Trauma and personality disorders in Danish treatment-seeking veterans

Abstract

Background & objectives

50.000 Danish soldiers have been deployed abroad during the last three decades. A recent study reported that 24 % were registered with either a psychiatric diagnosis or receiving psychotropic medication. We want 1) to explore the relationship between ICD-11 PTSD and complex PTSD (CPTSD), and severe personality pathology as well as clinical syndromes, and 2) to assess the mediating impact of social support, secondary trauma symptoms, and aggression for the two trauma diagnoses.

Method

142 male war veterans were assessed at intake to the outpatient services of a specialized trauma treatment facility. The participants completed an assessment battery. In addition to descriptive statistics, two multiple regression analyses were performed: one for PTSD and one for CPTSD.

Results

While less than half of the personality disorders and the clinical syndromes were associated with PTSD, almost all personality disorders and all the clinical syndromes were associated with CPTSD and to a much higher degree. 52 % of the variation of PTSD could be explained in the regression analysis with three significant factors in the final model: Anxiety, schizotypal personality disorder, and dissociation. In the similar analysis for CPTSD, 61 % of the variation was explained in the final model with Avoidant personality disorder and Negative Affectivity as significant factors. Social support and aggression were not significant predictors in either analysis.

Conclusion

The study supports the more severe sequelae associated with CPTSD compared to PTSD. The relationships between trauma disorders and personality pathology deserve more attention to inform the treatment of veterans.

Since 1992, the Danish military has deployed approximately 50.000 soldiers to missions in the Balkans, Iraq, and Afghanistan.¹ A registry study on Danish veterans’ mental health highlights the serious consequences of deployment, reporting that 24 % were registered with either a psychiatric diagnosis or receiving psychotropic medication, despite having no such records prior to deployment. Nearly 5 % were diagnosed with post-traumatic stress disorder (PTSD).¹

The PTSD diagnosis is characterized by three core elements: 1) re-experiencing of past traumas in the present, accompanied by feelings of horror or fear; 2) avoidance of trauma-related reminders; and 3) a persistent sense of threat, often manifested as excessive hypervigilance. Research shows that prevalence and severity of PTSD symptoms tend to increase over time² and some soldiers are at risk of developing complex, long-lasting psychiatric conditions, that are difficult to treat.³

The inclusion of complex post-traumatic stress disorder (CPTSD) in the 11th revision of the ICD-11⁴ provides an opportunity to screen for persistent and pervasive disturbances in self-organization (DSO) in addition to the symptoms of PTSD.⁵ A CPTSD diagnosis requires at least one symptom in each of the three self-organization domains: affect, negative self-concept, and relational disturbance. These symptoms can result in problems with emotion regulation and anger, a diminished sense of self-worth, and difficulties in maintaining social relationships.⁴

Studies have found CPTSD to be both more common³ and more debilitating than PTSD.^6,7 Veterans with CPTSD tend to be slower to seek help,⁷ have higher levels of functional impairment, a greater likelihood of being single, divorced, or widowed and are more likely to use psychotropic medication.⁶ Furthermore, the group experiences a higher number of comorbid mental health issues, including dissociation, anger, social isolation, and impaired functioning.⁷ The distinction in ICD-11 between PTSD and CPTSD has been found to be clinically relevant in a sample of Danish treatment-seeking military veterans.⁶

Research has examined the risk factors and comorbidities associated with ICD-11 PTSD and CPTSD across diverse trauma-exposed populations. Evidence points to early and persistent traumatic experiences in childhood and adulthood, as contributing to a heightened risk of developing CPTSD.^5,8,9 In a Danish sample, those with CPTSD were more likely than those with PTSD or no trauma disorder to exhibit symptoms of depression, anxiety, stress, and suicide attempts.³ Alcohol and substance abuse are often comorbid with CPTSD.¹⁰ More generally, experiencing injury or witnessing a comrade’s death increases the probability of subsequent contact with the psychiatric system.¹¹ These findings emphasize that veterans are a high-risk group for developing serious mental health challenges following deployment.

In 2010, Denmark introduced its first veteran policy, focusing on the mental health consequences of deployment.¹ This policy defines veteran support to include time before, during and after deployment granting veterans lifelong access to treatment and support. As part of this initiative, three highly specialized veteran treatment facilities were established within the Danish psychiatric system, to treat individuals suffering from severe psychological disorders.

The present study investigates the mental health characteristics of veterans in a treatment-seeking sample, using the same population as^{12(blinded for review)}. This sample offers valuable insights into the mental health status of veterans an average of 14.2 years (SD = 7.99) after their last deployment.

