Yoga as a pre-treatment of EMDR to treat childhood abuse-related PTSD: feasibility and pilot study

ABSTRACT

Background: Yoga, a mind-body practice that enhances emotional regulation and self-awareness, may be an effective stabilisation tool for trauma-focused therapy, particularly for patients with childhood abuse-related posttraumatic stress disorder (CA-PTSD), who exhibit strong avoidance strategies.

Objective: This study aimed to test the efficacy of group yoga sessions followed by individual EMDR sessions on PTSD symptoms. Secondary objectives included evaluating the efficacy of this combined treatment on the most common associated comorbidities, and measuring the adherence and effectiveness of EMDR sessions.

Methods: 32 adult women with CA-PTSD were randomly assigned to either a Yoga + EMDR group (10 weekly yoga sessions followed by 10 weekly EMDR sessions) or a Wait + EMDR group (12 weeks of waiting followed by 10 weekly EMDR sessions). Assessments were conducted at baseline, week 12, and week 24, primarily using validated self-report questionnaires.

Results: While PTSD symptoms and most comorbidities showed no significant group differences, the Yoga + EMDR group exhibited significant improvements in anxiety and emotional dysregulation, fewer dropouts, and more effective EMDR sessions compared to the Wait + EMDR group. The study faced a limitation as the overall therapy duration extended beyond the protocol, rapidly preventing new recruitments.

Conclusions: Yoga preceding EMDR shows promise in enhancing emotion regulation and EMDR effectiveness for CA-PTSD patients.

HIGHLIGHTS

• Prior group yoga sessions enhanced the adherence and efficacy of individual EMDR sessions.

• Participants who received yoga sessions demonstrated significant improvements in emotional regulation and reduced anxiety compared to participants who did not.

• The prolonged nature of EMDR therapy for CA-PTSD can strain the resources of medical centres.

• Integrating yoga as both a pre- and co-treatment may sustainably improve EMDR efficacy and ultimately shorten therapy duration.

1. Introduction

1.1. Specificities of CA-PTSD

Childhood physical and sexual abuse are significant risk factors for PTSD (Cougle et al., 2010; Kessler et al., 2017). Childhood abuse-related PTSD (CA-PTSD) is associated with more severe and persistent consequences than adult-onset PTSD, including difficulties in emotional regulation, interpersonal relationships, and lower self-esteem. These challenges can hinder access to therapy or limit its efficacy (Cloitre et al., 2005; Gekker et al., 2018; Gilbert et al., 2009; Messman-Moore & Bhuptani, 2017; Norman et al., 2012).

1.2. Trauma-focused therapy to desensitise traumatic memories

The standard treatment for PTSD is trauma-focused therapy (TFT), in which patients are voluntarily exposed to their traumatic memories (Ehring et al., 2014; Hamblen et al., 2019; Lewis, Roberts, Andrew, et al., 2020; Mavranezouli et al., 2020). This exposure is essential for memory remodelling and desensitisation, provided that patients feel safe enough to expose themselves to their traumatic memories and related vulnerabilities, which is often facilitated by a strong therapeutic alliance and patients’ ability to tolerate and manage strong emotions. These 2 conditions must be present and reinforced in therapy prior to any exposure to traumatic memory (Bremner et al., 2003; Cloitre et al., 2002, 2011; McGaugh, 1966; Rauch & Foa, 2006; Schroyens et al., 2023).

Eye Movement Desensitization and Reprocessing (EMDR), developed by Francine Shapiro (Shapiro, 1995), is an internationally recommended and widely used TFT (APA, 2017; Bisson et al., 2019; NICE, 2018; WHO, 2023). After the essential preparation and stabilisation stages of the EMDR protocol, which can be spread over several sessions, the sessions dedicated to desensitising the traumatic memory form the core of the protocol. They involve mental immersion in traumatic memories and their related sensations while receiving bilateral sensory stimulation to facilitate memory reprocessing. However, as with other TFTs, EMDR exhibits substantial failure and dropout rates, particularly in CA-PTSD (Ehring et al., 2014; Karatzias et al., 2019; Lewis, Roberts, Gibson, et al., 2020; O’Doherty et al., 2023; Wright et al., 2024).

1.3. Therapy challenges in CA-PTSD

Patients with CA-PTSD display a significantly impaired sense of security, leading to a pervasive feeling of danger from external and/or internal sources. Consequently, a significant proportion of these patients exhibit marked experiential avoidance of trauma-related stimuli, reporting intense anxiety and distress upon any recollection of the traumatic event, or any anticipation of recalling it. This thereby impedes engagement in exposure-based therapies, and various involuntary avoidance strategies are common during therapy sessions, such as topic avoidance or dissociative symptoms. Otherwise, patients regularly skip or forget to come to sessions, or even abandon therapy completely (Ehring et al., 2014; Karatzias et al., 2019; Lewis, Roberts, Gibson, et al., 2020; O’Doherty et al., 2023; Wright et al., 2024).

