Introduction
Obsessive-compulsive disorder (OCD) is a debilitating psychiatric disorder with an approximate lifetime prevalence of 1–2% (Diagnostic and Statistical Manual of Mental Disorders 5th ed., text rev.; DSM-5-TR; American Psychological Association, 2022). Left untreated, OCD can cause significant impairments in social, family, and occupational functioning (Abramowitz et al., Reference Abramowitz, Taylor and McKay2009). Current evidence-based treatment options for OCD are pharmacotherapy and/or cognitive behavior therapy that includes exposure and response prevention (CBT-ERP). Whilst OCD typically responds to serotonin reuptake inhibitors (SRIs) and selective serotonin reuptake inhibitors (SSRIs), pharmacotherapy is not always effective and approximately 40 – 60% of people with OCD are either unable to tolerate side-effects or only partially respond to these approaches (Gershkovich et al., Reference Gershkovich, Wheaton and Simpson2017). Discontinuation of medications is also associated with symptom relapse (Fineberg et al., Reference Fineberg, Reghunandanan, Simpson, Phillips, Richter, Matthews, Stein, Sareen, Brown and Sookman2015). Similarly, of those who respond to CBT-ERP, approximately 60% demonstrate at least partial relapse (Eisen et al., Reference Eisen, Sibrava, Boisseau, Mancebo, Stout, Pinto and Rasmussen2013; Simpson, Franklin, Cheng, Foa, & Liebowitz, Reference Simpson, Franklin, Cheng, Foa and Liebowitz2005) and a significant percentage of those treated for OCD (14–31%) are classed as non-responders (Foa et al., Reference Foa, Liebowitz, Kozak, Davies, Campeas, Franklin, Huppert, Kjernisted, Rowan, Schmidt, Simpson and Tu2005; Norberg et al., Reference Norberg, Calamari, Cohen and Riemann2008).
Given the percentage of non-responders and rate of relapse for those with OCD, noninvasive brain stimulation is being investigated as a therapeutic approach. Transcranial direct current stimulation (tDCS) is a low-cost, safe, and well tolerated noninvasive brain stimulation technique (Kuo et al., Reference Kuo, Chen and Nitsche2017) that involves delivering a low intensity electrical current (1–2mA) via two electrodes on the scalp, an anode (excitatory) and a cathode (inhibitory). Depending on the location of the electrodes, tDCS has been demonstrated to modulate local and network-level brain activity and alter maladaptive behavior (Nardone et al., Reference Nardone, Bergmann, Christova, Caleri, Tezzon, Ladurner, Trinka and Golaszewski2012).
It is feasible that tDCS may be effective in addressing deficits in the neural feedback loop thought to underly extinction and inhibitory learning, which are implicated in the onset and maintenance of anxiety and OCD (see Craske et al., Reference Craske, Liao, Brown and Vervliet2012; Lissek et al., Reference Lissek, Powers, McClure, Phelps, Woldehawariat, Grillon and Pine2005). The clinical efficacy of tDCS has been researched most commonly in relation to treatment-resistant depression (George et al., Reference George, Padberg, Schlaepfer, O’Reardon, Fitzgerald, Nahas and Marcolin2009). However, there is a growing body of evidence that suggests tDCS can improve OCD symptoms (Brunelin et al., Reference Brunelin, Mondino, Bation, Palm, Saoud and Poulet2018; Nuñez et al., Reference Nuñez, Zinbarg and Mittal2019; Pinto et al., Reference Pinto, Cavendish, da Silva, Suen, Marinho, Valiengo, Vanderhasselt, Brunoni and Razza2022).
