Alternating current cranial electrotherapy stimulation (CES) for depression

  • Protocol
  • Intervention



This is the protocol for a review and there is no abstract. The objectives are as follows:

To assess the effectiveness and safety of alternating current cranial electrotherapy stimulation (CES) compared with sham CES for acute depression.


Description of the condition

Unipolar depressive disorders constitute a category of mental illness characterized by disturbances in mood regulation, sleep and appetite biorhythms, interpersonal and occupational function, self care activities such as grooming and, in extreme cases, suicidal behaviors. The defining feature of depressive disorders is the persistent experience of emotional pain, which may encompass sadness, anxiety, and irritability. For some patients, the primary mood symptom is a profound blunting of emotion resulting in an inability to experience appropriate positive and negative emotions, with a consequent marked diminution of interest in activities previously associated with interest and pleasure. The mood disturbance is deemed out of proportion (in severity or duration, or both) to the circumstances that triggered the disorder. Inadequate coping skills (adaptive cognitive and behavioral responses to adversity) may be both a predisposing factor for depression and an expression of its occurrence (Akiskal 2009).

Unipolar depressive disorders are classified according to patterns of recurrence, chronicity, and severity. The two most widely used nosologic systems, the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revised (DSM-IV-TR; APA 2000), and the World Health Organization's International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Version 2007 (ICD-10; WHO 2007) both define two primary depressive disorders: major depression/depressive episode (which is the more acute and severe of the depressive disorders) and dysthymia (a chronic, low-grade depression).

In DSM-IV-TR, major depression is defined by the presence of significant depressed mood or diminished interest/pleasure in activities over at least two continuous weeks, together with four of the following additional symptoms resulting in functional impairment: change in appetite; insomnia/hypersomnia; objective psychomotor agitation/retardation; decreased energy; feelings of worthlessness or excessive guilt; diminished concentration or indecisiveness; and recurrent thoughts of death or suicidal ideation (APA 2000).

The corresponding ICD-10 diagnosis of acute depression encompasses mild, moderate, and severe depressive episode. The three core symptoms of all depressive episodes are depressed mood, loss of interest and pleasure, and lack of energy, which lead to increased fatiguability and reduced activity, persisting for at least two weeks. Additional symptoms may include impaired concentration, decreased self esteem and self confidence, ideas of guilt or worthlessness, bleak and pessimistic views of the future, ideas or acts of self harm and suicide, disturbed sleep (such as waking up several hours before the usual time), and diminished appetite. Grades of depression severity are distinguished by the number, severity, and types of symptoms present. Mild depressive episode is diagnosed if two of the three core symptoms occur, along with at least two of the other symptoms, and the severity of symptoms results in only mild impairment in usual function. In moderate depression, patients must have at least two of three core symptoms, plus three or four associated symptoms, and they must experience "considerable difficulty" in continuing with usual activities. Finally, in severe depression, all three core symptoms plus four of the additional symptoms, some to a severe degree, must occur, resulting in marked functional impairment (WHO 2007).

The incidence and prevalence of depression vary significantly across populations. In general, throughout the world, women are more likely to be diagnosed with depression than men. A review by the World Health Organization (WHO) conducted as part of the Global Burden of Disease (GBD) study estimated that the incidence of depressive episodes, defined using either ICD-10 or DSM-IV criteria, ranges from 2.6% in males in Southeast Asia to 7.2% among females in North America. The same review reported that prevalence ranges from 1.0% among Western Pacific males to 3.6% in North American females (Ustun 2004).

Depression exacts a heavy personal and societal cost. It has been estimated that, at the individual level, recurrent depressive illness results in the loss of 12 productive years of role functioning at work and home (Kopelowicz 2009). At the population level, depressive disorders comprise the fourth leading cause of disease burden globally, accounting for 4.4% of total disability-adjusted life-years (DALYs) and 12% of DALYs due to illness-related disability (Ustun 2004). They are projected to become the second leading cause of total DALYs by 2020 (Donohue 2007). The adverse economic effects of depression due to absenteeism, impaired work performance, and treatment costs are severe worldwide. In Europe, the cost was estimated at about EUR 118 billion in 2004 (Sobocki 2006) and in the United States was computed at USD 83 billion in 2000 (Donohue 2007).