In order to overcome these difficulties, phase-based treatments have been designed and tested for over two decades (Cloitre et al., 2002, 2011). The first phase aims to establish a robust therapeutic alliance while enhancing emotional regulation and adaptive coping skills prior to the implementation of trauma-focused interventions. This first phase primarily integrates cognitive and behavioural therapy tools to facilitate the development of emotion regulation strategies, trauma narrative processing, cognitive restructuring, anxiety and stress management, and interpersonal skills.

1.4. Complementary alternative therapies

Practices such as mindfulness meditation and yoga aim to cultivate present-moment self-awareness, and have been found effective for emotional regulation (Campelo et al., 2025; Goldberg, 2022; Lazzarelli et al., 2024). However, the relative immobility of meditation can easily lead to mind-wandering, rumination and distress, particularly in PTSD patients (Farias et al., 2020; Goldberg et al., 2020). In contrast, the sensorimotor engagement inherent in yoga may better sustain attention and facilitate present-moment focus, making it a potentially more accessible and effective preparatory practice for TFT, as supported by preliminary findings in CA-PTSD populations (Davis et al., 2020; Shatrova et al., 2024; Zaccari et al., 2023; Zhu et al., 2022).

Yoga is increasingly recognised as a complementary treatment modality for trauma. Recent literature reviews and meta-analyses highlight the potential efficacy of yoga in the treatment of PTSD and its commonly associated mental comorbidities (Macy et al., 2018; Nejadghaderi et al., 2024; O’Shea et al., 2022; West et al., 2017). Although these results are preliminary, as the number of rigorous studies examining yoga for trauma is still insufficient, this body of research generally concludes that trauma-informed yoga is a useful and safe approach for PTSD, particularly in integrative approaches where it is combined with TFTs.

1.5. Yoga as a pre-treatment of EMDR for CA-PTSD

Our yoga protocol for CA-PTSD followed the recommendations of trauma-informed yoga (Emerson, 2015; Macy et al., 2018). During yoga sessions, we emphasise body awareness through varied and moderately challenging movements and postures. This facilitates the development of focused attention on somatic sensations, providing an accessible anchor to the present moment. It also teaches patients to find their window of tolerance in postural discomfort, which prepares them to also find this window in emotional discomfort, particularly during TFT. Then, the integration of slow, deep breathing in yoga modulates the autonomic nervous system, promoting parasympathetic activation and a state of calm. Moreover, the practice fosters emotional acceptance and self-compassion by encouraging a non-judgmental and self-caring focus, thereby mitigating maladaptive thought patterns. Finally, the group format offers the pragmatic benefits of enhanced motivation, vicarious learning, and cost-effectiveness. We propose that group yoga sessions may offer an accessible approach to enhance emotional regulation and acceptance as well as sustain self-awareness in the present moment, preventing dissociation. These abilities would reduce avoidance of traumatic memories and make TFTs such as EMDR more effective.

1.6. Evaluating yoga as a pre-treatment for EMDR

This study investigated the impact of 10 weekly group yoga sessions followed by 10 weekly individual EMDR therapy sessions in CA-PTSD participants. This dual treatment was compared to a control group receiving 10 weekly individual EMDR sessions after an equivalent waiting period. This article presents the feasibility, acceptability, and preliminary results of this dual treatment. Our primary objective was to assess changes in PTSD symptom severity from baseline to post-treatment. Our Secondary objective was to evaluate changes in symptoms of common comorbidities, including dissociation, depression, anxiety, emotional dysregulation, and attentional deficits. Our third objective was to compare EMDR adherence and effectiveness between groups by measuring dropout rates, the number of participants having started traumatic memory desensitisation, and the mean number of such sessions received by participants. Our last objective was to measure the effect of yoga alone on PTSD and comorbid symptoms by comparing the intermediate period with baseline, and finally, to explore correlations between the effect of yoga and the effectiveness of subsequent EMDR sessions.

We hypothesised that receiving yoga in pre-treatment would lead to greater improvement of PTSD and comorbid symptoms. We also expected enhanced EMDR efficacy, with fewer dropouts, a higher proportion of participants having started traumatic memory desensitisation within the first ten EMDR sessions, and a greater number of such sessions per participant. We also expected to observe PTSD and comorbid symptom improvements by the intermediate period (post-yoga/pre-EMDR), and hypothesised a correlation between such improvements and EMDR efficacy. However, due to premature study termination and a resulting small sample size, correlation analyses were not feasible.

PTSD and comorbid symptoms were assessed using validated self-report questionnaires. PTSD symptoms were additionally assessed using the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) to mitigate the biases induced by self-assessment. Anxiety and emotional dysregulation were evaluated using both self-report questionnaires and physiological measures, specifically serum cortisol levels and heart rate variability, as biomarkers of stress. Sustained attention capacity was also measured using a validated neuropsychological test.

