Is EMDR an Evidenced-Based Treatment for Depression? A Review of the Literature

PTSD With Comorbid Depression

Numerous studies reporting the use of EMDR for PTSD and phantom limb pain have noted improvements in comorbid depression scores during treatment (Korn & Leeds, 2002; Maxfield & Hyer, 2002; Schneider, Hofmann, Rost, & Shapiro, 2008; van der Kolk et al., 2007); these studies are summarized in Table 1. When treating PTSD with EMDR, comorbid depression improves in almost all cases and EMDR appears to be better at treating this than imaginal exposure (Arabia, Manca, & Solomon, 2011), waiting list controls (Hogberg et al., 2007; Rothbaum, 1997), standard care (Marcus, Marquis, & Sakai, 1997), and fluoxetine or placebo pills (van der Kolk et al., 2007). Other studies have also reported significant improvements in comorbid depression where treatment with EMDR was at least as effective as, but not significantly different from, cognitive behavioral therapy (CBT; Narimani, Ahari, & Rajabi, 2010) or prolonged exposure (Ironson, Freund, Strauss, & Williams, 2002; Lee, Gavriel, Drummond, Richards, & Greenwald, 2002; Rothbaum, Astin, & Marsteller, 2005). Only one RCT found that comorbid depression did not improve when PTSD was treated with EMDR, relaxation, or biofeedback (Silver, Brooks, & Obenchain, 1995). All of the case studies and case series found an improvement in depressive symptoms when PTSD was successfully treated with EMDR (Chemtob, Nakashima, & Carlson, 2002; Korn & Leeds, 2002; Lobenstine & Courtney, 2013; Montefiore, Mallet, Levy, Allilaire, & Pélissolo, 2007; Raboni, Tufik, & Suchecki, 2006; Schneider et al., 2008; Silver, Rogers, & Russell, 2008; Tarquinio, Schmitt, & Tarquinio, 2011; Tarquinio, Schmitt, Tarquinio, Rydberg, & Spitz, 2011). Although each case study on its own is not generalizable because of the nature of the experimental design, the growing number of such reports adds weight to the premise that, in people with PTSD, comorbid depression can be alleviated by treatment with EMDR.

All 10 of the controlled trials, pilot RCTs, and the applied behavior analysis (ABA) controlled study were reviewed with respect to quality assessments using the critical appraisal skills program (CASP), RCT checklist, and the Cochrane Collaboration tool for assessing the risk of bias. The results of this review are summarized in Table 2.

As is often the case in EMDR research, the quality of studies varies. In a review of the controlled treatment outcome studies, Maxfield and Hyer (2002) highlight the need to maintain high standards of research to truly evaluate the EMDR method. This is necessary because there appears to be “a high correlation between methodology and outcome. As methodology became more rigorous, the treatment effect became larger” (Maxfield & Hyer, 2002).

In this review, none of the articles reported power or sample size calculations being done before the study took place to determine recruitment numbers and, as such, many could be argued to be underpowered, although only one (Lee et al., 2002) mentions this as a potential limitation. Sample sizes ranged from 18 (Rothbaum, 1997) to 88 (van der Kolk et al., 2007). Lack of power is not necessarily the case, however; Marcus et al. (1997) had 67 participants over two arms, which, considering the effect sizes expected, appears reasonable. Without a statistical sample size calculation, one cannot be sure, and the standard care treatment in that study was extremely variable. Two studies call themselves pilots (Arabia et al., 2011; Ironson et al., 2002), so here a small sample size is to expected.