While the pathophysiology of depression remains unclear, evidence indicates that both biological and psychosocial risk factors contribute to its occurrence (Rihmer 2009). Demographic risk factors associated with depression include female gender and age less than 45 years. Psychosocial and environmental factors associated with depression include: marital status and social support, lower socioeconomic status, urban residence, geography (northern latitudes associated with seasonal depression), and psychosocial stress (Rihmer 2009). Genetic factors are thought to account for 40% to 50% of the risk for depression (Lohoff 2010). Other biological risk factors include medical conditions such as cerebrovascular disease, cardiovascular disease, chronic pain, and dementia (Blazer 2005).

Description of the intervention

At present, the standard first-line therapies for depression include psychotherapy, antidepressant medications, or combination treatment using both modalities (APA 2010; NICE 2009). Electroconvulsive therapy (ECT) is considered a treatment of choice for severe, refractory depression, in patients with psychotic or catatonic features, and in persons with nutritional compromise or suicidality. While ECT has the highest reported rates of response and remission, the efficacy of this treatment must be weighed against the possibility of complications resulting from general anaesthesia use and the risk of adverse cognitive effects (APA 2010; NICE 2009).

The effectiveness of the most widely accepted therapies for depression, namely antidepressant medications and psychological therapies, is limited, leaving many patients with significant residual symptomatology and in need of alternative treatment options. Indeed, comprehensive meta-analyses of antidepressant trials indicate that only 50% to 60% of such trials indicate superiority of drug therapy over placebo (Pigott 2010). Also, some patients who may have benefited from antidepressant medications may experience adverse effects that lead to treatment discontinuation. At the same time, available evidence indicates that even the most rigorously evaluated psychotherapy protocols, such as cognitive-behavioral therapy, are effective in only about 60% of patients with depression (Mor 2009).

Cranial electrotherapy stimulation (CES) - also called 'cranial electrostimulation', 'electrosleep therapy', and 'electronarcosis' - is a non-pharmacological treatment in which low-intensity electrical stimulation is applied to the scalp. CES has been used for the treatment of depression, anxiety, and insomnia. It is distinguished from the other main form of low-intensity cranial electrical stimulation, transcranial direct current stimulation (tDCS), by the use of alternating current (AC) rather than direct current (DC) electricity. Available evidence indicates that these differences in electrical stimulation result in significant differences in biological effects (Zaghi 2010). For instance, in tDCS, the application of a unidirectional current between two scalp electrodes (from the anode to the cathode) results in polarization of brain electrical activity, with acutely increased brain electrical activity under the anode and suppressed brain electrical activity under the cathode (Rosa 2012). In contrast, the electrical effects of CES on brain activity are thought to be consistent across the area of stimulation. Furthermore, the neurochemical systems responsible for mediating the after-effects of tDCS are different from those implicated in CES effects; the NMDA (N-Methyl-D-Aspartate) neurotransmitter system is centrally involved in mediating tDCS effects, while, as noted below (see How the intervention might work), other chemical mediators such as norepinephrine, serotonin, and GABA ((γ-aminobutyric acid) have been implicated in CES effects (Bystritsky 2008; Zaghi 2010).

Initially developed and investigated in the early 1900s, preliminary studies indicated that such electrical stimulation produced sedation, and it was thus initially called 'electrosleep' or 'electronarcosis' (Gilula 2005; Klawansky 1995). Much of the early clinical work using CES was conducted in Eastern Europe and the Soviet Union in the 1950s, but because these clinical trials were largely uncontrolled or otherwise of poor quality, the clinical efficacy of CES remained uncertain (Klawansky 1995). A number of additional clinical trials of CES in the treatment of depression, anxiety, substance abuse, and other conditions have been completed since the introduction of CES to the United States and Western Europe in the 1960s. However, since these later studies are likewise of variable quality, the utility of CES in depression and other disorders continues to be unclear (Rosa 2012).