Until 2018, published studies involving tDCS for OCD included case reports (n = 8), open-label trials (Bation et al., Reference Bation, Poulet, Haesebaert, Saoud and Brunelin2016; Dinn et al., Reference Dinn, Aycicegi-Dinn, Göral, Karamursel, Yildirim, Hacioglu-Yildirim, Gansler, Doruk and Fregni2016; Najafi et al., Reference Najafi, Fakour, Zarrabi, Heidarzadeh, Khalkhali, Yeganeh, Farahi, Rostamkhani, Najafi, Shabafroz and Pakdaman2017), and one randomized controlled partial crossover trial (D’Urso et al., Reference D’Urso, Brunoni, Mazzaferro, Anastasia, de Bartolomeis and Mantovani2016). It is difficult to establish the efficacy of tDCS for OCD from these trials due to methodological variations, lack of protocol reporting, and highly mixed results. For example, two of the case reports involved 20 sessions of anodal tDCS to increase cortical activation of the pre supplementary motor area (pre-SMA) and cathodal tDCS to reduce activation of the orbital frontal cortex (OFC) (Hazari et al., Reference Hazari, Narayanaswamy, Chhabra, Bose, Venkatasubramanian and Reddy2016; Narayanaswamy et al., Reference Narayanaswamy, Jose, Chhabra, Agarwal, Shrinivasa, Hegde, Bose, Kalmady, Venkatasubramanian and Reddy2015). Both patients in Narayanaswamy et al.’s case report demonstrated reduced pre-post Yale Brown Obsessive Compulsive Scale (YBOCS) scores (40% and 46.7%) and this was maintained for 1 to 2 months. The participant in the Hazari et al., case report demonstrated a YBOCS reduction of 80% that was maintained for 7 months post-intervention with minor fluctuations (Hazari et al., Reference Hazari, Narayanaswamy, Chhabra, Bose, Venkatasubramanian and Reddy2016). D’Urso et al., (Reference D’Urso, Brunoni, Anastasia, Micillo, de Bartolomeis and Mantovani2016a) also applied anodal tDCS stimulation to pre-SMA across 10 sessions. They found that the participant’s Y-BOCS score initially increased (i.e., OCD symptoms worsened). In response, they changed the electrodes so that the cathode (inhibitory) was positioned over pre-SMA and reported an improvement in OCD symptoms (D’Urso et al., Reference D’Urso, Brunoni, Anastasia, Micillo, de Bartolomeis and Mantovani2016). D’Urso et al., extended their findings by conducting a follow-up crossover design study (N = 10) and found that cathodal (but not anodal) tDCS of pre-SMA led to significant improvements in Y-BOCS (D’Urso et al., Reference D’Urso, Brunoni, Mazzaferro, Anastasia, de Bartolomeis and Mantovani2016). Silva et al. (Reference Silva, Brunoni, Miguel and Shavitt2016) also applied cathodal tDCS over SMA in their case studies (n = 2). They reported a small, delayed improvement in Y-BOCS (18% at 6 months post-intervention) in one person, while the other demonstrated a 45% improvement of symptoms at 6 months post-intervention. Despite a number of studies reporting such improvements in OCD symptoms in response to tDCS, the small sample sizes, heterogeneity of tDCS protocols, and lack of methodological rigor reduces the quality of the findings (Brunelin et al., Reference Brunelin, Mondino, Bation, Palm, Saoud and Poulet2018). The optimal montage and efficacy of tDCS remains ambiguous.
A more recent systematic review and meta-analysis evaluated the efficacy of tDCS for OCD and the optimal tDCS montage using electric field monitoring (Pinto et al., Reference Pinto, Cavendish, da Silva, Suen, Marinho, Valiengo, Vanderhasselt, Brunoni and Razza2022). The review included eight studies (four open-label and four randomized controlled trials [RCTs]) up until 2021 and excluded several studies due to bias related to methodological issues. Pinto et al. concluded that tDCS is a promising intervention to reduce OCD symptoms, despite reporting no significant effect of active tDCS compared to a sham condition. Pinto et al., attributed the lack of a positive effect to there only being four RCTs included in their meta-analysis, which were limited by their small sample sizes and highly variable tDCS placement and dose (Pinto et al., Reference Pinto, Cavendish, da Silva, Suen, Marinho, Valiengo, Vanderhasselt, Brunoni and Razza2022). Since this review, there have been several further RCTs concerning tDCS for OCD published, adding to the available data for a new overarching review. To date, however, no systematic reviews or meta-analyses on this topic have used the Consolidating Standards of Reporting Trials (CONSORT) statement, the “gold standard” for assessing the reporting quality of clinical trials (Altman et al., Reference Altman, Moher, Egger, Davidoff, Elbourne and Schulz2001) to consider the quality of the RCTs being included in their reviews.
The aim of the present study was to conduct a systematic review of the literature to date involving RCTs of tDCS for OCD (including parallel and crossover design) and evaluate the reviewed studies using the CONSORT statement as a framework for understanding the quality of methodological and outcome reporting.
Materials and methods
This systematic review was prospectively registered with PROSPERO (CRD42023426005).