Despite a general problem of small sample sizes, the articles by the Arabia, Chemtob, Hogberg, Ironson, van der Kolk, and Lee research groups are high quality. The articles are missing the occasional piece of information such as attrition (Chemtob et al., 2002), equal treatment of the groups (Hogberg et al., 2007), homogeneity of the groups (Ironson et al., 2002), and failure to treat both groups the same during the treatment phase (van der Kolk et al., 2007), but overall, there are few concerns about the reports. As is often the case when psychotherapy is compared to a medication regime, in the study by van der Kolk and colleagues (2007), participants in the different groups received different amounts of input from professionals. Participants in the EMDR arm had eight weekly 90-min sessions, whereas the medication and placebo groups only received 20-min sessions. However, both Rothbaum’s articles suffer from a lack of homogeneity between the groups. In the early article, Rothbaum’s waiting list group was significantly more depressed than the EMDR group and in the 2005 article, the EMDR group had more severe PTSD than the waiting list or prolonged exposure groups. Marcus’ article from 1997 may be poorly written or maybe of low quality. Initial homogeneity of the groups is not reported, neither is attrition, although exclusion criteria are. There is no clearly focused research question reported, assessors were not blind to participant group allocation and the groups were not treated the same. Silver et al.’s (1995) article raises some serious concerns because participants choose their treatment group and there were 83 clients split unequally over four groups, with one group only containing 6 people. There is no clearly focused research question reported and the groups were not treated the same, the EMDR group had as little as one therapy session over a 3-month inpatient stay. The other two treatment groups received at least three sessions. Exclusion criteria, attrition, homogeneity, blinding of assessors, and confidence intervals and standard deviations are also not reported.

To report depression symptoms, seven articles used the Beck Depression Inventory (BDI), one used the Hamilton Rating Scale for Depression (HRSD; Hogberg et al., 2007), one used the Children’s Depression Inventory (CDI; Chemtob et al., 2002), and one did not report the scales used (Silver et al., 1995).

Depression as a Primary Diagnosis

Of the eight journal articles that focused on the treatment of a primary diagnosis of depression, six were case studies that cover various demographics in individuals whose depression ranges widely in severity. Publications regarding the use of EMDR to treat a primary diagnosis of depression are summarized in Table 3.

A report of two case studies using EMDR with Korean adolescents with mild/moderate depression had impressive results (Bae, Kim, & Park, 2008). In Mexico, three people with mild/moderate depression had a full remission of symptoms in an average of 12 sessions. As expected, the negative images, beliefs, emotions, and sensations associated with traumatic events ceased to be distressing and the participants recovered from their depression. EMDR was also reported to have a positive effect on the cognitive processing of emotional material and on conceptual organization in long-term memory storage (Rosas Uribe, López Ramírez, & Jarero Mena, 2010). In the United States, an adult patient with severe depression was treated successfully over 3 months with EMDR (Grey, 2011) and an adult with depression and attention deficit hyperactivity disorder (ADHD) since childhood recovered enough to stop taking his medication for both conditions (Broad & Wheeler, 2006). An elderly Chinese man with depression because of physical health problems and a fear of surgery was successfully treated with an intensive combination of EMDR to address his fear and mindfulness meditation over a 2-week inpatient stay (Sun, Wu, & Chiu, 2004). All these examples focused on acute phase depression and, because they are case studies, could not control for the passage of time or other factors associated with recovery.

Only one study has reported the treatment of a client with chronic depression. Chronic depression is usually treatment-resistant, 10%–20% of adequately treated patients with acute depression will continue to have symptoms for at least 2 years (National Collaborating Centre for Mental Health [NCCMH], 2010), and there is evidence to suggest that some forms of psychotherapy may be less affective at treating this form of the illness (Cuijpers et al., 2010). A severely and chronically depressed woman in India was treated with EMDR after she experienced strong negative reactions to antidepressant medication. After nine treatment sessions, her depression was reported to be “completely cured” and this was maintained at 6-month follow-up (Srivastava & Mukhopadhyay, 2008).

No randomized controlled studies could be found that were published in full English looking at using EMDR with a primary diagnosis of depression. Hogan (2001) conducted a small-scale trial for a doctor of philosophy (PhD) thesis but the full report does not appear to have been published. From the abstract, it appears that in the first four sessions of treatment, adult participants with a diagnosis of depression received one session of either CBT or EMDR. There were only 15 participants per arm and the abstract does not indicate what was done in the other three sessions. There was no statistically significant difference in the responses of the two groups, although it is reported that four of the EMDR group reported “near complete remission of symptoms” and that participants perceived EMDR to be “less negative than CBT” (Hogan, 2001). It is difficult to make a judgment of quality because only the abstract is available for this study, but it makes no mention of randomization methods and the study was probably very underpowered with only 15 people per treatment arm. The method is specific about the amount of therapeutic intervention that both arms received and it seems unlikely that only one session of either treatment would cause large changes in symptoms, especially because many of the case studies report time scales of 3–6 months (Bae et al., 2008; Grey, 2011; Rosas Uribe et al., 2010).