CES therapy is self administered using battery-powered electrical devices that may be held in one hand or clipped to a belt. These devices deliver a continuous flow of low-intensity alternating current (AC) electrical stimulation to the scalp using two adhesive electrodes moistened with a conducting solution (Gilula 2005). Since the electrodes are maintained in place through their adhesive properties (possibly reinforced with a headband) and the main device may be clipped to a belt, patients may engage in sedentary activity such as working on a computer, watching television, or reading while undergoing treatment, though some evidence suggests that results are enhanced when treatment is administered in relaxing, comfortable positions (Gilula 2005).

A standard depression therapy protocol might consist of 20 to 60 minutes of stimulation daily on the first three weeks "at a comfortable level of current," then "treatments every other day or on an as-needed basis for as long as necessary" (Gilula 2005), but considerable variability exists on treatment parameters. No consistent metric is used to describe electrical dose during a single CES administration, and trial reports only inconsistently provide details on the various parameters of electrical stimulation used. Across clinical indications, ranges for the major electrical parameters are the following: frequency, 0.5 Hz to 167 kHz; current amplitude, 100 μA to 4 mA; duration of stimulation per application, continuous application of stimulation from five minutes to six consecutive days (Zaghi 2010); recommended days of use, 30 days to the time needed to achieve benefit (Gilula 2005).

Since no comprehensive database on currently available CES devices exists, the number and variety of existing CES products is unclear. The United States Food and Drug Administration (FDA) indicates that there are 11 different CES devices cleared for marketing in the USA, but the numbers available elsewhere are uncertain. One of the more popular devices is the Fisher-Wallace Cranial Stimulator (model SLB500-B), which is the same device as the Liss Cranial Stimulator (model SLB201-M); this has been marketed for depression, anxiety and insomnia since the 1970s. This device uses scalp electrodes delivering currents of 0.5 to 1.0 mA, with rectangular pulses in three frequencies (15, 500, 15,000 Hz). Another of the more popular devices, the Alpha-Stim Stress Control System, available since the early 1980s, delivers rectangular pulses with frequencies of 0.5, 1.5 and 100 Hz, with a total current output of 10 to 600 μA, and is attached via ear clips. Other devices use different electrode placements (over the orbits, frontal areas, occipital regions) with electrical parameters varying in the ranges described above. In most countries, CES devices are available at pharmacies and through other vendors without a prescription. However, in some countries, due to regulations on medical device advertising and marketing, CES devices may be sold only as treatments for stress rather than as therapy for medical conditions such as depression or anxiety. The United States is the only country where CES devices require a prescription from a licensed healthcare practitioner.

How the intervention might work

The mechanism of action of CES is uncertain, but a number of different models have been proposed to account for the effects of CES on the brain (Kirsch 2002; Smith 2007). These models may be better understood with a review of some of the relevant biological models of depression, which remain controversial (Krishnan 2008).

The 'monoamine hypothesis' of depression, which has been popular since the 1960s, posits that depression results from reduced activity of key neurotransmitters (the monoamine molecules serotonin, norepinephrine, dopamine), the chemical signal molecules produced and released by neurons to communicate with other neurons. All of the currently marketed antidepressants work to increase availability of one or more of these neurotransmitters, generally by blocking their reuptake or degradation (Krishnan 2008).