2. Methods

2.1. Study design

The study is a two-arm randomised controlled trial with a 1:1 allocation ratio. The experimental arm (Yoga + EMDR) consisted of 10 weekly group yoga sessions followed by 10 weekly individual EMDR sessions. The control arm (Wait + EMDR) consisted of 12 weeks of waiting followed by 10 weekly EMDR sessions. To anticipate some participant absences while maintaining a target of 10 sessions of yoga as well as EMDR, 12 weekly sessions of yoga and EMDR were planned. Assessments were conducted at baseline, week 12, and week 24, primarily using self-report questionnaires.

All EMDR and yoga sessions were conducted at the psychotherapy centre for psycho-trauma, which is affiliated with the clinical research unit at Ville-Evard public health establishment (France). The study was registered in the French Trial Registry (NCT04431531) and approved by the French Sud-Est IV protection of individuals committee (approval number: 19.02.22.43216).

2.2. Participant selection criteria

Participants were included if they were women aged 18–70 who met the diagnostic criteria for CA-PTSD, using the PTSD Checklist for DSM-5 (PCL-5) with a score ≥33 and a history of childhood sexual and/or physical abuse. This abuse history was assessed at the first interview by the psychiatrist and by the CTQ self-questionnaire (see Table 1). All participants had to provide free, informed, and signed consent.

Table 1.

Characteristics of the questionnaires used in the study.

Name of the questionnaire	Abbreviation	References	No. of Items	Score range	Cronbach's alpha (at baseline in our sample)	Brief description & objective
Clinician-Administered PTSD Scale for DSM-5	CAPS-5	(American Psychiatric Association, 2013)	30	0–80 (Total Severity)	0.914	Gold-standard clinician interview assessing DSM-5 PTSD symptom severity (20 symptoms), onset, duration, distress, and functional impairment. Provides a total severity score.
PTSD Checklist for DSM-5	PCL-5	(American Psychiatric Association, 2013)	20	0–80 (Total Severity)	0.916	Self-report measure mirroring the 20 DSM-5 PTSD symptoms. Assesses symptom severity over the past month, yielding a total score often used for probable PTSD screening.
International Trauma Questionnaire for ICD-11	ITQ	(Cloitre et al., 2018)	18	0–24 (for both PTSD and DSO subscores)	0.885	Self-report measure assessing both PTSD and Complex PTSD (CPTSD) symptoms according to ICD-11 criteria, through Disturbances in Self-Organization (DSO) symptoms, allowing for a nuanced evaluation of trauma-related distress.
Beck Depression Inventory II	BDI-II	(Collet & Cottraux, 1986)	13	0–39	0.921	Self-report questionnaire widely used to assess the severity of depressive symptoms, covering cognitive, affective, somatic, and vegetative aspects. Higher scores indicate greater severity.
Multidimensional Assessment of Interoceptive Awareness	MAIA	(Mehling et al., 2012)	32	0–5 per item (8 subscales)	0.848	Self-report questionnaire assessing multiple dimensions of interoceptive awareness (e.g. noticing bodily sensations, not distracting from them, emotion regulation through body awareness).
Dissociative Experiences Scale	DES	(E. M. Bernstein & Putnam, 1986)	28	0–100 (Frequency)	0.960	Self-report measure assessing the frequency and range of dissociative experiences, including symptoms like amnesia, depersonalisation, derealization, and absorption.
State Trait Anxiety Inventory	STAI	(Spielberger et al., 1970)	40 (2 × 20)	20–80 for the 2 subscales	0.970 (STAI-A) 0.917 (STAI-B)	Self-report measure composed of two subscales: State Anxiety (current feelings) and Trait Anxiety (general disposition), providing insights into both immediate and general anxiety levels.
Difficulties in Emotion Regulation Scale	DERS	(Gratz & Roemer, 2004)	36	36–180	0.891	Comprehensive self-report measure assessing various facets of emotion dysregulation across six subscales (e.g. non-acceptance, impulse control, limited strategies). Higher scores indicate greater difficulties.
Childhood Traumatic Experiences Questionnaire	CTQ	(Bernstein & Fink, 1998)	28	5–125 for the total score (5–25 for each of the 5 subscales)	0.877	Self-report questionnaire used to screen for a history of childhood abuse and neglect across five forms: emotional, physical, and sexual abuse, and emotional and physical neglect.

Participants were excluded if they presented with neurological disorders, pregnancy, severe or unstable somatic pathologies, or severe motor impairments (e.g. paralysis). Exclusion also applied to patients not affiliated with social security or state medical aid (AME), those involuntarily hospitalised in psychiatric care or hospitalised at the request of a third party, those under guardianship or legal protection, those concurrently participating in another biomedical research study, or those not proficient in French. Individuals unable to provide consent were also excluded.

