The influence of identity-oriented psychotrauma therapy on Hashimoto disease activity: a randomised controlled trial

ABSTRACT

Background: Childhood adversity and trauma have been linked to altered thyroid function and hypothyroidism, yet few randomised controlled trials (RCTs) have examined the effectiveness of trauma-focused psychological interventions in patients with Hashimoto’s thyroiditis (HT).

Objective: This study aimed to evaluate the effectiveness of Identity-Oriented Psychotrauma Therapy (IOPT) in adults with HT, focusing on both immunological and psychological outcomes.

Method: This parallel-group RCT randomised 70 adults with HT (93.94% women; age 25–57) to receive treatment as usual (TAU) or TAU plus IOPT. The intervention comprised 10 bimonthly group sessions, one session every two weeks. Outcomes included thyroid peroxidase (TPO-ab) and thyroglobulin (TG-ab) antibodies, along with depression, anxiety, stress, dissociation, alexithymia, and anger, assessed at baseline, post-treatment, and 3-month follow-up. A series of 2-way mixed-model ANOVAs and Bonferroni-corrected post-hoc tests were used. Missing data due to attrition – particularly in the control group (37%) – were addressed via multiple imputation, and sensitivity analyses were conducted to test robustness.

Results: Initial analyses suggested that IOPT significantly reduced TPO-ab levels compared to TAU (p < .001), with effects maintained at follow-up (p = .01); however, these effects were not replicated in complete case analysis. IOPT led to significant improvements in dissociation (p = .03), alexithymia (p < .001), depression (p < .001), anxiety (p < .001), stress (p = .015), state anger (p = .009), anger-in (p = .009), and quality of life (p = .042) at follow-up.

Conclusion: These preliminary findings suggest that IOPT, when added to standard medical care, may offer psychological benefits for HT patients and could potentially influence immunological outcomes. However, given the attrition-related sensitivity of some results, further replication in larger, well-retained samples is necessary before drawing definitive conclusions.

Trial registration: ClinicalTrials.gov identifier: NCT04600349.

HIGHLIGHTS

• Over 90% of Hashimoto’s patients reported early-life trauma.

• IOPT may influence thyroid autoantibodies, though findings were not robust.

• IOPT plus standard care may reduce psychological symptoms more effectively than standard care alone.

Hashimoto thyroiditis (HT) is a chronic autoimmune disease in which the immune system attacks thyroid tissue, leading to progressive thyroid dysfunction (Mincer & Jialal, 2022). The condition is marked by elevated anti-thyroglobulin (TG-ab) and anti-thyroid peroxidase (TPO-ab) antibodies and is the most prevalent autoimmune disorder, disproportionately affecting women (Cooper & Biondi, 2012). Despite the availability of hormone replacement therapies, patients frequently experience psychological symptoms – such as depression, anxiety, fatigue, and cognitive disturbances – which are often underdiagnosed or insufficiently treated (Samuels, 2014; Siegmann et al., 2018).

Accumulating evidence indicates that early-life adversity, including childhood trauma, is linked to dysregulation of the hypothalamic-pituitary-thyroid (HPT) axis and may contribute to autoimmune processes (Marques-Feixa et al., 2021; Plaza et al., 2010). Psychological mechanisms such as alexithymia and dissociation, both frequently rooted in unresolved trauma, have been identified as mediators between early stress and immune dysfunction (Macarenco et al., 2021). These factors are particularly relevant in HT, where a significant proportion of patients report high levels of emotional numbing and affective disconnection.

Although few randomised trials have been conducted on psychological interventions for HT specifically, two studies have shown that psychological treatment can positively influence both mental health and thyroid-related biomarkers. Markomanolaki et al. (2019) reported that a stress management programme significantly improved thyroid antibodies, TSH levels, and psychological symptoms including stress, depression, and anxiety in women with HT. Similarly, Portokalidou et al. (2022) demonstrated that psychological support for HT patients improved emotional distress and health-related quality of life. Beyond HT, studies in other autoimmune conditions such as rheumatoid arthritis, systemic lupus erythematosus, and multiple sclerosis have shown that stress-reduction interventions (e.g. mindfulness-based stress reduction, Cognitive Behavioural Therapy (CBT), or supportive therapy) may improve mood symptoms, quality of life, and, in some cases, disease activity markers (Greco et al., 2004; Zautra et al., 2008). However, most of these interventions focus on coping and emotional regulation, rather than addressing trauma-specific processes such as dissociation and identity fragmentation. Considering this gap in the literature, the present study investigates the effectiveness of Identity-Oriented Psychotrauma Therapy (IOPT) in adult HT patients.

To our knowledge, no prior RCT has directly assessed a trauma-specific therapeutic intervention in an autoimmune population with elevated dissociative symptoms and alexithymia. Therefore, this study contributes a novel psychotherapeutic angle within a field that has traditionally emphasised physiological and behavioural mechanisms.

IOPT is a trauma-focused therapeutic method developed by Franz Ruppert, based on the premise that early relational trauma leads to identity fragmentation and enduring emotional dysregulation. Central to IOPT is the concept of the split psyche, comprising a healthy self, a traumatised part, and a survival self. Through structured sessions centred on a self-chosen therapeutic intention (e.g. ‘I want to be healthy’), clients explore unconscious dynamics related to trauma using interpersonal resonance and embodied awareness. Recent work suggests that such processes may activate neurobiological synchrony and emotional integration (Kinreich et al., 2017; Müller, 2022). IOPT addresses the structural effects of early relational trauma on identity, affect regulation, and dissociative functioning. Given that many HT patients exhibit alexithymia, dissociation, and fragmented emotional processing – often undetected in standardised diagnostic interviews – IOPT offers a tailored, non-directive therapeutic frame focused on internal conflict resolution rather than symptom exposure or behavioural modification. This makes it particularly suitable for autoimmune populations with complex trauma histories and emotion regulation deficits.

To our knowledge, no previous randomised controlled trial has evaluated a trauma-specific therapeutic intervention in an autoimmune population characterised by elevated dissociative symptoms and alexithymia. While IOPT has shown preliminary promise in treating individuals with chronic pain, dermatological, and rheumatologic conditions, its relevance for autoimmune thyroid disease remains unexplored. The present study is the first to systematically assess the feasibility, acceptability, and potential clinical efficacy of IOPT as an adjunct to treatment-as-usual (TAU) in patients with HT, offering a novel psychotherapeutic approach within a field historically dominated by physiological and behavioural paradigms.