There is some evidence that CES increases levels of monoamines and other neurotransmitters in the brain, but how it does so remains poorly defined. A study of CES application in dogs suggests that such brain stimulation works by increasing levels of the neurotransmitter dopamine in the central nervous system (Kirsch 2002). Furthermore, trials in humans indicate that CES alters brain levels of the monoamines serotonin and norepinephrine as well as the neurotransmitter GABA (Bystritsky 2008; Zaghi 2010). Finally, an experiment in which CES was found to attenuate opiate withdrawal in rats implies that CES may also increase endorphins, endogenous neurotransmitters that act on the same neuronal receptors as opiate drugs, and mediate a variety of functions including sleep, pain, and mood regulation (Smith 2007). All of these biochemical changes are presumably caused by the penetration of some of the applied electrical current through the scalp into brain tissues, where electrical impulses stimulate changes in neuronal activity such as increasing neurotransmitter release or production (Zaghi 2010). Notably, little if anything is known about the specific function of the various parameters of electrical stimulation, as study reports often omit details on the specific electrical parameters used (Zaghi 2010).

Electroencephalography (EEG) studies offer another perspective on CES effects on brain physiology, although many important aspects of the mechanism of action of CES remain unclear. EEG machines analyze regional and global electrical activity using a series of scalp electrodes attached to a computer. The activity of individual neurons, which is mediated through the neurotransmitters described above, involves generation of electrical impulses between neurons. EEG analysis can therefore provide information about brain function (Sterman 1996). Quantitative EEG researchers have reported EEG correlates of both normal physiological processes and of abnormal mental states associated with various neuropsychiatric disorders (Hughes 1999). A recent study of CES in healthy male volunteers found that CES produced changes in brain electrical activity similar to that produced by meditation (decreases in higher frequency alpha and beta waves, which are associated with stress and arousal), replicating findings of earlier trials (Schroeder 2001). However, the mechanisms whereby CES effects such changes are yet to be defined.

Why it is important to do this review

Because medical device regulation in the United States, and much of the rest of the world, has historically been weak or non-existent, CES device makers have been able to market their product without the submission of controlled clinical trial safety and efficacy results demanded of drugmakers. Indeed, until the 1976 passage of the Medical Device Amendments (MDA) to the federal Food, Drug and Cosmetic Act (FDCA), the United States FDA did not have any regulatory power over medical devices. In 1979, the FDA approved the marketing of CES devices for treatment of insomnia, depression, and anxiety but did not require submission of clinical trial data on safety and efficacy (Bystritsky 2008). Rather, CES devices and more than 1700 other devices already in commercial distribution in the United States, referred to as 'preamendments' devices, were exempted from the 'premarketing approval' process. 'Premarketing approval' would have required submission of clinical trial safety and efficacy data for review by a panel of scientific experts before approval for marketing. Furthermore, newer CES devices have been cleared for marketing without submission of clinical trial data through the '510(k)' provision of the FDCA; under 510(k), a new device may be granted marketing approval upon submission of evidence that it is "substantially equivalent" in its technological characteristics and intended use to an existing "predicate" device (i.e., a device already approved by the FDA) (Hines 2010).

In Europe, where device regulation "is 25 years behind the regulation of medical devices in the United States, and some 25 years behind the European regulations of pharmaceuticals" (Altenstetter 2003), CES devices were granted marketing approval in 1998. Notably, as in the United States, approval of medical devices for European marketing does not require randomized controlled clinical trials demonstrating efficacy (Altenstetter 2003). In Europe and many other countries throughout the world, CES devices are available without a prescription.

Three previous meta-analyses on psychiatric applications of CES have been published (Kirsch 2007; Klawansky 1995; Smith 2007). In 1995, Klawansky et al published a well-conducted meta-analysis on use of CES to treat anxiety, brain dysfunction, headache, and insomnia, but did not consider its use in depression (Klawansky 1995). More recently, Kirsch and Gilula reported on a meta-analysis of CES in depression but provided no information on pre-specified trial search strategy or inclusion criteria. Furthermore, their summary effect size statistic represents an unweighted average of the percentage improvement on various clinical rating scale scores only in patients receiving active CES treatment, rather than between-group comparisons between patients receiving active and sham treatment. In addition, because they pooled data from open-label, uncontrolled trials, and blinded, controlled trials, where quality of allocation concealment is not explicitly addressed, it is likely that their treatment effect is biased toward an over-estimate of effect size. Also, interpretability of their findings is further compromised by their inclusion of trials with marked heterogeneity of primary diagnoses (alcoholism, fibromyalgia, attention deficit-hyperactivity disorder (ADHD), insomnia). Finally, they excluded one negative controlled trial on the grounds that improvement in sham CES-treated patients (inactive treatment control group) "invalidated" the study (Kirsch 2007). Similar problems undermine the conclusions of a meta-analysis of CES trials in depression reported in Smith's monograph on CES (Smith 2007).

