pill count and was considered to be acceptable if less than 10% of the pills were returned.
OUTCOMES
All the assessments were performed by trained psychiatrists and psychologists who were unaware of the trial-group assignments. Efficacy and safety were measured during screening, at base-line, and at the end of weeks 3, 6, 8, and 10. The primary outcome was the change in the HDRS-17 score from baseline to 10 weeks.
Secondary outcomes included the changes from baseline in the Montgomery–Åsberg De-pression Rating Scale (MADRS) score (range, 0 to 60, with higher scores indicating more severe depression; minimal clinically significant differ-ence, 1.6 to 1.9 points),16 the Beck Depression Inventory score (range, 0 to 63, with higher scores indicating more severe depression; mini-mal clinically significant difference, 5 points),17,18 and several other scales that are listed in the protocol, the Supplementary Appendix (available at NEJM.org), and previous publications.19,20 Ad-ditional secondary outcomes were early improve-ment (defined as a change in the score from baseline to week 3 on the HDRS-17),21 clinical response (defined as a >50% reduction from the baseline HDRS-17 or MADRS score), and remis-sion (defined as an HDRS-17 score ≤7 or a MADRS score ≤10) at week 10.
ADVERSE EVENTS
Adverse events were assessed with the use of the Systematic Assessment for Treatment Emergent Effects questionnaire22 and a commonly used questionnaire regarding 39 adverse events that
N ENGL J MED 376;26 NEJM.ORG JUNE 29, 2017 | 2525 |
The New England Journal of Medicine
Downloaded from nejm.org on June 28, 2017. For personal use only. No other uses without permission. Copyright © 2017 Massachusetts Medical Society. All rights reserved.
T h e NEWENGL A ND JOUR NA L o f MEDICINE
have been associated with tDCS.23 At the end of weeks 3 and 10, patients were asked to fill out these questionnaires, describing the presence of an adverse event, its severity, and their opinion regarding its relationship to the trial regimen. We report data on all mild, moderate, and severe adverse events that were considered by the pa-tients to be at least remotely associated with the intervention.
The Young Mania Rating Scale (range, 0 to 60, with higher scores indicating a greater degree of manic features; minimal score to define new-onset mania or hypomania, 8 points) was used at the end of weeks 3 and 10 to assess mania or hypomania during the trial.24 Other events that were considered by the investigators to be poten-tially serious adverse events were hospitalization for a psychiatric cause, suicidality or attempted suicide, or events leading to major incapacity or a life-threatening condition. A brief neuropsy-chological evaluation consisting of seven tests (listed in the trial-design publication10 and Table S12 in the Supplementary Appendix) was per-formed at baseline and at 10 weeks to determine whether tDCS was associated with cognitive im-pairment.
BIOLOGIC MARKERS
Several biologic markers were investigated as predictors and mediators of clinical response. Two of these biologic markers have been ana-lyzed so far: heart-rate variability and motor corti-cal excitability (see the Supplementary Appendix). Other planned analyses have not yet been done.
STATISTICAL ANALYSIS
The sample size was estimated on the basis of results from our previous study, the Sertraline versus Electrical Current Therapy for Treating Depression Clinical Study (SELECT-TDCS),8 with the use of an attrition rate of 13%25 and a non-inferiority margin of 50% of the comparative efficacy of placebo versus escitalopram. The non-inferiority margin was based on our hypothesis that tDCS would be associated with at least 50% of the difference in efficacy of escitalopram as compared with placebo (see the Supplementary Appendix).26
For our primary hypothesis, we compared the decrease in the HDRS-17 score (the baseline score minus the score at 10 weeks, with a decrease in score indicating less depression) among the three
groups. A modified t-test was used to assess whether the difference in the scores between the tDCS group and the placebo group was more than 50% of the mean difference between the scores in the escitalopram group and the place-bo group.25 The noninferiority margin was based on the point estimate of 50% of the difference in the mean change in the HDRS-17 score in the comparison of placebo with escitalopram. This approach was chosen because our hypothesis was that tDCS would be associated with at least 50% of the effectiveness of escitalopram. If the lower boundary of the confidence interval around the mean difference between the scores in the tDCS group versus the escitalopram group exceeded this value, then noninferiority could be claimed (see the statistical analysis plan provided with the protocol). After the noninferiority assess-ment, t-tests were used for superiority analyses, comparing the decrease in the HDRS-17 score for escitalopram versus tDCS, for escitalopram versus placebo, and for tDCS versus placebo.
A mixed-model analysis of variance was con-ducted to assess a reduction in symptoms over time. Logistic regression was performed to as-sess the rates of response and remission between groups. The chi-square test or Fisher’s exact test was used to compare the frequency and severity of adverse events, new-onset mania or hypoma-nia, and serious adverse events between groups. The number of adverse events between groups at 10 weeks was compared with the use of the Kruskal–Wallis test. General linear models were used to assess predictors of response. These ex-ploratory analyses were not corrected for multi-ple comparisons (see the protocol).