The present study

This prospective RCT has a parallel arm design with two conditions: one control group receiving treatment as usual (TAU) and one experimental group receiving the IOPT in addition to TAU, following the guidelines for RCTs (Boutron et al., 2008). This study evaluates whether IOPT decreases the level of autoantibodies in HT patients effectively and reduces psychological symptoms associated with HT. The null hypothesis is that IOPT has no significant effect on auto antibodies levels or on the psychological dimensions. Rejecting the null hypothesis (alpha < .05) supports the alternative hypotheses, that IOPT decreases the level of autoantibodies and improves depression, anxiety, stress, dissociation, alexithymia, anger, and quality of life. We hypothesised greater therapeutic gains in the experimental condition, sustained at 3-months follow-up. This is the first RCT that integrates a trauma-focused therapy into HT treatment.

Materials and methods

Transparency and openness

Our study protocol was approved by the University of Bucharest Committee of Ethics in Research (39/27.02.2019) following the CONSORT guidelines for parallel group randomised trials. All participants signed a written informed consent before participation in the study.

Participants and procedure

The study included adults with a confirmed HT diagnosis, recruited via a Facebook ad from March to September 2019. Participants responded by emailing the principal investigator, who then requested medical documents to confirm the HT diagnosis. A research assistant randomly allocated the participants to one of the two conditions using a 1:1 sequence generated by the software Randomiser.org and placed the randomisation results in sealed envelopes. Seventy outpatients aged 25–57 years (M = 41.29, SD = 7.95), were randomised to either TAU or IOPT plus TAU.

Masking of outcome assessors

During the study, an endocrinologist blinded to the two conditions assessed the participants’ biological markers at baseline, post-treatment, and follow-up.

Eligibility criteria

Inclusion criteria were as follows: (1) adults with a confirmed HT diagnosis; and (2) at least one biological marker of TPO-ab and TG-ab exceeding the reference range.

Exclusion criteria were as follows: (1) psychotic symptoms; (2) current psychological treatment; (3) current psychotropic medication; (4) neurodevelopmental or neurocognitive disorders; (5) substance abuse; (6) serious legal or health issues preventing regular attendance, and (7) both biological markers within the reference range.

The IOPT group

The experimental group was further divided into five smaller groups, each led by a clinical psychologist who specialised in IOPT and was trained and supervised by the method’s author, Franz Ruppert. The IOPT method and the protocol were introduced in the first session, after which each participant worked on an Intention related to their current illness. The protocol required participants to attend 10 IOPT modules every two weeks. During each module, every participant spent approximately 50–60 minutes working on their own Intention.

As a trauma-focused method, IOPT centres on the use of an Intention – a short, client-generated sentence that reflects a psychological issue or therapeutic goal. In the first session, each participant was asked to formulate their own Intention, focusing on the illness or related symptoms (e.g. ‘I want to understand my Hashimoto’), and write it visibly on a flip chart or whiteboard. The client then selected group members to resonate with specific words from the Intention – each acting as a resonator for one element. For instance, the client might ask: ‘Can you please resonate with the word ‘Hashimoto’ from my Intention?’ The resonance process began with non-verbal attunement and was followed by verbal sharing, through which resonators reported spontaneous bodily sensations, emotions, or mental images. These experiences were used to explore the client’s unconscious dynamics, such as internal conflicts, psychological splits, and trauma-related entanglements. By observing the resonators’ responses, clients could gain insight into dissociated or suppressed parts of the self, promoting emotional integration and therapeutic resolution. The session concluded with the client formally releasing each resonator (e.g. ‘Thank you for resonating with the word ‘Hashimoto’; now you are Christina again’).

Treatment as usual group (TAU)

Participants in the TAU group continued their usual endocrinological care for HT, which included administration of Levothyroxine, as prescribed by their treating physicians. As the psychological intervention was the primary focus of this study, we did not interfere with or systematically monitor individual dosage adjustments during the trial. All participants in the TAU group met the eligibility criteria at baseline, which included the absence of current psychological treatment or psychotropic medication. However, during the study, several TAU participants began psychological therapy or psychiatric medication. These participants were excluded from follow-up analyses, in accordance with the study protocol, as these treatments introduced potential confounding effects. This contributed to a 37.1% attrition rate in the TAU group, which is further addressed in the Limitations section.

Measurements

Measurements were collected at baseline, post-treatment, and 3 months follow-up. Baseline data included participants’ sociodemographic characteristics and trauma history (Table 1). Participants provided medical analyses of TPO-ab and TG-ab at each time point.

Table 1.

Sociodemographic and clinical characteristics of the participants in each treatment group.

	Experimental group N = 31	Control group N = 35	p
AgeM (SD)	40.7 (7.9)	41.7 (8.3)	.572
Age at disease onset M (SD)	32.7 (7.9)	32.1 (9.1)	.760
GenderN (%)			1
Female	29 (93.5)	33 (94.3)
Male	2 (6.5)	2 (5.7)
Socioeconomic status			1
Employed	29 (93.5)	33 (94.3)
Not employed	2 (6.5)	2 (5.9)
Education			.115
High school	0	4 (11.4)
Bachelor’s Degree	19 (61.3)	15 (42.9)
Postgraduate Degree	12 (38.7)	16 (45.7)
Marital status			.572
Married	19 (61.3)	19 (54.3)
Single	3 (9.7)	6 (17.1)
Divorced	4 (12.9)	3 (8.6)
Living with a partner	5 (16.1)	6 (17.1)
Ever undergone counselling or therapy			.642
Yes	15 (48.4)	20 (57.1)
No	16 (51.6)	15 (42.9)
Duration of counselling or therapy			.192
None	16 (51.6)	15 (42.9)
Several sessions	2 (6.5)	8 (22.9)
Several months	6 (19.4)	6 (17.1)
Several years	8 (25.8)	6 (17.1)
Trauma typeN (%)
Emotional neglect	24 (77.4)	30 (80)	.527
Non-verbal emotional abuse	29 (93.6)	33 (91.4)	1
Parental physical maltreatment	23 (74.2)	19 (57.1)	.193
Parental verbal abuse	24 (77.4)	25 (71.4)	.582
Peer emotional abuse	17 (54.8)	28 (82.9)	.030
Peer physical bullying	19 (61.3)	22 (62.9)	.802
Physical neglect	10 (32.3)	13 (37.14)	1
Sexual abuse	17 (54.8)	20 (60)	.805
Witnessing interparental violence	15 (48.4)	15 (45.7)	.805
Witnessing violence to siblings	12 (38.7)	4 (11.4)	.019
Reported multiple trauma	29 (93.6)	32 (91.4)	.961
Participants monitoring
Complete ORS/SRS	23 (74.19) / 22 (70.97)	–	–
Premature termination ORS/SRS	7 (22.58) / 6 (19.35)	–	–
Missing data	1 (3.23) / 3 (9.68)	–	–