In the absence of a systematic review and meta-analysis on the topic, this review may help to elucidate the scientific evidence on the safety and effectiveness of CES, so that healthcare providers and patients may make informed choices on the use of CES as a treatment option for depression.


To assess the effectiveness and safety of alternating current cranial electrotherapy stimulation (CES) compared with sham CES for acute depression.


Criteria for considering studies for this review

Types of studies

Randomized, double-blinded (in which both participants and outcome assessors are blinded to treatment allocation), parallel-group trials will be included in this review.

In sham CES treatment, devices that are visually identical to those used in active treatment deliver either no current or only a brief, low-intensity current sufficient to produce skin tingling at the site of electrode attachment (O'Connell 2011).

Cluster-randomized trials will be eligible for inclusion (although this trial design is rarely used in studies of biological therapies for depression).

Cross-over trials will be included, but only data from the first treatment period will be used.

Types of participants

Adults (persons aged 18 to 75) with a primary diagnosis of a depressive disorder diagnosed according to standardized diagnostic criteria (e.g., DSM-IV (APA 2000), ICD-10 (WHO 2007)).

Trials in which participants have common comorbid psychiatric disorders, such as anxiety or substance misuse disorders, will be included if they meet all other inclusion criteria. Trials in which participants have comorbid medical conditions known to affect mood or response to antidepressant treatment will be included if they meet all of the other inclusion criteria.

Studies in which participants have dementia or other conditions characterized by structural brain damage or brain neurodegeneration, psychotic disorders, or personality disorders, will be excluded.

Since this review will focus on treatment of acute depression, trials of CES in chronic or refractory depression will be excluded.

Types of interventions

Experimental intervention

Participants are treated with a CES device to administer a dose of AC electrical stimulation thought to be sufficient to penetrate the skull and induce changes in brain electrical activity. Electrodes are applied to the head via ear-clips or scalp electrodes held in place by a head band or adhesive material. Participants undergoing CES may either be resting comfortably or engaged in sedentary activity. As noted above (under Description of the intervention), electrical parameters used in active CES treatment vary across studies, as does the duration of treatment (both the amount of time CES is applied per treatment session and frequency of treatment sessions). Since there may be only a few trials eligible for inclusion in this review, trials will not be excluded based on the electrical parameters used or frequency/duration of CES treatments, and dose ranging trials will be included if they contain a sham CES control arm.

Control/comparator intervention

Participants receiving the control intervention receive sham CES. In sham CES, a CES unit visually identical to the active CES device is used to administer electrical stimulation to the skin under the electrodes sufficient to induce local tingling but insufficient to penetrate the skull. Local skin irritation can be achieved with electrical stimulation at a current intensity significantly less than is needed to pass through the underlying muscle and bone. In the most rigorous trial designs, the superficial current in the sham treatment is individually adjusted for each participant to match the tingling caused by active treatment, such that participants are unable to distinguish between the sham and active interventions (Marshall 1975).

Excluded interventions

Head-to-head trials comparing CES with another active intervention and no sham treatment arm will be excluded.

Types of outcome measures

Validated self rated and observer-rated depression scales are used to generate continuous variable measures of depression severity. They may also be used to generate categorical outcomes variables such as treatment response/non-response and remission/non-remission. Thus, a common definition of depression treatment 'responder' is a participant experiencing a greater than 50% improvement in the baseline depression score. Remission is usually defined by the occurrence of a depression score below a specific value designating minimal residual depressive symptoms. Remission is arguably a more important and robust treatment goal, but may not be consistently reported.