Search strategy and selection of studies
The Prospero Registered proposal aimed to include RCTs, open trials, case-reports/studies, and case-series. However, a preliminary search yielded multiple RCTs, which allowed for a more focussed review of RCTs only.
Data for this systematic review was collected in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (Page et al., Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann, Mulrow, Shamseer, Tetzlaff, Akl, Brennan, Chou, Glanville, Grimshaw, Hróbjartsson, Lalu, Li, Loder, Mayo-Wilson and Moher2021). A systematic database search was conducted on 28th June, 2023 of Ovid® MEDLINE, Emcare, Global Health, Embase, PsycInfo, PsycArticles, PubMed, Cochrane, and Google Scholar. The reference lists of earlier relevant reviews were searched by hand. ClinicalTrials.gov was also searched for unpublished results to reduce the risk of bias. The final search was updated on July 24th, 2023.
The initial “advanced search” was conducted using the keywords (Transcranial direct current stimulation) AND (obsessive compulsive disorder), limits/filters included human*, English language, additional limits were adults 18+ OR all adults (19 plus years), adulthood 18+, and randomized controlled trial OR control* OR sham. A total of 120 articles were identified, which were reduced to 68 records after screening and removal of duplicates. The remaining records were then assessed for eligibility.
Selection criteria
We included RCTs involving tDCS for OCD conducted in adults who met the diagnostic criteria for OCD in accordance with the Diagnostic and Statistical Manual of Mental Disorders versions IV-TR and 5-TR or the World Health Organization’s International Statistical Classification of Diseases and Related Health Problems (ICD) 10th and 11th ed. (World Health Organisation, 1992, 2021) criteria.
Exclusion criteria included (a) case reports and case studies; (b) studies without a standardized OCD outcome measure; (c) studies where investigators were targeting their treatment for another condition (i.e., OCD was a comorbid condition); (d) noninvasive brain stimulation techniques other than tDCS. See Figure 1 for identification, screening, and inclusion criteria.
PRISMA flow diagram.
Quality assessment
The quality of the included studies was evaluated in accordance with the CONSORT statement (Moher et al., Reference Moher, Hopewell, Schulz, Montori, Gøtzsche, Devereaux and Altman2012). The CONSORT statement is a 25-item checklist which was developed to improve the quality of reporting of RCTs. First published in 1996, it has been updated as the CONSORT 2010 Statement (Moher et al., Reference Moher, Hopewell, Schulz, Montori, Gøtzsche, Devereaux and Altman2012) with extensions to include randomized controlled nonpharmacologic treatment trials and other trial designs. The guidelines facilitate critical appraisal and interpretation of trial reports. Data was extracted and tabulated by the first author (PG), and results were cross-checked by a postgraduate psychology research student. Any disparities in quality assessment were settled by one of the co-authors/supervisors (R.A.A, A.M. L). As the replication and translation of tDCS, as undertaken in these studies, is only possible when detailed information about methods is provided, this study also collated data on the degree to which the tDCS protocol was reported on a range of relevant factors (e.g., electrode size, contact medium, dose stability).
Outcomes
At its inception, this study sought to examine the clinical effectiveness of tDCS for OCD via a meta-analysis. We elected not to undertake this due to the heterogeneity in protocols and the conflicting and often lack of empirical bases for the choice of stimulation sites. There was limited overlap in procedures in terms of electrode montages, dose (i.e., duration and frequency) and, in particular, stimulation sites. The lack of consistency in protocols represents a significant limitation that would cast doubt over any meta-analysis results, and render interpretations less meaningful (Kriston, Reference Kriston2013). A descriptive summary has been provided of the study outcomes and conclusions as an alternative.
Results
Systematic review
Eleven RCTs with 363 participants were included in this review. See Table 1 for a summary of the key characteristics of the eleven RCTs.
Main characteristics and findings of randomised controlled trials involving transcranial direct current stimulation (tDCS) for obsessive compulsive disorder (OCD)
Note: tDCS = transcranial direct current stimulation, L = left, R = right, RCT = randomized controlled trial, DLPFC = dorsolateral prefrontal cortex, NR = Not Reported, OFC = orbitofrontal cortex, (pre) SMA = (pre) supplementary motor area, mA = milliampere, μA = microampere, hrs = hours, min: minutes, ms = millisecond, Y-BOCS = Yale-Brown Obsessive Compulsive Scale, OLE = open-label extension, (SD) standard deviation, TCA = tricyclic antidepressant, SGA, BDZs, SSRI = selective serotonin reuptake inhibitor, CBT-ERP = cognitive behavioral therapy with exposure and response prevention, N/A = Not applicable, no true sham condition.