Trauma history was assessed using the 52-item Maltreatment Abuse and Exposure Scale (MAES), a shorter version of the Maltreatment and Abuse Chronology of Exposure Scale (MACE; Teicher & Parigger, 2015), translated into Romanian by an authorised translator. The translation process involved forward translation by a bilingual expert, backward translation by another independent translator, and reconciliation of discrepancies with the input of a clinical psychologist. We selected this innovative measure for its suitability in research settings and robust psychometric qualities. It allows simultaneous evaluation of childhood adversity severity and multiplicity, reflecting the variety of events. Both MAES and MACE have been used in various cultural contexts, including Romania (Oltean et al., 2022) and have demonstrated strong psychometric properties across different populations. The MAES assess exposure to 10 types of abuse, including emotional neglect (Cronbach’s $\alpha \mathrm{in}\mathrm{this}\mathrm{sample}$ = .7); non-verbal emotional abuse ($\alpha$ = .56); parental physical maltreatment ($\alpha$ = .76) parental verbal abuse ($\alpha$ = .78) peer emotional abuse ($\alpha$ = .73) peer physical bullying ($\alpha$ = .72); physical neglect ($\alpha$ = .59); sexual abuse ($\alpha$ = .64), and witnessing inter-parental violence ($\alpha$ = .78). A sample item is ‘Before 18 years old, a parent, stepparents or other adults living in the house acted in a way that made you afraid that you might be physically hurt’.

The Romanian version (Morariu et al., 2013) of The Toronto Alexithymia Scale – 20 (TAS-20) (Parker et al., 1993) was used to assess alexithymia. The questionnaire is rated on a 5-point Likert scale and a sample item is: ‘It is difficult for me to find the right words to describe my feelings.’. The TAS-20 uses cutoff scoring: ≤ 51 = non-alexithymia, ≥ 61 = alexithymia, 52–60 = possible alexithymia. The Cronbach’s alpha was .86 in this sample.

Dissociation was measured using the Romanian version (Curșeu, 2006) of the 28-item Dissociative Experiences Scale (DES) (Carlson & Putnam, 1993), a self-report tool for assessing dissociation in clinical and nonclinical populations. A sample item is: ‘Some people have the experience of finding themselves in a place and having no idea how they got there’. Respondents estimate the percentage of time these experiences occur using an 11-point visual analogue scale ranging from 0% to 100%. The total score is the sum of all items. In this sample, the scale has excellent test-retest reliability (r = .93), and a high internal consistency (Cronbach’s $\alpha$ =  .88).

Anger was measured using the Romanian version (Pitariu & Iliescu, 2006) of The State-Trait Anger Expression Inventory (STAXI-2) (Spielberger, 2010), a 57-item tool covering: state anger ($\alpha$ = .96) trait anger ($\alpha$ = .88), anger control out ($\alpha$ = .76), and anger control in ($\alpha$ = .7). The scale is rated on a 4-point Likert scale and a sample item is: ‘I tend to wear spades without telling anyone’.

The Depression Anxiety Stress Scales (DASS-21) (Lovibond & Lovibond, 1995) measured perceptions of depression ($\alpha$ = .82), anxiety ($\alpha$ = .8), and stress ($\alpha$ = .88) levels. The Romanian version of the DASS-21 was utilised in this study. The scale is a self-reporting questionnaire using a 4-point scale and a sample item is: ‘I was unable to become enthusiastic about anything’. Participants rated how much each item applied to them over the past week, from 0 = ‘did not apply to me at all’ to 3 = ‘applied to me very much, or most of the time.’ Higher the scores indicate greater emotional distress.

The Romanian version of the 26-item The World Health Organisation Quality of Life (WHOQOL-BREF version; WHO, 1995) was used to assess quality of life. WHOQOL-BREF is a self-report measure covering physical health, psychological health, social relationships, and environment. A typical question is ‘How would you rate your quality of life?’ Participants rate each question on a 5-point scale (1 = very poor, 5 = very good). Higher scores indicate better quality of life. The global QOL has excellent internal consistency in this sample ($\alpha$ .91).

Every module, The Outcome Rating Scale and Session Rating scales ORS-SRS (Miller et al., 2003) were administrated. ORS-SRS are very brief, feasible measures for tracking client well-being and the quality of the therapeutic alliance. The ORS consists of a four-item visual analogue scale measuring the client’s experience of well-being in his or her individual, interpersonal, and social functioning. Like the ORS, the SRS is a visual analogue scale that assesses four interacting elements, such as the quality of the relational bond, the degree of agreement between the client and therapists on the goals, methods, and overall therapy approach. Participants were asked to fill in the ORS at the beginning of each module and the SRS at the end of each module.

Data analysis

G-Power 3.1 was used a priori to determine sample size, indicating that 30 participants per condition would be sufficient to detect changes at 5% significance with 80% power for both immunological and psychological variables. To account for potential dropout, 5 additional participants were randomly assigned to each group, resulting in 35 participants per group.

Subsequently, data was analyzed using R Software v. 4.2.2. Group differences in demographic characteristics and trauma types were assessed using independent samples t tests for continuous variables (e.g. age and age at onset), Exact Fisher’s test for categorical variables with 2 × 2 contingency table (e.g. gender, socioeconomic status, ever undergone therapy, and trauma type), and permutation tests for 2 × 3 or 2 × 4 contingency tables (e.g. education, marital status, and duration of therapy), conducted via the coin package (Version 1.4.3). Permutation tests employ resampling techniques to generate null distributions, providing a robust approach for smaller sample sizes. Missing data imputation was performed prior to statistical analyses using the mice package (Version 3.15.0), applying the Predictive Mean Matching method (PMM). Consequently, a series of a 2-way mixed model ANOVAs, adjusted for potential violations of assumptions (e.g. sphericity) and bootstrapped by 1000 samples, were conducted to analyse the effects of group, measurement time point, and their interaction. Mauchly’s test for sphericity, as well as the Greenhous-Geisser correction for violations of this assumption, were performed using the rstatix package in R (Version 0.7.2).

The analyses used a 2 (group: IOPT vs. TAU) × 2/3 (time: baseline, post-treatment, follow-up) design to assess treatment effects over time on the following dependent variables: TPO-ab and TG-ab levels, depression, anxiety, stress, dissociation, alexithymia, quality of life and anger. To account for potential outliers and deviations of normality, a robust 2-way mixed ANOVA based on 5% trimmed means was also performed using the bwtrim function from the WRS2 package (Version 1.1.4). This approach provides Q-values as test statistics for main effects and interactions, reflecting robust comparisons between groups and time points under trimmed mean assumptions. These analyses offer additional confidence in results when standard ANOVA assumption may be compromised.

Post-hoc analyses with Bonferroni correction were applied to determine the specific nature of significant differences. Effect sizes for standard ANOVAs were reported using generalised eta squared (ges), where values of .02, .13, and .26 + represent small, moderate, and large effects. Significance was determined at p < .05, with all analyses assuming a 95% confidence interval.