Beyond comorbidity, little is known about whether personality pathology and clinical syndromes are differentially and specifically associated with the core components of PTSD versus the DSO components of CPTSD. We want first to examine relationships between demographic and deployment-related variables on the one side and PTSD, CPTSD, and other indicators of mental health and wellbeing on the other side. Our second aim is to explore the relationships between ICD-11 PTSD and CPTSD, and severe personality pathology as well as clinical syndromes using the ITQ and the MCMI-III. These findings might contribute to improving treatment strategies for veterans experiencing chronic conditions several years post-deployment.

Methods

Participants

Data was collected among a group of Danish male war veterans (N = 142), at intake to the outpatient services of a specialized trauma treatment facility (ATT), in the Region of Southern Denmark, between September 2017 and May 2021. All participants were asked to complete a self-administered, paper-based questionnaire, that assessed their current symptoms, trauma history, psychotropic medication use, and demographic details. Each participant received information about the study and provided informed consent. A small number of participants were excluded for various reasons. The study has been approved by the University Data Protection Agency (RIO at SDU; #16.510).

Measures

Demographics and details on deployment

Data were collected on basic demographic information, such as age, gender, parental and marital status, number of deployments, most recent country of service and experiences of severe injury, life-threatening danger and dread during military service as well as received debriefing. Additionally, information on recent major life events (e.g. divorce, relocation, illness), use of psychotropic medication, and weekly alcohol intake was gathered.

PTSD-symptoms (ITQ)

To assess symptoms of PTSD and CPTSD in accordance with ICD-11, we used the International Trauma Questionnaire (ITQ).¹³ This brief diagnostic self-report measure captures a limited but core set of symptoms through simply worded items and thus maximizes the clinical utility and international applicability in accordance with the organizing principles of ICD-11. PTSD symptoms were assessed across three clusters: Re-experience, Avoidance, and Sense of Threat, each containing two items evaluated over the past month. Disturbance of self-organization (DSO) was also measured using three clusters with two items each: Affective Dysregulation, Negative Self-concept, and Disturbances in Relationships. The items were answered in accordance with how the participant typically felt, thought about themselves, and related to others. The three items for measuring the level of functional impairment for each domain (PTSD and DSO) were excluded from this study. All items were rated on a five-point Likert scale from (0) “Not at all” to (4) “Extremely”, resulting in a total of 12 items and a possible total ITQ score ranging from 0–60. Scores ≥2 (”Moderately”) indicated symptom presence, in accordance with standard practice in trauma research.^14,15 The ITQ have demonstrated strong psychometrical properties in multiple studies.^16-18 In this study the alpha values for the ITQ were α = 0.85, for the PTSD-subscale α = 0.80 and for the DSO subscale α = 0.81.

Negative affectivity, somatization and dissociation (TSC-26)

Three dimensions: negative affectivity, somatization and dissociation, were measured using the revised 26-item Danish version of the Trauma Symptom Checklist (TSC-26), a validated instrument for assessing psychological symptoms related to trauma.¹⁹ The TSC-26 rated the items on a four-point Likert scale ranging from (1) “no”, (2) “yes, sometimes”, (3) “yes, often”, to (4) “very often”. Possible score ranges are 10–40 for negative affectivity; 11–44 for somatization; 5–20 for dissociation, and 26–104 in total. In this study, the alpha values for the subscales were as follows: negative affectivity α = 0.84, somatization α = 0.77, dissociation α = 0.78, and total TCS α = 0.90.

Aggressive behavior (AS)

Previous research has shown elevated aggression levels among veterans with PTSD, and that both PTSD diagnosis and symptoms can predict higher aggression.²⁰ In this study, the levels of aggression were measured using a short, 6-item aggression scale, developed by Taft and colleagues at the National Center for PTSD, Boston. This scale assesses aggression across a spectrum from verbal violence and threats of violence (e.g., verbal abuse, threatened someone with a weapon) to actual perpetration of violence (e.g., physical fighting, using a weapon against someone). Participants reported aggressive incidents toward others over the four months prior to assessment, using a 7-point scale ranging from 'never' (=0), 'once' (=1), 'twice' (=2), '3 to 5 times' (=3), '6 to 10 times' (=4), '11 to 20 times' (=5), to 'more than 20 times' (=6). The total aggression frequency score was computed by summing these values. The alpha value for the total aggression scale in this study was α = 0.62.