1 Between group difference observed at 12 weeks post-intervention F (1, 84.06) = 84.06, p<0.03, d = 0.62
2 Between group difference observed at 12 weeks post-intervention F (1, 75) = 4.60, p<0.035, η 2 = 0.058
CONSORT evaluation
Table 2 presents a summary of the CONSORT evaluation of the 11 studies.
CONSORT evaluation of treatment studies in chronological order of publication date
Note: Not Reported 
, Present with limitations 
, Reported 
, *See table 3 for tDCS replication.
Inclusion of information needed to interpret findings and replicate tDCS in real world conditions
Note: Not Reported , Present with limitations , Reported , Contact medium (saline solution, electrode gel), *If dose stability was maintained throughout treatment and follow-up periods.
Trial design
Six of the eleven studies were randomized double-blind sham-controlled trials. One was a randomized sham-controlled trial. One was a multicenter randomized, double-blind sham-controlled design, and one was a randomized, double-blind, sham-controlled, cross-over, multicenter trial. Two were randomized controlled cross-over trials. One was a quasi-experimental trial with pre-post testing.
Recruitment and sample size
All but one study reported where participants were recruited from (e.g., universities/research institutes [n = 4], clinics [n = 3], and hospitals [n = 3]), with the exception of Harika-Germaneau, Heit, et al. (Reference Harika-Germaneau, Heit, Drapier, Sauvaget, Bation, Chatard, Doolub, Wassouf, Langbour and Jaafari2024), it was unclear where tDCS took place and whether the sample were inpatients or outpatients at the time. A common shortcoming was the lack of dates defining the recruitment period, which was only reported by five studies (Akbari et al., Reference Akbari, Hassani-Abharian and Tajeri2022; Gowda et al., Reference Gowda, Narayanaswamy, Hazari, Bose, Chhabra, Balachander, Bhaskarapillai, Shivakumar, Venkatasubramanian and Reddy2019; Harika-Germaneau et al., Reference Harika-Germaneau, Gosez, Bokam, Guillevin, Doolub, Thirioux, Wassouf, Germaneau, Langbour and Jaafari2024; Yoosefee et al., Reference Yoosefee, Amanat, Salehi, Mousavi, Behzadmanesh, Safary and Salehi2020; Silva et al., Reference Silva, Brunoni, Goerigk, Batistuzzo, Costa, Diniz, Padberg, D’Urso, Miguel, pedes and Shavitt2021).
Participants
A total of 363 participants were enrolled across the eleven trials. The number of participants in each trial varied from 12 to 80. Age ranged from 18–70 years (Mean = 37years). The gender of participants was reported by nine trials which comprised of 136 male participants (45%) and 164 female (55%) overall. Baseline OCD symptom severity (as measured by the Y-BOCS) was reported by ten studies and ranged from 19 (moderate) to 30 (severe) with a mean Y-BOCS of 27 (severe). The duration of OCD was reported by six studies and ranged from 7.4 to 31.1 years.
Eligibility
All trials recruited adults only. The age range of participants varied across trials and included 18 years or older, 20–45 years, 18–45 years (n = 2), 18–65 years (n = 3), 18–60 years, and 18–70 years (n = 2). One study did not report an age range. All trials required that OCD meet the diagnostic criteria of the DSM-IV, DSM-5 (later studies), or ICD-10 or -11, with eight stipulating a minimum Y-BOCS symptom severity of 16. Two trials recruited participants with a slightly higher minimum Y-BOCS of ≥20, and one trial did not specify symptom severity in their eligibility criteria.
Eight out of the eleven trials recruited treatment resistant participants, but the definition of treatment resistance was inconsistent across studies. Most trials described treatment resistance as a lack of response to SSRIs. However, the number of SSRIs trialled without response varied from one to at least two SSRIs, with one study specifying either no response to two SSRIs or one SSRI and clomipramine. Three studies included a lack of response to CBT by a trained psychologist in addition to a lack of response to SSRIs. Three out of the eight trials also stipulated that an SSRI be trialled for at least 3 months at a maximum tolerated dose, with no response. By contrast, one trial’s criteria were a lack of response to an SSRI trialled for 2 weeks prior the intervention. Three of the trials included participants prescribed psychotropic medications with a proviso that the medication dose had been stable prior to the inclusion in the study. Studies varied in their definition of medication stability from 4 to 12 weeks.