Inclusions took place between October 2020 and September 2023.

2.3. Variables

PTSD symptoms were assessed at baseline, mid-treatment and post-treatment using the PCL-5 and CAPS-5. Validated self-report questionnaires measured depression, dissociation, anxiety, emotion dysregulation, and interoceptive awareness (Table 1). Due to the lack of a validated French translation at the time of designing the study, the International Trauma Questionnaire (ITQ), which is validated by the 11th version of International Classification of diseases (ICD-11) to assess complex PTSD, was not used, even though CA-PTSD is the largest subgroup of complex PTSD (Cloitre et al., 2018; Maercker et al., 2013). However, the DSO (disturbances in self-organisation) sub-score, a recognised feature of complex trauma (Sarr et al., 2024; Seiler et al., 2023), was used.

Blood samples were collected to measure routine physiological variables and cortisol. Heart rate variability (HRV: McCraty, 2022) was also measured to assess stress and emotional dysregulation. The d2 Test of Attention (Brickenkamp & Zillmer, 1998) assessed sustained attention. Socio-demographic, clinical information, and a childhood trauma questionnaire (CTQ) were collected at baseline.

To assess EMDR adherence and effectiveness, we measured the number of: (i) dropouts, (ii) patients who began desensitisation of traumatic memories, and (iii) traumatic memory desensitisation sessions per patient. Given the brevity of the protocol and the complexity of CA-PTSD, we expected very few completely desensitised traumatic memories per patient by the end of the protocol (their therapy would then continue for this purpose). This led us to focus on the number of desensitisation sessions rather than the number of desensitised memories.

2.4. Intervention protocol

The intervention comprised 10 weekly 75-minute closed-group trauma-sensitive yoga sessions (n ≈ 8 women per group, range 4–12) followed by 10 weekly individual EMDR sessions. Closed groups ensured consistent progression for all participants.

Yoga sessions were facilitated by a certified trauma-sensitive yoga instructor who designed a yoga protocol based on two recognised and broadly similar methods: that of Dr Lionel Coudron (France; Coudron & Miéville, 2018) and that of David Emerson (USA; Emerson, 2015). A psychologist served as co-therapist in all yoga sessions. Each yoga session consisted of the same sequence (∼50 min): first gentle movements while seated on a chair, then progressively more demanding movements while standing and then on a yoga mat on the floor. The emphasis is on grounding, body and breathing awareness with repeated invitations for non-judgmental acceptance of sensations. Sessions concluded with relaxation (7 min), brief mindfulness meditation (3 min), and a post-session discussion (10–15 min). Sessions were conducted without hands-on adjustments and with the emphasis on the choice of whether or not to do the suggested postures or their adaptations.

Individual EMDR sessions were delivered by certified EMDR practitioners who had all passed both training levels from the French Institute of EMDR. The therapists were instructed to record the number of traumatic memory desensitisation sessions carried out for each participant, with the explicit instruction that this recording should not lead them to alter the pace of progress of the therapy, in order to avoid any performance bias.

2.5. Measuring HRV coherence ratio

Heart Rate Variability (HRV) and coherence were assessed using the HeartMath emWave2 and Inner Balance devices. The Inner Balance application incorporates an LED photoplethysmography (PPG) sensor, typically attached to the earlobe, to monitor blood volume pulse via light absorption. HRV data were collected over a 2-minute period, a duration previously established as sufficient for reliable coherence ratio assessment (McCraty et al., 2018). Coherence ratio was calculated using the emWave2 software by identifying the maximum power within the 0.04–0.26 Hz frequency band. Peak power was determined by integrating the spectrum within a 0,030 Hz window centred on this maximum peak. The coherence ratio was then computed as (Peak Power / (Total Power − Peak Power)). This nonlinear method provided a robust measure of HRV coherence that allowed for the nonlinear nature of the HRV waveform over time (McCraty, 2022).

2.6. Data analysis

The sample size was calculated based on data from two studies on PTSD patients: (i) Jindani et al. (2015) investigating the effect of yoga on PTSD symptoms, and (ii) McLay et al. (2016) evaluating EMDR efficacy for PTSD. We hypothesised a 10-point mean reduction difference (standard deviation = 12) in PCL scores in the Yoga + EMDR group compared to the Wait + EMDR group. For a two-sided alpha risk of 5% and a power of 90%, 31 patients per group (N = 62) were required to detect this difference. Accounting for an estimated 5% dropout rate, a total of 66 patients (33 per group) should have been included.

However, the final sample comprised 32 (16 per arm, Figure 1) participants due to recruitment limitations (see section 3.3). Due to the high number of dropouts in the Wait + EMDR group, we were unable to carry out an intention to treat but only a per-protocol analysis.

Figure 1.

CONSORT patient flow.