Results

Participants’ characteristics

Sociodemographic and clinical characteristics are presented in Table 1. Sixty-six participants (6.06% males and 93.94% females), aged 25–57 years (M = 41.24, SD = 8.03) were assessed. Most were employed (78.78%), married or living with a partner (74.25%) and 93% had a university or post-university degree. Approximately 54.55% had undergone counselling or therapy: 15.15% for a few sessions, 18.18% for several months and 21.21% for years. Regarding the trauma history, 93.55% of the experimental group and 91.43% of the control group reported multiple exposures, from emotional and physical neglect to sexual abuse and witnessing violence. There was no significant age difference between the experimental and control groups. Similarly, no significant differences were found in gender, socioeconomic status, education level, marital status, duration of therapy, prior experience with therapy, or most trauma types, except for peer emotional abuse and witnessing violence to siblings.

Attrition

A total of 70 participants were randomised (35 per group). Four participants from the IOPT group withdrew before baseline assessment, resulting in 66 participants with baseline data (35 in the TAU group and 31 in the IOPT group). At post-treatment, 12 participants from the TAU group and 6 from the IOPT group were lost to follow-up due to failure to return biological or psychological data or initiating psychological treatment (TAU group). One additional participant from each group was lost at the 3-month follow-up. Consequently, 24 participants (22 TAU, 2 IOPT) did not complete the full protocol, and complete case data across all three time points were available for 46 participants (22 TAU, 24 IOPT). Despite this attrition, multiple imputation was applied to address missing data, and all primary analyses were conducted on the full randomised sample (N = 66) with baseline data. Participant flow and exclusion criteria are detailed in Figure 1.

Figure 1.

CONSORT flow diagram showing subject allocation to the therapy conditions.

Outcomes

Descriptive statistics and comparison of group and time point effects for immunological and psychological variables are presented in Table 2, while the main effects of the mixed ANOVA results, including between-subjects, within subjects, and interaction effects (e.g. group x moment) are shown in Table 3. Results revealed no significant baseline differences between groups.

Table 2.

Descriptive statistics and comparison of group and time effects on immunological and psychological variables.

Variable	Time	Group	N	Min	Max	Median	Mean	SD	Group p	Time p
Anti-TPO	Baseline	C	35.0	3.0	1,912.0	312.6	472.6	477.0	.288	.650
		E	31.0	3.8	1,344.7	235.0	305.0	293.8		.004
	Post-Treatment	C	35.0	31.5	1,549.7	423.0	504.1	401.4	.001
		E	31.0	10.0	377.0	149.0	135.7	96.3
	Follow-up	C	35.0	12.1	1,912.0	423.0	582.5	457.0	.001
		E	31.0	18.0	280.5	94.3	126.1	88.7
Anti-TG	Baseline	C	35.0	1.0	482.0	58.6	141.3	160.1	1	.038
		E	31.0	1.2	419.0	71.0	126.2	135.9		.446
	Post-Treatment	C	35.0	1.0	454.0	53.2	111.0	137.5	.99
		E	31.0	1.6	453.0	71.0	146.1	153.9
	Follow-up	C	35.0	3.9	361.0	42.9	71.3	78.8	.414
		E	31.0	1.0	454.0	26.1	96.8	115.2
Alexithymia	Baseline	C	35.0	25.0	77.0	51.0	49.8	13.7	1	1
		E	31.0	31.0	76.0	48.0	47.6	11.0		.032
	Post-Treatment	C	35.0	25.0	69.0	52.0	47.3	12.5	1
		E	31.0	29.0	69.0	44.0	45.6	10.3
	Follow-up	C	35.0	27.0	75.0	53.0	50.7	12.9	.001
		E	31.0	30.0	67.0	38.0	40.2	9.3
Depression	Baseline	C	35.0	0.0	13.0	5.0	6.2	3.7	1	1
		E	31.0	1.0	14.0	7.0	7.00	3.7		.001
	Post-Treatment	C	35.0	1.0	20.0	6.0	6.5	4.1	1
		E	31.0	0.0	15.0	6.0	6.2	4.1
	Follow-up	C	35.0	0.0	20.0	6.0	6.7	4.3	.001
		E	31.0	1.0	10.0	2.0	3.3	2.4
Anxiety	Baseline	C	35.0	1.0	14.0	6.0	6.6	3.9	1	1
		E	31.0	0.0	13.0	8.0	7.0	4.0		.001
	Post-Treatment	C	35.0	1.0	18.0	5.0	6.3	4.1	1
		E	31.0	0.0	13.0	7.0	6.8	3.4
	Follow-up	C	35.0	1.0	18.0	6.0	7.0	3.9	.001
		E	31.0	0.0	9.0	3.0	3.7	2.7
Stress	Baseline	C	35.0	3.0	18.0	10.0	10.8	3.8	1	.766
		E	31.0	2.0	19.0	12.0	11.0	4.3		.034
	Post-Treatment	C	35.0	2.0	19.0	9.0	9.6	4.4	1
		E	31.0	3.0	16.0	10.0	10.4	3.6
	Follow-up	C	35.0	2.0	19.0	11.0	10.8	3.7	.015
		E	31.0	3.0	18.0	8.0	8.4	3.1
DES	Baseline	C	35.0	1.8	27.5	10.0	10.5	6.6	.546	1
		E	31.0	2.5	31.1	12.9	13.0	8.2		.026
	Post-Treatment	C	35.0	1.8	23.9	11.8	11.1	6.2	1
		E	31.0	2.0	27.1	10.7	10.7	5.7
	Follow-up	C	35.0	2.0	23.2	11.4	11.0	5.5	.072
		E	31.0	1.7	18.2	6.8	8.0	5.1
Anger State	Baseline	C	35.0	10.0	36.0	12.0	16.4	8.1	1	.240
		E	31.0	10.0	32.0	12.0	15.9	6.5		.001
	Post-Treatment	C	35.0	10.0	35.0	14.0	14.5	5.1	.051
		E	31.0	10.0	25.0	11.0	11.9	2.9
	Follow-up	C	35.0	10.0	31.0	14.0	14.1	4.3	.027
		E	31.0	10.0	24.0	11.0	11.6	2.8
Anger Trait	Baseline	C	35.0	15.0	38.0	23.0	24.4	5.6	.336	.768
		E	31.0	16.0	34.0	27.0	26.6	5.3		.044
	Post-Treatment	C	35.0	16.0	38.0	22.0	23.8	5.4	1
		E	31.0	16.0	37.0	23.0	25.0	5.1
	Follow-up	C	35.0	17.0	36.0	22.0	23.4	4.1	1
		E	31.0	16.0	36.0	23.0	24.1	5.0
Anger Out	Baseline	C	35.0	15.0	36.0	25.0	25.1	5.4	.720	.956
		E	31.0	17.0	37.0	27.0	26.6	5.5		1
	Post-Treatment	C	35.0	17.0	31.0	24.0	24.1	4.1	.255
		E	31.0	17.0	38.0	25.0	26.0	4.5
	Follow-up	C	35.0	16.0	40.0	24.0	24.4	4.5	.522
		E	31.0	17.0	40.0	24.0	26.1	5.5
Anger In	Baseline	C	35.0	10.0	21.0	15.0	14.9	2.8	1	.164
		E	31.0	9.0	22.0	15.0	15.0	3.1		.034
	Post-Treatment	C	35.0	11.0	20.0	14.0	14.7	2.5	1
		E	31.0	10.0	21.0	15.0	15.2	3.1
	Follow-up	C	35.0	11.0	23.0	16.0	15.6	2.9	.009
		E	31.0	10.0	20.0	13.0	13.5	2.5
Anger Control	Baseline	C	35.0	7.0	28.0	20.0	19.3	5.5	1	.402
		E	31.0	11.0	27.0	18.0	18.7	4.2		.984
	Post-Treatment	C	35.0	14.0	28.0	20.0	20.7	4.1	.432
		E	31.0	9.0	26.0	19.0	19.2	3.9
	Follow-up	C	35.0	12.0	28.0	20.0	20.5	4.2	1
		E	31.0	9.0	28.0	20.0	19.6	3.9
Quality of life	Baseline	C	35.0	60.0	112.0	94.0	91.0	13.0	1	1
		E	31.0	56.0	113.0	87.0	88.6	14.2		.056
	Post-Treatment	C	35.0	69.0	115.0	94.0	92.6	10.4	1
		E	31.0	65.0	115.0	95.0	93.5	12.3
	Follow-up	C	35.0	55.0	113.0	93.0	89.7	12.9	.04
		E	31.0	69.0	116.0	100.0	97.7	12.7