We will provide a descriptive summary of adverse events, as reported in the clinical trials, and we will examine rates of adverse events.

Primary outcomes
  1. The primary benefit outcome will be differences in change scores between intervention and control groups on validated rating scales such as the following:

    1. Observer-rated depression rating scale such as the Montgomery-Asberg Depression Rating Scale (MADRS) (Montgomery 1979) and the Hamilton Rating Scale for Depression (HAM-D) (Hamilton 1960).

    2. Self rated depression scale such as the Beck Depression Inventory (BDI) (Beck 1988; Beck 1993).

  2. The primary harm outcome will be differences in rates of discontinuation due to adverse events between active treatment and control groups.

Secondary outcomes
  1. Response and remission rates.

  2. A validated global change measure such as the Clinical Global Impression of Change scale (CGIC) (Guy 1976).

  3. Differences in change scores on anxiety rating scales such as the Hamilton Anxiety Rating Scale (HAM-A) (Hamilton 1959).

  4. Rates of adverse events.

  5. Discontinuation rates due to lack of efficacy and all causes.

Timing of outcome assessment

Outcomes may be reported at different follow-up periods. In general, since most CES treatment protocols for depression involve multiple treatments per week over at least a few weeks, it is expected that outcome assessments will occur after one to two weeks of treatment, then every one to two weeks until trial completion (2 to 10 weeks). In some cases, longer-term outcomes of three, six, nine, or 12 months may be reported.

Search methods for identification of studies

The Cochrane, Depression, Anxiety and Neurosis Review Group's Specialized Register (CCDANCTR)

The Cochrane Depression, Anxiety and Neurosis Group (CCDAN) maintain two clinical trials registers at their editorial base in Bristol, UK, a references register and a studies-based register. The CCDANCTR-References Register contains over 31,500 reports of randomized controlled trials in depression, anxiety, and neurosis. Approximately 65% of these references have been tagged to individual, coded trials. The coded trials are held in the CCDANCTR-Studies Register and records are linked between the two registers through the use of unique Study ID tags. Coding of trials is based on the EU-Psi coding manual. Please contact the CCDAN Trials Search Co-ordinator for further details. Reports of trials for inclusion in the Group's registers are collated from routine (weekly), generic searches of MEDLINE (1950-), EMBASE (1974-), and PsycINFO (1967-); quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL), and review-specific searches of additional databases. Reports of trials are also sourced from international trials registers c/o the World Health Organization's trials portal (ICTRP),, drug companies, the handsearching of key journals, conference proceedings, and other (non-Cochrane) systematic reviews and meta-analyses.

Details of CCDAN's generic search strategies can be found on the Group's website.

Electronic searches

The Trials Search Co-ordinator (TSC) at CCDAN will run searches of their specialized registers (CCDANCTR-Studies and CCDANCTR-References) using the following search terms:


Condition = (depress* or dysthymi* or "affective disorder*" or "affective symptom*" or "mood disorder*" or "adjustment disorder*") and Intervention = (electrotherapy or "electric stimulation" or "cranial electrostimulation" or "electric treatment" or "electrosleep" or "electronarcosis" or "cranial alpha stimulation")


Free-text = ((depress* or dysthymi* or "affective disorder*" or "affective symptom*" or "mood disorder*" or "adjustment disorder*") and ("electric stimul*" or "electric therap*" or "electric treatment*" or "electrical stimul*" or "electrical therap*" or "electrical treatment*" or electrotherap* or CES or "electrosleep" or "electronarcosis" or "cranial alpha stimulation"))

Additionally, the authors will search the WHO trials portal ( and

Searching other resources

We will search references from related articles identified through the primary search for relevant clinical trials. In addition, we will contact the manufacturers of FDA-approved CES devices for any further unidentified clinical trials. We will also contact trial authors and experts in the field for other references.