Due to the non-normal distribution of the data, the Mann-Whitney U test was used for all between-group comparisons of changes in variables. Specifically, this test was applied to assess the change in measures between baseline (W0) and post-treatment (W24) for our primary objective (and objectives #2 and #3). The same statistical test was also used for between-group comparisons of changes in variables between baseline (W0) and mid-treatment (W12) for objective #4. When between-group comparisons of changes were significant, the Wilcoxon signed-rank test was used to assess changes over time within each group. Fisher’s exact test was used for between-group comparisons of changes in categorical variables when chi-square assumptions were violated (McHugh, 2013). Since non-parametric tests were employed, the effect size was estimated using the r correlation coefficient, calculated from the Z value of the test statistic divided by the square root of the total sample size (r = Z/N) (Rosenthal, 1991). All analyses were conducted using SPSS version 23.0.0.0, with a significance level of p < .05 (two-tailed). We applied a Holm-Bonferroni correction for multiple comparisons (Holm, 1979).

3. Results

3.1. Participant flow

As detailed in Figure 1, 41 of 73 initially assessed women were excluded (non-meeting criteria, non-response, declined participation), leaving 32 who consented. A variable delay (days to 5 months) between consent and randomization (due to closed group formation) resulted in 5 further exclusions at baseline (2 no longer met PCL-5 criteria, 2 with disabling dissociative symptoms hindering assessment, 1 lost to follow-up). Consequently, 14 patients commenced yoga and 13 entered the 12-week waitlist. During this period, 2 dropouts occurred in each group. Subsequently, 12 patients proceeded to EMDR after yoga, and 11 after waiting. Notably, no dropouts occurred in the Yoga + EMDR group during the 12-week EMDR phase, compared to 4 dropouts in the Wait + EMDR group (p < .05, detailed in Section 3.6). The final sample completing the protocol comprised 19 patients (12 Yoga + EMDR, 7 Wait + EMDR).

3.2. Sample characteristics

The mean age of participants was 36,9 ± 12,3 years (range: 19–59). At baseline, none had prior yoga or EMDR experience, reported physical limitations to yoga practice, or lacked French fluency. As expected, the mean Childhood Trauma Questionnaire (CTQ1) z-score was significantly elevated (3,47 ± 1,82; (Scher et al., 2001)), indicating a history of childhood trauma. Regarding education, 66,7% (n = 18) held a university degree (bachelor's or master's), and 29,6% (n = 8) were unemployed, including 2 permanently disabled individuals.

3.3. Feasibility and acceptability

Protocol feasibility faced limitations as the centre reached capacity after the first cohorts initiated EMDR (three years post-implementation), with subsequent therapist caseload saturation due to extended CA-PTSD follow-up. Given constraints on resources, this prevented further recruitment.

Intervention acceptability was high. Yoga + EMDR showed robust adherence with 2 dropouts out of 14 (both due to participants having to take university exams). The 14 participants averaged 8.3 ± 2.7 yoga sessions (rising to 9.2 ± 1.4 for the 12 completers). No yoga-related adverse events were reported, though initial sessions were challenging (regarding non-judgmental self-observation and the final relaxation period triggering intrusive thoughts). Participant feedback on yoga was overwhelmingly positive.

Wait + EMDR had a low dropout rate during the waiting phase (2/13; one suicide attempt, one debilitating dissociation). However, EMDR phase dropout was higher (4/11), resulting in 7 completers, consistent with known TFT attrition. As expected, several participants reported distress during trauma questionnaire completion and EMDR desensitisation to traumatic memory, particularly during initial exposures.

3.4. PTSD symptom evolution

Contrary to expectations, both groups showed improvement in PCL-5, CAPS-5, and ITQ DSO scores from baseline to post-treatment, with no significant between-group differences (Table 2).

Table 2.

Comparison between-group of PTSD symptoms evolution between baseline and post-treatment.

	Descriptives						Comparisons
	Baseline (W0)			Post-treatment (W24)			W24 – W0
	n	Mean (SD)	Median (Min; Max)	n	Mean (SD)	Median (Min; Max)	U	p	r
PCL-5
Yoga + EMDR	14	59.1 (11.4)	61 (39; 76)	12	38 (17.8)	32 (14; 66)	50	.885	−0.03
Wait + EMDR	13	57.9 (8.0)	60 (44; 69)	8	37 (17.0)	38 (13; 59)
CAPS-5
Yoga + EMDR	14	49.6 (14.8)	50 (27; 70)	12	36.8 (13.4)	31 (18; 59)	46.5	.662	−0.03
Wait + EMDR	12	45.1 (12.3)	47 (29; 71)	7	40.7 (18.2)	46 (15; 58)
DSO (ITQ)
Yoga + EMDR	14	17.2 (5.4)	19 (5; 24)	12	11.7 (6.7)	12 (3; 21)	41	.587	−0.12
Wait + EMDR	13	17.9 (3.3)	18 (10; 23)	8	12.4 (6.1)	14 (3; 21)

Abbreviations: PCL-5, PTSD Checklist for DSM-5; CAPS-5, Clinician-Administered PTSD Scale for DSM-5; DSO (ITQ), Disturbances in Self-Organization of the International Trauma Questionnaire for CIM-11; SD, Standard Deviation; U value of the Mann-Whitney test; the p-value for significance; p < .05 are considered significant; the r value is a correlation coefficient to estimate the effect size.