Note. E = Experimental; C = Control; Group = Group Differences; Time = Measurement Moment Differences.

Table 3.

Main effects of mixed ANOVA: between-subjects, within-subjects, and interaction effects.

Variable	Effect	dfn	dfd	F	p	ges	Q (df b/w)
Anti-TPO	Group	1	64	23.4	<.001	0.19	23.21 (1, 39.7)*
	Moment	1.6	99.0	1.2	=0.31	0.01	1.3 (2, 40.8)
	Group × Moment	1.55	99.0	5.3	=0.001	0.03	4.05 (2, 40.8)*
Anti-TG	Group	1	64	23.4	<0.001	0.19	0.46 (1, 60.5)
	Moment	1.6	99.0	1.2	=0.31	0.01	3.63 (2, 56.1)*
	Group × Moment	1.6	99.0	5.3	=0.01	0.03	0.92 (2, 56.1)
Alexithymia	Group	1	64	5.3	=0.02	0.04	5.88 (1, 63.6)*
	Moment	1.7	108.1	1.8	=0.17	0.01	1.85 (2, 48.9)
	Group × Moment	1.7	108.1	3.9	=0.03	0.03	7.06 (2, 48.9)*
Depression	Group	1	64	1.6	=0.21	0.02	1.54 (1, 62.8)
	Moment	2	128	7.5	<.001	0.03	6.81 (2, 50.2)*
	Group × Moment	2	128	12.1	<.001	0.05	9.70 (2, 50.2)*
Anxiety	Group	1	63	0.5	=0.48	0.01	0.66 (1, 63.1)
	Moment	2	126	8.8	<.001	0.04	5.30 (2, 51)*
	Group × Moment	2	126	14.9	<.001	0.06	6.56 (2, 51)*
Stress	Group	1	64	0.4	=0.5	0.00	0.39 (1, 63.8)
	Moment	2	128	4.3	0.02	0.02	4.71 (2, 54.0)*
	Group × Moment	2	128	6.9	0.03	0.03	10.16 (2, 54.0)*
Dissociation	Group	1	64	0.1	=0.81	0.00	0.08 (1, 62.7)
	Moment	1.8	116.2	4.5	=0.01	0.02	3.95 (2, 56.3)*
	Group × Moment	1.8	116.2	6.6	0.003	0.03	5.92 (2, 56.3)*
Anger State	Group	1	58	1.6	=0.21	0.01	3.62 (1, 59.4)
	Moment	1.1	63.3	15.3	<.001	0.10	6.57 (2, 43.9)*
	Group × Moment	1.1	63.3	1.6	=0.21	0.01	0.62 (2, 43.9)
Anger Trait	Group	1	64	1.6	=0.21	0.02	1.82 (1, 63.5)
	Moment	1.8	112.2	5.3	=0.009	0.02	4.64 (2, 55.2)*
	Group × Moment	1.8	112.2	1.0	=0.38	0.00	0.90 (2, 55.2)
Anger Out	Group	1	63	4.2	=0.04	0.04	3.88 (1, 59.8)*
	Moment	2	126	1.1	=0.34	0.01	0.95 (2, 53.7)
	Group × Moment	2	126	0.1	=0.93	0.00	0.03 (2, 53.7)
Anger In	Group	1	64	0.7	=0.39	0.01	0.69 (1, 64.0)
	Moment	2	128	0.8	=0.46	0.00	1.09 (2, 50.4)
	Group × Moment	2	128	7.2	<.001	0.04	8.53 (2, 50.4)*
Anger Control	Group	1	64	1.3	=0.25	0.01	1.25 (1, 62.8)
	Moment	1.8	114.2	2.1	=0.13	0.01	1.60 (2, 55.2)
	Group × Moment	1.78	114.2	0.3	=0.70	0.00	0.20 (2, 55.2)
Quality Of Life	Group	1	64	0.8	=0.36	0.01	0.84 (1, 57.4)
	Moment	2	128	2.8	=0.06	0.02	1.99 (2, 54.0)
	Group × Moment	2	128	4.5	=0.01	0.03	3.46 (2, 54.0)*

Note. group = between-subjects effect, moment = within-subjects effect, dfn = degrees of freedom numerator, dfd = degrees of freedom denominator, ges = generalised eta-squared, Q (df b/w) = robust analysis (degrees of freedom between/within); *p < 0.05.

Analyses of primary outcomes at post-treatment revealed a significant decrease in TPO-ab for IOPT vs TAU (F _{(1, 64)} = 24.8, padj < .001, $\eta$²_ges = .28), with gains largely maintained at 3-months follow-up (F _{(1, 64)} = 29.9, p < .001, $\eta$²_ges = .32). In the IOPT group, significant differences in TPO-ab were observed across the measurement time points (F _{(1.06, 31.7)} = 10.9, padj = .004, $\eta$²_ges = .17), being significantly decreased at post-treatment (padj = .002) and follow-up (padj = .001) compared to baseline. Significant differences for TG-ab were found in the TAU group across measurement time points (F _{(2, 66)} = 4.23, padj = .038, $\eta$²_ges = .06), with a decrease observed at follow-up compared to baseline (padj = .036), but no effect relative to the IOPT group.