The exclusion criteria varied across studies ranging from only one exclusion criteria (psychotropic medications) to a more extensive list of eight different exclusion criteria. The most common exclusion criteria were history of psychotic disorder, head/brain injury, metal or implanted device in brain, substance abuse, and pregnancy.
Interventions
The CONSORT Intervention item consists of 4 parts including a description of different components, and whether/how - interventions were standardized; adherence to protocol was assessed or enhanced; and adherence of participants was assessed or enhanced. Table 3 provides an overview of the level of information in each of the trials required to replicate a tDCS protocol in real-world conditions and understand the generalizability of findings. All studies included a description of the different components of the intervention to varying degrees, but not all studies included enough detail for replication. For example, two studies did not report the size of electrodes used, and half of the studies did not indicate how the electrodes were secured to the head. Five studies did not provide an adequate description of previous treatment, whether pharmacotherapy was concomitant and, if it was, if the dose remained stable throughout the trial and was maintained during the follow-up period.
Four out of eleven studies reported the use of standardized diagnostic procedures and assessment-tools and described the psychometric properties of the outcome measures. However, none of the studies included how the interventions were standardized. Adherence to the protocol was assessed and enhanced by only one of the trials. Yoosefee et al. (Reference Yoosefee, Amanat, Salehi, Mousavi, Behzadmanesh, Safary and Salehi2020) reported that tDCS was delivered by a doctor and nurse trained in the technique, and procedures were monitored to ensure blinding of participants to the intervention was maintained. Silva et al. (Reference Silva, Brunoni, Goerigk, Batistuzzo, Costa, Diniz, Padberg, D’Urso, Miguel, pedes and Shavitt2021) was the only trial of the eleven trials to address the participant’s adherence to the interventions by arranging for medications to be prescribed and provided daily at the outpatient clinic where brain stimulation took place to ensure the dose remained stable during the trial.
Outcomes measures and follow-up
Overall, the pre-specified primary and secondary outcomes were well reported. A reduction in symptom severity, as measured by the Y-BOCS, was the primary outcome for ten trials. The other trial by Yekta et al. (Reference Yekta, Rostami and Fayyaz2015) indicated that decision making, and a reduction of obsession symptoms, assessed using the Y-BOCS, were their primary outcomes. CONSORT standards also require trials to identify whether there were any changes to trial outcomes after the trial commenced. This was an area that was poorly done and only explicitly addressed by Yoosefee et al. (Reference Yoosefee, Amanat, Salehi, Mousavi, Behzadmanesh, Safary and Salehi2020). In terms of adverse outcomes of tDCS, all except two trials (Akbari et al., Reference Akbari, Hassani-Abharian and Tajeri2022; Yekta et al., Reference Yekta, Rostami and Fayyaz2015) described assessing adverse events and reported that tDCS was generally well tolerated. The most common adverse effects reported were mild headache, local redness, and itching or tingling of the skull at the electrode site. Three moderate severity events were reported by Silva et al. (Reference Silva, Brunoni, Goerigk, Batistuzzo, Costa, Diniz, Padberg, D’Urso, Miguel, pedes and Shavitt2021), including drowsiness, change in appetite, and muscle tic, but the authors noted that none of the events required any specific intervention. Follow-up periods were included in the design of six trials and varied from 4 weeks to 3 months.
Randomization and blinding
A common limitation across all studies was the reporting on randomization. Despite all trials involving randomization, the details of the random allocation sequence were often missing or unclear. Five out of eleven studies described the type of randomization used, and the method used to generate the random allocation sequence. Only two provided details of any restrictions (such as blocking and block size). Likewise, the implementation of random allocation was poorly addressed. Half of the studies reported who generated the random allocation sequence, and who was responsible for enrolling and assigning participants to interventions. However, only four studies described the methods used to conceal the random allocation sequence. Blinding on the other hand was well reported with ten of the trials providing a description of who (e.g., participants, those administering interventions, and/or those assessing the outcomes) was blinded after assignment to interventions.