While the Yoga + EMDR group showed a larger improvement in CAPS-5 scores (baseline: 49.6 ± 14.8 vs post-treatment: 36.8 ± 13.4) compared to Wait + EMDR (baseline: 45.1 ± 12.3 vs post-treatment: 40.7 ± 18.2; U = 46.5, p = .662), this difference was not statistically significant.

Baseline CAPS-5 scores were notably lower than PCL-5 scores in both groups, consistent with prior observations (Resick et al., 2023; Zaccari et al., 2023).

The lack of significant between-group differences in PTSD symptom evolution is likely attributable to the small sample size (see Discussion).

3.5. Comorbid symptom evolution

Both groups demonstrated improvements in dissociation (DES) and depression (BDI-II) from baseline to post-treatment, with no significant between-group differences (Table 3).

Table 3.

Comparison between-group of comorbid symptoms evolution between baseline and post-treatment.

	Descriptives						Comparisons
	Baseline (W0)			Post-treatment (W24)			W24 – W0
	n	Mean (SD)	Median (Min; Max)	n	Mean (SD)	Median (Min; Max)	U	p	r
Clinical measures:
DES
Yoga + EMDR	14	38.1 (25.0)	33 (3; 88)	12	21.5 (15.2)	21 (6; 43)	40.5	.772	−0.07
Wait + EMDR	13	38 (20.2)	38 (9; 64)	10	24.5 (16.5)	21 (1; 45)
BDI-II
Yoga + EMDR	14	18.5 (10.4)	16 (3; 37)	12	10.8 (6.8)	12 (0; 23)	44	.473	−0.16
Wait + EMDR	13	21.3 (6.7)	23 (9; 31)	9	16.7 (6.3)	18 (7; 25)
MAIA-SR
Yoga + EMDR	14	2.29 (1.9)	2.38 (0.25; 3.75)	12	2.69 (1.4)	2.5 (1.25; 3,5)	35.5	.123	−0.33
Wait + EMDR	13	2.28 (1.2)	2.13 (1.25; 3.25)	9	2.13 (1.4)	2.1 (0.5; 4.25)
STAI-A
Yoga + EMDR	14	64.7 (13.4)	68.5 (30; 40)	12	50.7 (13.8)	52 (29; 74)	16	0,007*	−0.59
Wait + EMDR	13	64 (14.8)	71 (39; 80)	9	61.6 (13.2)	66 (39; 76)
DERS
Yoga + EMDR	14	118 (22.3)	121 (82; 150)	12	103.5 (21.1)	108 (75; 137)	22.5	0,049*	−0.44
Wait + EMDR	13	109.3 (16.1)	110 (89; 134)	8	114.4 (20.8)	114 (82; 151)
Physiological measures:
HRV-CR
Yoga + EMDR	9	0.9 (0.3)	0.8 (0.6; 1.3)	8	1.2 (0.5)	0.9 (0.6; 1.8)	9	.338	−0.29
Wait + EMDR	10	1.3 (0.6)	1.4 (0.4; 2.1)	4	1.1 (0.4)	1 (0.6; 1.6)
Cortisol (nmol/L)
Yoga + EMDR	8	324 (156)	313 (91; 616)	6	351 (170)	312 (214; 687)	21	.329	−0.24
Wait + EMDR	6	424 (108)	376 (306; 568)	4	358 (119)	313 (275; 533)
Cognitive measure:
D2 KL score
Yoga + EMDR	14	54.8 (28.8)	44.1 (13.6; 95.5)	12	74.5 (20.3)	70.9 (46; 99.9)	35	.554	−0.13
Wait + EMDR	13	42.5 (29.3)	42.1 (2.3; 97.7)	7	51.4 (35.9)	50 (8.1; 97.7)

Abbreviations: DES, Dissociative Experiences Scale; BDI-II, Beck Depression Inventory II; MAIA-SR, Self-Regulation subscale of the Multidimensional Assessment of Interoceptive Awareness; STAI-A, state anxiety scale of the State Trait Anxiety Inventory; DERS, Difficulties in Emotion Regulation Scale; HRV-CR, Coherence Ratio of the Heart Rate Variability; KL, concentration level score of the D2 test of attention; SD, Standard Deviation; U value of the Mann-Whitney test; the p-value for significance; *p < .05 are considered significant; the r value is a correlation coefficient to estimate the effect size.