A larger trial was conducted in China comparing clients with depression who were taking sertraline to those who were taking sertraline and had additional treatments of EMDR (Song & Wang, 2007). Here, 64 patients (32 per arm) were randomly allocated to treatment group for a 6-week period. At the end of this period, both groups had improved significantly but there was no significant difference between the two. In earlier weeks in the trial, there was significant difference with the EMDR and sertraline group doing better than the sertraline alone group. The authors conclude that “sertraline combined with EMDR takes effect faster and has higher safety and better compliance in the treatment of depression (Song & Wang, 2007).” However, only the abstract is available in English and adverse effects and compliance issues are not sufficiently addressed. Again, the lack of a full article makes a judgment with regard to the quality of the research difficult.

Several books and book chapters are devoted to how EMDR can be expanded beyond just a treatment for PTSD. These are often written by EMDR professionals using examples from their clinical experience. Five books, containing 10 relevant chapters, have been published between 1998 and 2013 that report examples of EMDR being used to treat depression (Knipe, 2009; Manfield, 1998a, 1998c; Oppenheim, 2009; Servan-Schreiber, 2004; F. Shapiro & Forrest, 2004; R. Shapiro, 2009a, 2009c, 2009d, 2009e). All except one (which describes the process and how it can help depression without the need for medication or other psychotherapies [Servan-Schreiber, 2004]) demonstrate the treatment of depression by clinical example. This is helpful to EMDR practitioners and these chapters provide additional examples of clinical case studies to suggest that EMDR may indeed be an effective treatment option for depression. However, unlike the case studies in journals, the grey literature does not commonly report standardized symptom scales to demonstrate efficacy. It, instead, relies on process measures during EMDR (i.e., subjective units of disturbance [SUDs] and validity of cognition [VOC]). This makes it more difficult to compare the case studies in books to the ones in the peer-reviewed literature. In addition, some of these books fail to give a scientific rationale for the treatment. Others, which do give a rationale, have not been peer-reviewed, so its relevance and accuracy may be in doubt. Knipe (2009) outlines the way the AIP (F. Shapiro & Forrest, 2004; R. M. Solomon & Shapiro, 2008) relates not just to PTSD but also to depression and gives examples of using EMDR to treat shame-based depression. R. Shapiro (2009c) goes further and links the AIP model to the vagus nerve theory. However, this link appears tenuous. Although she uses six publications to portray the vagus nerve theory as a near certainty, only one of these has been peer reviewed. She makes no attempt to explain how AIP, which is in essence a cognitive learning theory, relates to the vagus nerve theory.

At the EMDR UK&I National Workshop in 2012, Hofmann presented work he and his colleagues are currently undertaking in Germany. Hofmann presented a pilot study using a matched pairs design of German inpatients (14 recurrent depression, 12 single episode) compared treatment as usual (TAU—psychodynamic therapy and medication) to TAU plus four or five sessions of EMDR. That study found that the TAU + EMDR group did significantly better than the TAU group. A second pilot study on German outpatients with depression (12 recurrent, 12 single episode) comparing CBT with CBT + five sessions of EMDR is still running. Currently, none of this work has been published (Hofmann, 2012). Although this work has been reported at conferences, the lack of publications in peer-reviewed journals means that tests of quality cannot be undertaken.

The latest project starting in 2012 is the European Depression and EMDR Network (EDEN) RCT, which will involve patients from six European countries with recurrent depression randomly assigned to medication alone, EMDR and medication, or CBT and medication. This study has a recruitment target of 350 participants. The study will use the standard EMDR protocol and will measure improvement using the BDI II (Beck, Steer, & Brown, 1996) and social functioning scores. Clients will also be followed up at 5 years to look at long-term gains and relapse prevention.

The present author is currently engaged in an intensive clinical replication series treating people with long-term depression with EMDR. The Sheffield EMDR and Depression Investigation (SEDI) aims to see if these clients respond to EMDR not only with an improvement in depressive symptoms but also in social functioning. The study will also investigate whether the participants respond in the same ways that PTSD patients have responded, with changes in memory narrative, heart rate variability, and skin conductance response. Finally, SEDI will interview participants to elicit information about the experience of having EMDR for depression (Wood, 2012).