For secondary outcomes, participants receiving IOPT reported significant reductions in state anger at post-treatment for IOPT vs control group (F _{(1, 64)} = 5.97, padj = .051, $\eta$ ² _ges = .10), with gains largely maintained at 3-months follow-up (F _{(1, 64)} = 7.35, padj = .027, $\eta$²_ges = .10). Participants assigned to IOPT also reported substantial reductions in dissociation (F _{(1.81, 116.16)} = 6.63, p = .03, $\eta$ ²_ges = .03), alexithymia (F _{(1, 64)} = 13.9, padj = .001, $\eta$ ² _ges = .18), depression (F _{(1, 64)} = 15.8, padj < .001, $\eta$ ² _ges = .19), anxiety (F _{(1, 64)} = 15.3, padj < .001, $\eta$ ² _ges = .19) and stress (F _{(1, 64)} = 8.33, padj = .015, $\eta$ ² _ges = .115) at follow-up relative to the TAU group. Significant IOPT vs TAU effects were observed at follow-up for anger in (F _{(1, 64)} = 9.44, padj = .009, $\eta$ ² _ges = .13) but not for trait anger (F _{(1, 64)} = 0.39, padj = 1, $\eta$ ² _ges = .01), anger control out (F _(1,64) = .83, padj = 1, $\eta$ ² _ges = .01) or anger out (F _(1,64) = 1.89, padj = .522, $\eta$² _ges = .029). The IOPT group also reported significant effects in quality of life at follow-up (F _{(1, 64)} = 6.42, p = .042, $\eta$² _ges = .091) relative to the control group. No significant differences were observed between the measurement time points within the TAU group for any measures. In contrast, the IOPT group showed significant changes across the measurement time points (e.g. baseline, post-treatment, follow-up) for dissociation (F _{(2, 90)} = 4.58, padj = .026, $\eta$²_ges = .09), alexithymia (F _{(2, 90)} = 4.36, padj = .032, $\eta$²_ges = .09), depression (F _{(2, 90)} = 9.73, padj = <.001, $\eta$ ²_ges = .18), anxiety (F _{(2, 90)} = 9.30, padj = <.001, $\eta$²_ges = .17), stress (F _{(2, 90)} = 4.25, padj = .034, $\eta$²_ges = .09), state anger (F _{(1.63, 49)} = 13.4, padj = <.001, $\eta$²_ges = .17), anger trait (F _{(1.44, 43.2)} = 4.78, padj = .044, $\eta$²_ges = .04), anger in (F _{(2, 60)} = 4.37, padj = .034, $\eta$²_ges = .07), but not anger control out (F _{(2, 60)} = .72, padj = .984, $\eta$²_ges = .01), or anger out (F _{(2, 60)} = .24, padj = 1, $\eta$²_ges = .01).

Post-hoc results

The IOPT group reported significant reductions at post-treatment compared to baseline for anger state (padj = .002) and at follow-up compared to baseline for depression (padj = <.001), anxiety (padj = <.001), stress (padj = .019), dissociation (padj = .010), and anger state (padj = <.001). The IOPT group also reported significant increase of quality of life at follow-up compared to baseline (padj = .023). Moreover, significant reductions were reported between post-treatment and follow-up for depression (padj = .004) and anxiety (padj = .001).

ORS/SRS results

The analysis of ORS and SRS revealed an overall upward trend across all IOPT sessions (range: 5.76–8.90 and 8.81–9.95, respectively), indicating consistent improvement in both the participants’ well-being and satisfaction. No notable external events or significant distress factors were observed to impact the participants’ experience throughout the study. Data imputation was not performed due to the subjectivity of the data, as imputing could have introduced bias. Analyses were conducted only on participants with complete data across all sessions (see Figures 2 and 3 for a detailed representation).

Figure 2.

Evolution of ORS scores across IOPT sessions (N = 23).

Figure 3.

Evolution of SRS scores across IOPT sessions (N = 22).

Sensitivity analysis

To assess the robustness of the findings, a sensitivity analysis was conducted using complete case data. The overall pattern of results was largely consistent with those obtained through multiple imputation. However, several differences emerged. Notably, the previously observed between-group effects for TPO-ab at post-treatment and follow-up were no longer statistically significant, and no within-group change was found for TPO-ab in the IOPT group. In contrast, Anger Control, which was not significant in the imputed model, showed a significant between-group difference at follow-up (p = .012) and a significant within-group improvement in the IOPT group (p = .001). These findings suggest that while the general trends are stable, some outcomes may be sensitive to analytic assumptions and should be interpreted with caution.

Taken together, the primary outcomes, post-hoc results, and sensitivity analysis provide a complex picture of the intervention’s impact – suggesting promising effects for several psychological and immunological variables, while also highlighting areas where results are contingent on data assumptions. These issues are discussed in more depth below.

Discussion

Summary of findings and theoretical implications

This RCT implemented a 5-month IOPT intervention in HT patients to assess its effectiveness on the thyroid autoantibodies, mental health, and quality of life. The primary objective was to assess whether IOPT leads to reductions in the levels of thyroid autoantibodies, specifically thyroglobulin antibodies effectively, TG-ab, and TPO-ab. The second purpose was to assess whether IOPT reduces depression, anxiety, stress, dissociation, alexithymia, and anger, while improving participants’ quality of life.

First, initial analyses indicated that IOPT significantly decreased TPO-ab levels in HT patients at post-treatment, with effects maintained at 3-month follow-up. This is a relevant finding, as elevated TPO-ab levels are associated with poorer physical and psychological well-being and may predict negative future health outcomes (Müssig et al., 2012). The pattern is consistent with prior studies showing that psychological interventions can reduce markers of autoimmune activity (Markomanolaki et al., 2019; Oppermann-Schmid et al., 2022). However, in the sensitivity analysis using complete case data, these effects were no longer statistically significant, suggesting that changes in TPO-ab may be sensitive to the method used for handling missing data. This limitation should temper interpretation and underscores the need for replication in larger and more stable samples.

By contrast, although TG-ab levels showed a numerical decline post-treatment in the IOPT group, this effect was not statistically significant. One possible explanation is that TG-ab may be a more stable or less treatment-sensitive immunological marker, requiring a longer intervention period to elicit significant changes. This interpretation aligns with psychoneuroimmunology frameworks suggesting that immune system recalibration – particularly for chronic, low-grade autoimmune responses – may lag behind psychological improvements and require sustained stress reduction (Danese & Lewis, 2017; Slavich & Irwin, 2014). Additionally, studies suggest that reductions in inflammatory biomarkers typically follow extended emotional processing and affect regulation (Lopresti, 2017). Thus, future trials might explore whether increased therapeutic duration or intensity enhances the likelihood of significant TG-ab reduction.