Statistical methods
Generally, most trials (ten) described what statistical methods were used to compare groups for primary and secondary outcomes, and where applicable, reported methods utilized for additional analyses. Seven trials provided sufficient information to be included in a meta-analysis (mean and standard deviation) pre- to post-intervention. Six trials reported an estimated effect size for between group differences for primary and secondary outcomes, and only four studies also reported the precision of the effect size (i.e., 95% confidence intervals [CI]).
Sampling issues and participant flow
Sample size was only addressed by four trials Harika-Germaneau, Heit, et al. (Reference Harika-Germaneau, Heit, Drapier, Sauvaget, Bation, Chatard, Doolub, Wassouf, Langbour and Jaafari2024); Gowda et al. (Reference Gowda, Narayanaswamy, Hazari, Bose, Chhabra, Balachander, Bhaskarapillai, Shivakumar, Venkatasubramanian and Reddy2019); Silva et al. (Reference Silva, Brunoni, Goerigk, Batistuzzo, Costa, Diniz, Padberg, D’Urso, Miguel, pedes and Shavitt2021) and Yoosefee et al. (Reference Yoosefee, Amanat, Salehi, Mousavi, Behzadmanesh, Safary and Salehi2020) who reported utilizing a power analysis to determine the sample size necessary to observe a significant treatment effect. Four trials did describe their small sample size as a limitation in the discussion section of their trials, whereas three trials did not address sample size, or the issues associated with an underpowered trial.
A participant flow diagram was used by seven out of eleven trials to illustrate the number of participants who were randomly assigned, losses and exclusions after randomization together with reasons, and the numbers who received the intended treatment and were analyzed for the primary outcome. None of the trials reported the delay between randomization and the initiation of the intervention. Also, no trial explicitly reported why the trial ended or was stopped, leaving the reader to assume that they stopped due to achievement of sample size and/or follow up was completed.
Discussion section items
The limitations, generalizability, and interpretation of the results in most trials were not adequately addressed in the discussion section of articles. All but one study identified at least one limitation associated with their trial. Four trials identified potential sources of bias, recognized imprecision as a limitation (associated with locating stimulation sites without assistance of a neuro-navigation tool), and possible issues associated with blinding (i.e., who was and who was not blinded).
The generalizability of the results in terms of intervention, comparators, patients, care providers and centers involved in the trial was also limited. Most studies (eight out of eleven) discussed generalizability in their papers by acknowledging their patient characteristics in relation to established norms. They did not, however, discuss generalizability in terms of the characteristics of the care providers and unique features of the treatment settings involved with their trials.
The interpretation of findings in the discussion sections was not always consistent with results. Indeed, half of the studies made concluding statements that were incongruent with their results. Despite the majority of trials reporting no significant between group differences in the post-intervention active versus sham outcome measures, tDCS was still described as a promising approach for OCD. Furthermore, some studies only discussed selective results for example, one study reported an improvement in a decision-making task, one of their primary outcomes, but in the discussion stated that obsessive-compulsive symptoms were also significantly reduced without reporting the post intervention Y-BOCS scores or including the statistical analyses to support the statement. The premise of their interpretation was that impaired decision-making is associated with OCD.
Other information section
The other information section is comprised of trial registration, access to full protocol, and funding. Half of the studies reported whether their trial was registered and provided the registration number and name of the registry. Only half of the studies reported where the full trial protocol could be accessed. Seven of the studies addressed funding, reporting where applicable, the sources, role of funders, and if there was any additional support in kind (such as supply of drugs).