Data collection and analysis

Selection of studies

Two of the authors of this review (HK and KL) will independently review the results of the search and identify trials potentially eligible for inclusion. Disagreements on trial selection will be resolved by discussion between the first two authors and where necessary by consultation with a third author (KC).

Data extraction and management

Study design

Two authors (HK and KL) will independently tabulate in a data extraction form study design, sample size, participant characteristics (demographics, diagnoses including comorbid conditions, history of prior antidepressant treatment), outcome measures, parameters of the interventions, trial duration, outcome measures and time points at which participants are assessed with outcome measures, statistical methods used, and assessment of potential sources of bias.

Participant characteristics

We will record participant characteristics that might impact treatment outcome including severity of illness, presence of comorbid psychiatric and medical conditions, history of treatment refractoriness, and compliance with the treatment protocol. We will also record methodological factors impacting outcomes including electrical parameters, frequency and duration of treatment, choice of outcomes measures, and quality of blinding of assessors and participants.

Main planned comparisons
  1. Alternating current cranial electrotherapy stimulation versus sham treatment

Assessment of risk of bias in included studies

CES trials with inadequate blinding of either participants or raters present insurmountable problems with bias. Participants must be blinded to treatment allocation because participant self report constitutes an important element in clinician-rated scales of depression and adverse events.

Two review authors (HK and KL) will independently assess the risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The domains will be as follows.

  1. Sequence generation (checking for possible selection bias).

  2. Allocation concealment (checking for possible selection bias).

  3. Participant blinding (checking for possible performance bias and detection bias).

  4. Outcome assessor blinding (checking for possible performance bias and detection bias).

  5. Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations).

  6. Selective reporting bias (checking if expected outcomes were reported).

  7. Other sources of bias (such as stopping the trial early or changing methods during the trial).

We will rate studies according to whether risk of a particular type of bias is adequately controlled for or not with the following assessments: low risk of bias, high risk of bias, or unclear risk of bias.

Measures of treatment effect

Continuous outcomes

Since different trials are likely to use different continuous outcomes scales, we will compute standardized mean differences (SMD) together with 95% confidence intervals (CI) as the summary measure of continuous outcomes. We will generate separate SMD values for self reported and observer-rated scales.

Dichotomous outcomes

For dichotomous outcomes, the summary statistic will be the risk ratio (RR) together with 95% confidence intervals (CI).

Unit of analysis issues

Cross-over trials

Given the documented risk of carry-over from CES (Flemenbaum 1974; Kirsch 2002) we will only use first treatment period data from cross-over trials in our analyses (Higgins 2011).

Cluster-randomized trials

Although randomization in cluster-randomized trials occurs at the level of clusters rather than individuals, results from such trials are often incorrectly analyzed as though randomization had occurred at the individual participant level. In such cases, we will attempt to correct for these analytical errors by computing the design effect and estimating the 'effective sample size' using the intraclass correlation coefficient (ICC) where reported by trialists (Higgins 2011).

Multiple-armed trials

In trials including more than two treatment arms, such as two different 'doses' of CES (e.g., different electrical parameters) versus sham CES, we will combine the two different CES groups into one active CES treatment group compared with sham CES.

Dealing with missing data

In cases where data are missing, we will contact the authors of the trial report to obtain missing data. If standard deviations remain unavailable, we will estimate them using reported group means, standard errors, or P values. If a P value is used to compute standard deviations but is imprecisely reported (e.g., P < 0.05), we will use the limiting value for this statistic to approximate the actual value (e.g., in the case of a reported P < 0.05, we will set P = 0.05) (Higgins 2011).

In cases where missing data are imputed using 'last observation carried forward' (LOCF) or other approaches, we will explore assumptions for data imputation and the potential impact on findings in the discussion section (Higgins 2011).