Second, the findings show a significant reduction of dissociation and alexithymia in the IOPT group at follow-up. Previous research (Terock et al., 2018; Vonderlin et al., 2018) strongly associates traumatic experiences to the aetiology of dissociation and alexithymia. The high prevalence of multiple early trauma, dissociation, and alexithymia in our sample supports the idea that alexithymia develops in the context of childhood abuse, consistent with recent research (Zorzella et al., 2020). These findings also show that reducing alexithymia and dissociation scores following trauma therapy may impact biological markers, reinforcing the view that these defense mechanisms, rather than trauma itself, contribute to the development of autoimmune diseases, including HT (Macarenco et al., 2021).

Also, this research shows significant treatment effects on anxiety and depression symptoms in the experimental group, consistent with recent studies of stress management and trauma therapy for anxiety and depression in HT (Markomanolaki et al., 2019) and multiple sclerosis (Wallis & de Vries, 2020). Our findings also found significant stress reduction and improved quality of life in the IOPT group at follow-up, supporting previous research showing that trauma treatments can alleviate distress and enhance quality of life in chronically ill patients by reducing negative emotions and promoting positive ones (Murray, 2016; Portigliatti Pomeri et al., 2020).

Moreover, these findings show a significant decrease in state anger in the IOPT group at post-treatment, maintaining at follow-up, supporting the Wilson and colleagues (2001), who found that trauma treatment lowers anger measures. Further, significant IOPT vs TAU effects were observed at follow-up for anger-in, meaning patients were more aware of their anger, finding more healthy ways to express it. These results support previous studies showing that anger-in, the tendency to recognise but suppress anger (Spielberger, 2010), is linked to higher scores in dermatologic autoimmune conditions (Gulec et al., 2009), but also in other autoimmune conditions (Russell et al., 2016), including HT (Pankowski et al., 2021). Anger levels have also been associated with treatment effectiveness (Bunevicius & Prange, 2000).

Contrary to the expectations, the most significant secondary outcome measures were observed at follow-up, rather than post-treatment. This suggests that HT patients undergoing trauma-focused therapy, may require additional time to consolidate emotional and cognitive changes initiated during treatment. One contributing factor may be the structure of the intervention: participants worked on one Intention every two weeks over a 5-month period, which allowed for gradual integration but may also have delayed symptom resolution. In addition, the delayed improvement pattern may reflect the so-called sleeper effect – a well-documented phenomenon in psychotherapy whereby therapeutic benefits become more pronounced after treatment ends, as internal processing continues (Flückiger & Del Re, 2017). Moreover, symptom reduction in trauma therapy often occurs non-linearly, in the form of sudden gains (Tang & DeRubeis, 1999) which have been observed across disorders, age groups, and settings (Shalom & Aderka, 2020). Taken together, these factors may account for the lag between treatment completion and measurable psychological improvement.

In conclusion, this pilot study provides preliminary evidence that IOPT may contribute to reductions in thyroid autoantibodies in HT patients, particularly TPO-ab, with effects maintained at 3-month follow-up in the primary analysis. However, these immunological effects were not confirmed in the sensitivity analysis and should be interpreted with caution. More consistently, IOPT was associated with improvements in psychological symptoms often reported in HT, including depression, anxiety, stress, state anger, and anger suppression. Furthermore, the intervention appeared to reduce dissociation and alexithymia – two trauma-related processes frequently implicated in the pathogenesis of autoimmune disease (Macarenco et al., 2021). Many participants demonstrated greater awareness of traumatic experiences and improved emotional processing, which coincided with gains in quality of life. These findings suggest that, when combined with standard medical treatment, IOPT may offer added value in addressing the complex psychophysiological burden of HT. Replication in larger, more diverse samples is needed to substantiate these preliminary outcomes.

Practical implications

Research on autoimmune diseases is fundamental as their pathogenesis is not fully understood, and current treatments only provide palliative relief. Severe early life stress can lead to long-term consequences for immune dysregulation (Fagundes et al., 2013) and is an independent risk factor for inflammation, marked by increased proinflammatory cytokines in adulthood (Danese et al., 2007, 2011). Autoimmune patients, including HT patients, require mental health services frequently (Kuriya et al., 2020). The current study emphasises the importance of addressing chronic stress from early adversity, which could be valuable for medical doctors and mental health professionals treating HT patients.

The findings of this study suggest that IOPT may have the potential to influence thyroid autoantibodies in patients with Hashimoto’s thyroiditis, particularly TPO-ab, though this effect was not replicated in the sensitivity analysis and should be considered exploratory. According to recent literature (Li et al., 2024), both TPO-ab and TG-ab levels are inversely associated with general health and vitality, and positively linked to inflammatory markers and psychiatric symptom severity. Chronic psychological stress is known to dysregulate the hypothalamic – pituitary – adrenal (HPA) axis, promoting low-grade inflammation. Although there is no conclusive epidemiological evidence directly linking chronic stress to HT onset, it is plausible that reducing emotional stress and enhancing emotion regulation may help modulate immune activity. Through this mechanism, trauma-focused psychotherapy such as IOPT might contribute indirectly to immunological regulation. Previous evidence has shown that such interventions can alleviate physical symptoms associated with chronic pain (Grant & Threlfo, 2002), dermatologic disorders (Gupta & Gupta, 2002), and other physical symptoms linked to adverse life experiences (Shapiro, 2014). However, rigorous trials are needed to clarify whether these effects generalise to HT and under what conditions psychotherapeutic intervention may affect autoimmune parameters.

This study shows that over 90% of the HT participants reported multiple early trauma, ranging from emotional and physical neglect to sexual abuse and witnessing violence. Mental health professionals should therefore screen for childhood abuse, which is often linked to autoimmune disease risk and has significant implications.

These findings also underscore the importance of addressing dissociation and alexithymia in trauma therapy for HT patients who are often survivors of childhood abuse. In our study, IOPT appeared effective in reducing these symptoms, though further research is needed to confirm this in broader clinical populations. The development of autoimmune thyroiditis can be influenced by alexithymia and dissociation, and failing to address these issues can undermine patient motivation and limit treatment effectiveness (Macarenco et al., 2021).

In this study, trauma-focused therapy with IOPT was associated with reductions in anxiety and depressive symptoms commonly observed in HT patients. Given the high prevalence of affective symptoms in this population, routine screening for psychological distress by both physicians and mental health professionals is strongly recommended. Integrating medical and psychological approaches may be important not only for symptom management, but also for improving overall patient functioning and quality of life. In this context, IOPT – through its emphasis on emotional awareness and the processing of unresolved trauma – shows promise as a complementary intervention. By helping patients identify and work through past experiences that may contribute to current symptoms, IOPT may enhance treatment responsiveness and patient engagement. While these findings are preliminary, they offer initial support for the clinical utility of IOPT in HT care and highlight the need for further research in larger and more diverse samples.