Outcomes summary of tDCS for OCD
There was a high variability in electrode montages. While the intensity of stimulation was consistent across studies, the stimulation sites, electrode sizes, electrolyte medium, method of securing the electrodes to the scalp, tDCS devices, and the duration and frequency of stimulation were variable (see Table 3). We note that while one randomized sham-controlled study met criteria for inclusion in this review (Fineberg et al., Reference Fineberg, Cinosi, Smith, Busby, Wellsted, Huneke, Garg, Aslan, Enara, Garner, Gordon, Hall, Meron, Robbins, Wyatt, Pellegrini and Baldwin2023), the authors had explicitly stated that it was a feasibility trial to inform the design of future studies, and to gauge safety, acceptability, and the size of any treatment effect. This trial demonstrated no statistically significant difference between active and sham conditions. Eight further studies involved a sham-controlled design. Of these, Gowda et al. (Reference Gowda, Narayanaswamy, Hazari, Bose, Chhabra, Balachander, Bhaskarapillai, Shivakumar, Venkatasubramanian and Reddy2019) reported a greater reduction in YBOCS scores and a significantly higher response rate in the active condition, with four out of twelve in the active tDCS group improving, compared to no responders in the sham group. Akbari et al. (Reference Akbari, Hassani-Abharian and Tajeri2022) reported a significant difference between the post-test mean scores of OCD symptoms, favoring the active condition (F = 3.56, P < 0.05, η 2 = 0.24). Two trials reported no between group difference at immediate post, or at 6-week post-intervention, however, there was a significantly larger reduction in Y-BOCS in tDCS group compared to the sham at 12 weeks F (1, 84.06) = 84.06, p < 0.03, d = 0.62 moderate effect size (Silva et al., Reference Silva, Brunoni, Goerigk, Batistuzzo, Costa, Diniz, Padberg, D’Urso, Miguel, pedes and Shavitt2021), and F (1, 75) = 4.60, p < 0.035, η 2 = 0.058 medium effect (Harika-Germaneau, Gosez, et al., Reference Harika-Germaneau, Gosez, Bokam, Guillevin, Doolub, Thirioux, Wassouf, Germaneau, Langbour and Jaafari2024). This suggests that perhaps there is a delayed effect of tDCS and highlights the importance of longer follow-up periods. No significant between group (active vs. sham) pre- to post-test differences in Y-BOCS was found in the other four trials. The remaining two studies did not have a true sham-control condition to determine between group differences.
Discussion
This aim of this study was to conduct a systematic review to identify the current research base of RCTs involving tDCS for OCD; to evaluate the quality of the reporting using the CONSORT criteria for the reporting of RCTs of nonpharmacologic treatment, and, to examine the outcomes of tDCS for OCD in RCTs. Eleven RCTs were included in the evaluation. The results indicated low levels of overall compliance with the CONSORT standards highlighting the need for improvement in reporting. It is important to note that research in this field is relatively new, and while none of the studies to date met all of the criteria of the CONSORT, the level of reporting appears to have improved with time. Examination of the outcomes of tDCS for OCD revealed that only two trials found a significant between group (active vs. sham) differences, however, one was limited in terms of the quality of reporting. The trial that showed significant pre- to post-treatment differences in Y-BOCS, and demonstrated a more robust reporting of the trial design, involved anodal stimulation over the pre-SMA and cathodal stimulation over the right supra-orbital area (Gowda et al., Reference Gowda, Narayanaswamy, Hazari, Bose, Chhabra, Balachander, Bhaskarapillai, Shivakumar, Venkatasubramanian and Reddy2019). Despite the paucity of supporting outcomes in well-controlled trials, and the narrow specificity of stimulation sites that have found between group differences favoring active over sham conditions using statistical and clinically significant change methods, favorable reporting of the “promise” of tDCS is common.
The ability to draw any meaningful conclusions about tDCS for OCD, including via a meta-analysis, appears premature, given the variability of stimulation protocols and target sites. These conclusions are further complicated by the methodological limitation of many of the studies. Further issues included small sample sizes (especially in the context of the heterogeneity of OCD), lack of protocol standardization and assessment to ensure adherence to the protocol, and the absence of neuro-navigation for consistent location of targeted stimulation sites between sessions, participants, and trials. We acknowledge that, although ideal, the latter would add significant cost to any trial though and thus may not be feasible. Most trials recruited treatment resistant patients who continued prescribed medication throughout the trial. However, the criteria for what was deemed “dose stability” (i.e., the minimum period of time required between commencing pharmacotherapy and commencing a trial) varied from 2 to 12weeks. As the efficacy of most antidepressants indicated for OCD have been observed to increase by about 1.5 times across weeks 4–12 (Cheng et al., Reference Cheng, Huang, Xu, Li, Li, Shen, Zheng and Li2019), this would need be taken into account when evaluating outcomes. Another limitation was the lack of maintenance and follow-up. tDCS studies in other mental health conditions, such as depression, indicate that maintenance sessions may improve clinical outcomes and duration of effects (Martin et al., Reference Martin, Alonzo, Ho, Player, Mitchell, Sachdev and Loo2013).