The Yoga + EMDR group showed significant improvements between baseline and post-treatment compared to the Wait + EMDR group in state anxiety assessed with the STAI-A, as well as in emotion regulation assessed with the DERS (U = 16, p = .007 and U = 22.5, p = .049, respectively; see Table 3). The estimated effect sizes were large and medium-large respectively (r = −0.59 for STAI-A and r = 0.44 for DERS). A Wilcoxon test for within-group changes indicated that only the Yoga + EMDR group exhibited significant differences on these two scores between baseline and post-treatment (Yoga + EMDR: p = .002 for STAI-A, p = .010 for DERS; Wait + EMDR: p = .258 for STAI-A, p = .906 for DERS).

Contrary to expectations, no post-treatment difference was found in MAIA scores (although significant regarding the self-regulation sub-score at mid-treatment: see Section 3.7, Supplemental Table 2). No significant between-group differences were observed in cortisol levels (Table 3), potentially due to high inter-individual variability, nor in HRV coherence ratio at post-treatment (though significant at mid-treatment, see Supplemental Table 2). Sustained attention (d2 KL score) improved over time in both groups without significant between-group differences (Table 3).

3.6. EMDR adherence and effectiveness

As previously stated, all 12 Yoga + EMDR participants completed their subsequent EMDR sessions, while only 7 of 11 Wait + EMDR participants did so and this difference was statistically significant (Fisher's test, p = .041). Yoga + EMDR participants completed a mean of 9.7 ± 1.3 EMDR sessions, close to the 10-session target. Wait + EMDR participants who started EMDR averaged 8.4 ± 1.8 sessions (9.1 ± 1.7 for completers). Thus, excluding dropouts, EMDR attendance was high in both groups.

EMDR effectiveness differed significantly in the Yoga + EMDR group compared to Wait + EMDR. A higher proportion of participants in the Yoga + EMDR group received at least one desensitisation session within the first 10 EMDR sessions: 10 out of 12 (83.33%) compared to the Wait + EMDR group's 2 out of 11 (18.18%), a difference confirmed by Fisher's exact test (p = .008). Furthermore, the Yoga + EMDR group also showed a higher average number of desensitisation sessions per participant, with 2.8 ± 2.2 sessions compared to the Wait + EMDR group's 1.4 ± 2.5 sessions, with a Mann-Whitney U test indicating a significant difference (U = 27, p = .030) and a large estimated effect size (r = −0.49).

3.7. Intermediate symptom evolution (Yoga alone)

Contrary to expectations and prior yoga PTSD studies (Davis et al., 2020; Shatrova et al., 2024; Zaccari et al., 2023; Zhu et al., 2022), PTSD symptoms showed non-significant decrease at mid-treatment in the Yoga group compared to the Waiting group (Supplemental Table 1), potentially due to small sample size (see Discussion). In addition, more participants in the Waiting group had their medication adjusted compared to the Yoga group (5/13 vs 2/14, non-significant), which might explain the symptom improvements observed in this group.

However, the Yoga group showed significant improvements from baseline to mid-treatment in state anxiety (STAI-A) scores compared to the Waiting group (U = 23.5, p = .009). The estimated effect size was large (r = −0.55). Changes in MAIA self-regulation (MAIA-SR) sub-scores, measuring patient's ability to regulate their psychological distress through body awareness, were also significantly higher for the Yoga group compared to the Waiting group (U = 37, p = .044). The estimated effect size was medium-large (r = −0.55). Finally, HRV coherence ratio (HRV-CR) -one of the physiological markers of emotional regulation capacity significantly- improved at mid-treatment in the Yoga group compared to the Wait group (U = 4, p = .005). The estimated effect size was large (r = −0.72; see Supplemental Table 2). Apart from the MAIA-SR sub-scores, within-group change assessments revealed that only the Yoga + EMDR group exhibited significant differences in these scores at mid-treatment compared to baseline (Yoga: p = .041 for STAI-A, p = .054 for MAIA-SR, p = .027 for HRV-CR; Wait: p = .878 for STAI-A, p = .566 for MAIA-SR, p = .108 for HRV-CR).

Changes in the other scores measured were not significantly different between the two groups over this period (Supplemental Table 2).

4. Discussion

This pilot study offers preliminary support for the integration of yoga as a first phase intervention to enhance the efficacy of trauma-focused therapies, particularly EMDR, in the treatment of CA-PTSD.

Although we did not observe any statistical difference between the two groups with regard to the evolution of trauma symptoms or some of the most common comorbid symptoms, we did obtain statistically significant differences between the groups with regard to anxiety reduction and emotional regulation, as early as the intermediate assessment, measured using validated self-questionnaires (STAI-A, DERS, MAIA) and a physiological marker of stress level, the HRV coherence ratio.

These findings align with prior research demonstrating yoga's positive impact on emotional regulation in individuals with severe childhood trauma (Ali et al., 2023; Laplaud et al., 2023; O’Doherty et al., 2023; Zhu et al., 2022).