In the same time, its implementation in real-world settings presents both opportunities and challenges. One strength of the intervention is its structured format and focus on individual trauma processing, which facilitates personal relevance and therapeutic alliance. However, the treatment requires a degree of emotional readiness and psychological insight that may not be feasible for all patients, particularly those in acute medical distress or with limited access to psychotherapeutic services.

Furthermore, applying IOPT in Consultation-Liaison (CL) contexts demands effective collaboration between medical and mental health professionals. This integration is often hindered by systemic barriers, such as limited communication between disciplines, unclear treatment ownership, or differences in clinical paradigms. Future implementation efforts should explore structured referral pathways, interdisciplinary training, and collaborative case management models to support integration. This would increase the feasibility and accessibility of trauma-informed care in autoimmune populations.

Limitations and future research directions

While this study has significant implications, it also has limitations. First, the sample size was relatively small. Power calculations showed that this sample size could detect group differences, but a larger sample may have yielded more differences between the interventions.

Second, psychological measurements relied on self-report scales which, although well-validated, may reduce reliability due to potential biases such as recall failure, socially desirable responses, and limited self-awareness – particularly in individuals with high levels of alexithymia and dissociation. Future studies should consider complementing self-report instruments with structured or semi-structured clinical interviews. For example, while the DES was used for its feasibility in clinical research, it may not adequately capture the complex phenomenology of dissociative states, especially those embedded in developmental trauma. The use of interviews such as the SCID-D could yield richer, more embodied accounts of dissociative symptoms and better elucidate the connection between trauma-related psychological processes and immune dysregulation. Such an approach would contribute to refining the theoretical model linking dissociation and autoimmune reactivity.

Third, certain subscales of the MAES (e.g. non-verbal emotional abuse and physical neglect) showed relatively low internal consistency ($\alpha$ = .56 and $\alpha$ = .59, respectively), which may limit their interpretability. However, it is important to note that these subscales were used solely for descriptive purposes in assessing trauma history. As they were not included in the primary statistical analyses, they did not impact the main findings of the study. Fourth, therapists blinding was not possible, as in any clinical trial testing psychotherapeutic interventions. However, an endocrinologist blinded to randomisation groups assessed participants’ biological markers at baseline, post-treatment, and follow-up.

Additionally, the ability of individuals to tolerate emotionally demanding trauma treatment and sustain long-term engagement is an important consideration for future studies. In our trial, 19% of participants in the experimental group dropped out before the post-treatment phase. Based on therapist reports and follow-up contact, this attrition appears to have been associated with emotional discomfort during early IOPT sessions – such as distressing emotional reactions, heightened vulnerability, or fear of reactivation. Such responses are frequently observed in trauma-focused psychotherapy and may reflect insufficient psychological readiness or difficulties with affect regulation during the initial phases of treatment (Creswell, 2017; Melville et al., 2010). While these reactions may ultimately be part of the change process, they pose challenges for early retention and require appropriate preparatory strategies in future interventions. In the TAU group, attrition reached 37.1% by follow-up. Most dropouts failed to complete biological assessments or self-report questionnaires. Several also reported initiating psychotherapy during the study, thus violating eligibility criteria and being excluded from follow-up. Although subjective reasons were not systematically collected, this pattern may reflect perceived insufficiency of medical treatment alone or lack of engagement without a psychological component. Previous studies indicate that participants allocated to control groups without active interventions are at greater risk for dropout, especially when expectations for therapeutic benefit are unmet (Brueton et al., 2014). Future research should incorporate structured exit interviews and proactive engagement strategies to improve retention across all study arms.

Moreover, while attrition is common in clinical trials (Dumville et al., 2006) its magnitude and pattern here raise concerns about potential self-selection bias. Several participants in the TAU group began psychological therapy or psychiatric medication during the trial, which disqualified them from continued participation under the original eligibility criteria. These cases were excluded from follow-up analyses. As Schulz and Grimes (2002) note, attrition exceeding 20% warrants concern regarding bias, particularly when drop-out is not random but associated with symptom severity or treatment-seeking behaviour. This may have led to an underestimation of psychological symptoms in the remaining TAU sample, thereby inflating between-group differences. Although multiple imputation was applied to address missing data, this method cannot fully account for systematic attrition. Future studies should consider participant engagement strategies, such as small-value incentives (Abdelazeem et al., 2022) as well as analytic approaches that formally model attrition-related bias.

The study has a limited external validity, meaning that it is not appropriate to apply the findings to patients in other countries without further replications. We need to conduct more studies in different populations to determine whether the response pattern to the treatment is consistent across all populations or whether it varies. Similarly, we cannot generalise the results of the study to other autoimmune diseases. Further randomised controlled trials are needed to reach a conclusive understanding of the IOPT efficacy for different autoimmune diseases.

Finally, we acknowledge the absence of qualitative data as a limitation of the current study. While quantitative outcomes provide important insights into treatment effectiveness, they offer limited insight into the emotional, relational, and contextual factors influencing treatment engagement, adherence, and attrition. Future research should incorporate qualitative methodologies – such as in-depth interviews or focus groups – with treatment completers, dropouts, refusers, and even clinicians. Such data would enable a more nuanced understanding of how participants experience and interpret trauma-focused therapy and could clarify barriers to treatment acceptability and retention. Integrating qualitative insights can strengthen the ecological validity and translatability of psychological interventions for autoimmune populations.

Conclusions

In summary, the findings of this pilot study provide preliminary evidence that IOPT, additionally to TAU, may have beneficial effects for HT patients. Initial analyses suggested that IOPT was associated with reductions in thyroid peroxidase antibodies (TPO-ab), alongside improvements in several psychological outcomes including dissociation, alexithymia, depression, anxiety, stress, and anger regulation. However, sensitivity analyses indicated that the observed immunological effects – particularly for TPO-ab – were not robust across analytic strategies. These findings therefore require cautious interpretation and highlight the need for replication in larger samples with improved retention.

Despite these limitations, the consistent psychological benefits observed at follow-up suggest that trauma-focused interventions such as IOPT may hold promise for addressing the emotional and physiological burden of autoimmune thyroid conditions. This study contributes to an emerging field of inquiry at the intersection of psychotraumatology and immunology and lays the groundwork for future investigations into integrative treatment approaches for autoimmune diseases.

Disclosure statement

No potential conflict of interest was reported by the authors.

Data availability statement

The data that support the findings of this study are openly available in OSF at https://osf.io/c9rhs/?view_only=24c0f0497c8d417daff6dbe35166a597

Ethical approval

Our study protocol was approved by the University of Bucharest Committee of Ethics in Research (39/27.02.2019) following good clinical practice guidelines. All participants signed a written informed consent before participation in the study.