For biomarker analyses, 78 analyses were per-formed, and 3 or 4 positive results were expected by chance (see the Supplementary Appendix). For the clinical, demographic, neuropsychological, psychological,27 and adverse-event analyses, 4 or 5 positive results were expected by chance, be-cause 85 analyses were performed.
We performed noninferiority analyses in the intention-to-treat and per-protocol populations.28 Analysis was also performed in the population of patients who had high adherence to the trial visits (patients who had <2 missing visits). Missing data were considered to be missing at random and were imputed with the use of regression models, in which baseline depression and main demo-graphic characteristics were used as variables.
2526N ENGL J MED 376;26 NEJM.ORG JUNE 29, 2017
The New England Journal of Medicine
Downloaded from nejm.org on June 28, 2017. For personal use only. No other uses without permission. Copyright © 2017 Massachusetts Medical Society. All rights reserved.
ELECTRICAL DIRECT-CURRENT THERAPY VS. ESCITALOPRAM
Table 1. Demographic and Clinical Characteristics of the Patients at Baseline.*
| Placebo | Escitalopram | tDCS |
Characteristic | (N = 60) | (N = 91) | (N = 94) |
Female sex — no. (%) | 41 (68) | 61 (67) | 64 (68) |
Age — yr | | | |
Current | 40.9±12.9 | 41.8±12.5 | 44.6±11.8 |
At onset of depression | 25.7±11.3 | 26.4±12.0 | 26.4±11.7 |
Type of depression — no. (%) | | | |
Recurrent | 44 (73) | 59 (65) | 59 (63) |
Chronic | 29 (48) | 46 (51) | 42 (45) |
Severe | 22 (37) | 25 (27) | 28 (30) |
Melancholic | 22 (37) | 37 (41) | 34 (36) |
Atypical | 18 (30) | 25 (27) | 30 (32) |
Any anxiety disorder — no. (%)† | 38 (63) | 46 (51) | 56 (60) |
Family history of psychiatric disorder — no. (%) | 39 (65) | 57 (63) | 64 (68) |
History of treatment for depression | | | |
No. of treatment failures | | | |
In current episode | 1.0±1.4 | 0.9±1.5 | 1.0±1.2 |
Over lifetime | 4.9±4.3 | 4.5±3.9 | 4.8±3.8 |
Treatment resistance — no. (%) | 19 (32) | 25 (27) | 30 (32) |
Current use of benzodiazepines — no. (%) | 17 (28) | 20 (22) | 31 (33) |
HDRS-17 score‡§ | 22.7±4.3 | 21.7±3.5 | 21.8±3.9 |
MADRS score§¶ | 28.1±6.8 | 26.2±6.0 | 27.4±7.0 |
Beck Depression Inventory score§‖ | 31.1±11.1 | 29.4±8.8 | 30.9±9.2 |
* Plus–minus values are means ±SD. No significant between-group differences were observed.
† Any anxiety disorder was defined as generalized anxiety disorder, specific phobias, social anxiety disorder, or panic disorder.
ØScores on the 17-item Hamilton Depression Rating Scale (HDRS-17) range from 0 to 52, with higher scores indicating more depression; a score of 24 or more indicates severe depression (minimal clinically significant difference, 3 points).
ØThe variables include depression characteristics, such as recurrence (>3 previous episodes), chronicity (current episode with ≥12-month duration), severity, and treatment resistance (≥1 treatment failure in the current episode or >4 treatment failures over the patient’s lifetime).
ØScores on the Montgomery–Åsberg Depression Rating Scale (MADRS) range from 0 to 60, with higher scores indicating more severe depression (minimal clinically significant difference, 1.6 to 1.9 points).
ØScores on the Beck Depression Inventory range from 0 to 63, with higher scores indicating more severe depression (minimal clinically significant difference, 5 points).
R E SULT S
PATIENTS
Of 1479 patients screened, 245 were enrolled and underwent randomization. A total of 60 patients were assigned to receive placebo, 91 to receive escitalopram, and 94 to receive tDCS. Of these 245 patients, 202 received all 22 planned ses-sions of actual or sham tDCS and completed the week-10 assessment (55 patients in the placebo group, 75 in the escitalopram group, and 72 in the tDCS group) (Fig. S1 in the Supplementary Appendix). Withdrawal rates did not differ sig-
nificantly among the three groups (χ2 = 4.77, P = 0.09). The reasons for withdrawal are de-scribed in Table S1 in the Supplementary Ap-pendix. The characteristics of the patients are described in Table 1, and in Tables S2, S12, and S14 in the Supplementary Appendix.
PRIMARY OUTCOME
As compared with baseline, the mean depression scores, as measured by the HDRS-17, decreased (with greater decreases indicating less depres-sion) by 11.3±6.5 points in the escitalopram group, by 9.0±7.1 points in the tDCS group, and
N ENGL J MED 376;26 NEJM.ORG JUNE 29, 2017 | 2527 |
The New England Journal of Medicine
Downloaded from nejm.org on June 28, 2017. For personal use only. No other uses without permission. Copyright © 2017 Massachusetts Medical Society. All rights reserved.