In terms of study reporting, areas that were identified as generally well reported included trial designs, recruitment settings, inclusion/exclusion criteria, and blinding. More than half of the trials included a participant flow diagram (as recommended by CONSORT), making it easier for the reader to see whether losses/exclusions occurred after randomization, and the number of participants included in the outcomes analyses. However, the period of time between recruitment and the intervention was not reported. The inclusion of the participant flow is important to identify any potential risk of bias. For example, a loss of numbers after randomization could be due to the participant’s inability to continue due to an exacerbation of symptoms or harm from the treatment. Likewise, if the period of time between recruitment (i.e., inclusion/exclusion measures) and the intervention is not reported, it casts doubt over the validity of the baseline data (i.e., collected no more than 2 weeks prior to commencement of intervention) and the overall result.
The CONSORT criteria of Intervention were reasonably met, with most protocols described at a level that allowed the reader to gain a broad understanding of how a trial took place. However, more detail is required for researchers and clinicians to not only determine the quality and clinical significance of a trial, but also for replication of a trial. For example, electrode type, size, electrolyte medium, and how they are secured to the head all need to be reported. Electrodes deliver the current from the tDCS device to the scalp and can be metal or conductive rubber and can vary in size between highly focal to more dispersed charge. An electrolyte medium (i.e., saline solution, gel, or cream) is used as a buffer between the scalp and the electrodes to prevent skin injury and optimize delivery of current. The electrolyte medium can be placed directly on the electrode or, where electrodes are placed in a sponge case, saline solution is used. The volume of saline solution should be measured and described to ensure consistency and reproducibility of stimulation both within-, and between-participants, and to protect the integrity of the results (i.e., if sponges are over-saturated, saline spreads to an area greater than the sponge and targeted brain site). The method used to affix the electrodes to the scalp should also be reported. Elastic straps are often used, but if these are not tight enough, the electrodes can move during a session and change the distribution of current delivery. If the elastic straps are too tight, there is an increased likelihood of saline solution spread and dissipation of the current across the scalp. To ensure consistency throughout a trial, operators should be trained in standardized tDCS techniques, and adherence to established tDCS protocols should be monitored for internal validity and replicability of results. Only one of the trials addressed adherence to intervention protocol, and this is therefore an area that needs improvement in future trials.
While recruitment settings were generally well reported, it was unclear in the majority of trials where treatment took place, and whether the sample were inpatients or outpatients at the time. Inpatients receive around the clock care, medication compliance is often monitored, the likelihood of exposure to OCD triggers, and everyday life-stressors (i.e., work, managing a household, raising a family etc.) may be minimized, as are factors like travel, finding parking, and attending appointments on time. With the conditions/environment of inpatients being more strictly controlled, the results are less generalizable. Likewise, inadequate reporting of participant demographics (education level, employment status, socio economic status), access to- and affordability of treatment for OCD, and the level of treatment acceptability impacts on the capacity to generalize.
Future directions
To draw any meaningful conclusions about the effectiveness of tDCS for OCD, there is a need for appropriately powered, randomized clinical trials that include a sham condition to explore possible placebo effects. To address limitations properly requires investment to facilitate multicenter collaboration to enhance recruitment. Future studies should also include participants with a larger range/grouping of OCD symptom severity to identify the characteristics of potential responders as possible predictors of response, include maintenance and follow-up periods, and would ideally incorporate neuro-navigation techniques to improve precision of locating stimulation sites.
Conclusion
This systematic review and evaluation of the reporting standards of the literature involving tDCS for OCD revealed low levels of overall compliance with the CONSORT standards highlighting a need for improvement in reporting. Aside from the limitations and lack of generalizability of the results in many of the trials, interpretations were often incongruent with the results, and conclusions contained misleading statements suggesting tDCS is a “promising approach” for OCD. Given the limited robust evidence to suggest that any change in Y-BOCS scores may be anything other than a placebo effect, which is in itself interesting, it is timely to consider the value in continuing to conduct underpowered tDCS trials for OCD. Future researchers must conduct appropriately powered, randomized sham-controlled clinical trials with longer follow-up periods, and reported in accordance with the CONSORT statement, to determine whether tDCS is an efficacious intervention for OCD.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1355617724000602.
Funding statement
This research was supported by an Australian Government Research Training Program (RTP) Scholarship.
Competing interests
The authors have no known conflicts of interest to disclose.
Open access