In addition, in the Yoga + EMDR group, we found a significant increase in the effectiveness of EMDR sessions, with a significantly greater number of patients having started traumatic memory desensitisation within the initial ten protocol-measured sessions (83% of Yoga + EMDR patients versus 18% of Wait + EMDR). There was also double the number of traumatic memory desensitisation sessions in the Yoga + EMDR group compared to the Wait + EMDR group (3/10 versus 1.5/10). Therefore, the Yoga + EMDR group benefited from an earlier initiation of the core of EMDR therapy, the work on traumatic memory, than the Wait + EMDR group.

Finally, the Yoga + EMDR group had a significantly lower dropout rate compared to the Wait + EMDR group (0% versus 36% respectively). Participant feedback on the yoga sessions was also highly positive, which is consistent with other studies reporting good yoga adherence among trauma survivors (Macy et al., 2018; Nejadghaderi et al., 2024; O’Shea et al., 2022; van der Kolk, 2014; West et al., 2017).

These results strengthen the value of phase-based treatments for Complex PTSD, as recommended by international guidelines (Bisson et al., 2019; Cloitre et al., 2002, 2011). Our study expands on these models by incorporating a somatic element into the preparatory phase, as advocated by some CA-PTSD experts (Macy et al., 2018; Nejadghaderi et al., 2024; O’Shea et al., 2022; van der Kolk, 2014; West et al., 2017), whereas traditional phase-based treatments use more cognitive and behavioural therapy tools. However, the objectives remain the same: developing emotional regulation and grounding strategies to bolster patients’ readiness for intensive trauma work.

Yoga, particularly trauma-sensitive yoga, appears to fulfil this function through postural exercises combined with slow, deep breathing and non-judgmental self-awareness, which help patients to remain grounded, to accept and regulate their emotions, thereby increasing their tolerance of distress. As expected, this seems to facilitate the patient's exposure to their traumatic memory, which is the core of trauma-focused therapy (TFT). Therefore, integrating yoga as a stabilisation tool seems particularly well-suited to overcome the marked emotional dysregulation and experiential avoidance often observed in CA-PTSD patients, which are prominent barriers to therapy engagement (Cloitre et al., 2002, 2011; Ehring et al., 2014; Herman, 2015; Karatzias et al., 2019; Lewis, Roberts, Gibson, et al., 2020; O’Doherty et al., 2023; Wright et al., 2024).

4.1. Limitations and future directions

The primary limitation is the small sample size due to the particularly long duration of CA-PTSD TFT, which extended well beyond the six months of the research protocol and which, combined with the small capacity of our centre, led to the saturation of therapist slots three years after implementation and therefore prevented the recruitment of new participants.

This lack of significant differences in the evolution of PTSD symptoms and most comorbidities between groups could be due to our small sample rather than the ineffectiveness of the therapies or the characteristics of our sample population, since studies of yoga and/or TFT, with larger samples (n > 60) and a similar population, have found significant improvements in all of these symptoms (Bayley et al., 2022; Davis et al., 2023; McLay et al., 2016; McLean et al., 2022; Shatrova et al., 2024; Yi et al., 2022; Zaccari et al., 2023; Zhu et al., 2022). In addition, the clinical variability inherent in post-traumatic stress disorder, which is reflected in the wide range of cortisol levels measured and in the results of certain questionnaires (e.g. the ITQ DSO sub-score), requires larger sample sizes in order to detect statistically significant effects.

Another important limitation is the short duration of both the therapies and the monitoring of their effects, with only 10 sessions of yoga and 10 sessions of EMDR measured in the protocol (EMDR therapy continued outside the protocol for the patients) and the last evaluation carried out immediately at the end of the 10th EMDR session and no longer afterwards. This prevents us from determining the long-term sustainability of the enhanced EMDR effectiveness observed with yoga pretreatment, and whether it can ultimately reduce the overall duration of EMDR therapy. Future research should address these limitations through larger multicenter studies, with longer-term therapies and follow-up. Such studies could also explore incorporating yoga as a co-treatment alongside EMDR to further increase its effectiveness and potentially shorten overall therapy duration.

The use of closed yoga groups was convenient to ensure consistent progression for all participants, but it also presented a limitation, as the time needed to set up a group led to loss to follow-up. Implementing multiple weekly yoga groups at varying levels could mitigate this issue by allowing for more flexible participant inclusion.

5. Conclusions

This study provides preliminary evidence that group yoga sessions preceding individual EMDR therapy may enhance emotional regulation and improve EMDR adherence and effectiveness in CA-PTSD patients. While the study faced limitations, the findings suggest that yoga holds promise as a valuable first-phase intervention for EMDR in this population. Further research is warranted to confirm these findings and explore the optimal integration of yoga into EMDR treatment protocols.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

DOI:.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/20008066.2025.2530917.