Assessment of heterogeneity

Qualitative assessment of heterogeneity will be based on consideration of possible clinical and methodological differences between trials. We will assess clinical heterogeneity between trials by comparison of depression severity, age and gender of participants, and presence/absence of psychiatric and medical comorbidity. We will assess methodological heterogeneity between trials by comparison of electrical parameters and outcome measures. We will perform subgroup meta-analyses using clinical and methodological subgroups if there are at least two trials per subgroup.

We will assess the I2 statistic as a quantitative estimate of heterogeneity. I2 values will be interpreted as follows: 0% to 40%, might not be important; 30% to 60%, may represent moderate heterogeneity; 50% to 90%, may represent substantial heterogeneity; 75% to 100%, considerable heterogeneity. Assessment of the importance of the I2 statistic will depend on the magnitude and direction of effects as well as the strength of the evidence supporting this estimate of heterogeneity (P value of the associated Chi2 test) (Higgins 2011).

Assessment of reporting biases

We will attempt to minimize the impact of reporting bias by searching multiple sources, as indicated in our search methods above, including trial registries and unpublished trial results. In cases of incomplete results reporting, we will attempt to obtain data omitted from published reports by contacting the clinical trialists. Finally, in our discussion, we will consider the potential impact of location, language, citation, duplication, outcome reporting, and publication bias on meta-analytic results (Higgins 2011).

We will use qualitative (visual inspection) and quantitative (formal statistical tests) assessment of funnel plots for asymmetry to explore for publication bias if there are at least 10 trials in the meta-analysis (Higgins 2011). With continuous outcome variables, we will use the test proposed by Egger to assess for funnel plot asymmetry (Egger 1997; Higgins 2011). With dichotomous outcomes reported using odds ratios, we will use the test recommended by Peters (Peters 2006) to assess funnel plot asymmetry (Higgins 2011). Significant asymmetry indicated through these approaches will be interpreted as possible evidence of reporting bias.

Data synthesis

Trials using different stimulation parameters (different doses of electrical stimulation) will be included in a comprehensive meta-analysis to obtain a summary estimate of CES effect size in depression.

The decision to use fixed-effect or random-effects meta-analysis will be based on the extent to which studies are deemed similar in populations, interventions, comparators, outcomes, and settings. We will perform fixed-effect meta-analyses if included trials, or a subgroup of trials, share key clinical and methodological features. We will use random-effects models to generate summary meta-analytic results otherwise.

In trials with serial assessments on participants, we will compute separate summary statistics for different follow-up periods as applicable: one week, two weeks, four weeks, six to eight weeks, 12 weeks, 24 weeks (Higgins 2011).

Subgroup analysis and investigation of heterogeneity

If trial data are available, we will conduct subgroup analyses according to the following clinical or methodological features.

  1. Depression severity

  2. Gender

  3. Age

  4. Presence/absence of medical comorbidity

  5. Trial outcomes measures (i.e., self rated versus observer-rated depression scales)

Sensitivity analysis

We will conduct a series of sensitivity analyses using the following criteria.

  1. Excluding trials with possible bias (as per 'Risk of bias' tables) (specifically, trials rated to have inadequately or unclearly controlled for one or more type of bias will be considered possibly biased)

  2. Excluding trials reporting only per protocol analyses

  3. Excluding trials reporting only observed cases analyzed

  4. Excluding cross-over trials

  5. Excluding cluster-randomized trials

  6. Excluding trials with using imputed data

  7. Excluding trials with imputed standard deviations

  8. Excluding trials in which sham CES delivers no electrical current


CRG Funding Acknowledgement:
The National Institute for Health Research (NIHR) is the largest single funder of the Cochrane Depression, Anxiety and Neurosis Group.

The views and opinions expressed therein are those of the authors and do not necessarily reflect those of the NIHR, NHS or the Department of Health.

Contributions of authors

HK devised protocol and will undertake data extraction, data management (including entering data into RevMan (RevMan 2012)), data analysis and interpretation of results, and writing completed review. KL will participate in data extraction, interpretation of results, and writing up of completed review. KC will participate in data extraction, interpretation of results, and writing completed review.

Declarations of interest