Social support (CSS)

The importance of social support in post-trauma recovery has been reflected in numerous studies on trauma and is generally regarded as a potential mediator of distress and a predictor of psychological well-being among individuals who have experienced traumatic events. In this study, the experience of social support was measured using the Danish version of the Crisis Support Scale,²¹ a 7-item scale measuring social support after the occurrence of a crisis. The items assessed include: 1) perceived availability of someone to listen; 2) contact with people in a similar situation; 3) the ability to express oneself; 4) received sympathy and support; 5) practical support; 6) the experience of being let down; and 7) general satisfaction with social support. Responses were rated on a 7-point Likert scale, ranging from (1) “never” to (7) “always”, which gives the total score a possible range of 7–49. The CSS has been used in multiple trauma studies, and the scale demonstrates high robustness, with strong reliability, validity, internal consistency, and item discrimination. Elklit et al.²² have verified the psychometrical reliability and validity of the Danish version. In the present study, the alpha value for the total CSS score was α = 0.67.

Personality traits and psychopathology

The Millon Clinical Multiaxial Inventory-III (MCMI-III) is a self-report inventory that describes personality traits, severe personality pathology as well as clinical syndromes. It consists of 175 true-false items, organized into 24 scales, three modifying indices (Disclosure, Desirability, Debasement), and one validity index. These items correspond broadly to Axis I and Axis II disorders as outlined in the DSM-IV. The PTSD subscale was excluded for two reasons: 1) the items in the subscale do not include hypervigilance, and 2) the subscale might interfere with the analyses, since it seeks to assess the same symptoms as the ITQ-PTSD scale. Since raw scale scores are more appropriate for parametric tests of significance, these are used instead of corrected base rate scores.²³ The MCMI-III has been validated in Danish²⁴ and shows advantages in its relatively short administration time, its foundation in Theodore Millon’s theoretical-evolutionary model of personality, and the inclusion of modifying indices.²⁵

Data analysis

Descriptive analyses were conducted to calculate the frequency, mean, and standard deviation for all variables in the total sample. Reliability coefficients were calculated. Correlations between the total scales, subscales and three background factors (age, number of deployments and time since last deployment) were also calculated. Bivariate association between the MCMI-III scales for clinical and personality disorders were calculated. Due to the high number of independent variable, several hierarchical regression analyses were run preliminary to explore the contribution of each variable within the class of factors. Then, two hierarchical regression models were built, one for PTSD and one for CPTSD, both with four steps. The first included psychiatric medication and recent major life events. The second step consisted of clinical disorders that were strongly associated with the dependent outcome measures. The third step included personality disorders. The fourth and final step included the subscales for negative affectivity, somatization, and dissociation plus aggression and social support. The order of the steps rests on the assumption that clinical disorders are more likely to be changeable than personality disorders, and that the TSC subscales, the AS and the CSS might contribute with intraindividual and interpersonal variables that were not included in the clinical or personality disorders.

Results

Demographic statistics

The mean age of the sample was 42.4 years (SD = 8.7). The majority of participants were married or cohabiting with a partner, and most were parents with more than one child. Data on prior deployments indicated that over half of the participants had been deployed more than once, with an average of 14.2 years (SD = 7.99) since the most recent deployment. A substantial proportion reported experiencing life-threatening danger to themselves or others, personal injury, or witnessing injury to others. Approximately two-thirds reported reactions of intense fear, helplessness, or horror related to war experiences. More than half of the participants did not receive any debriefing after deployment. Just under one-third reported three or more recent major life events, almost half were using psychotropic medication, and a minority consumed more than 14 alcohol units per week (see Table 1).

Table 1.

Demographics and stressors.

	N	Valid %	Mean (SD)	Median
Age	138		42,67 (8.7)	44.5
22–37		32.6 %
31–40		32.6 %
41–50		34.8 %
Marital Status	139
Unmarried		25.9 %
Married/cohabiting		58.3 %
Widower/divorced		15.8 %
Children	136
Yes		76.5 %
No		23.5 %
Number of children	106		2 (1.2)	2
1		36.8 %
2		39.6 %
3–8		23.6 %
Number of deployments	138		2.5 (2.2)	2
1		45.6 %
2–3		35.5 %
3–15		18.9 %
Years since last deployment	133		14.2 (7.99)	13
0–9		36.8 %
10–19		29.4 %
20–29		33.8 %
Experience of life-threatening danger to self or others	138
Yes		87 %
No		13 %
Injured or witnessed others being injured	137
Yes		72.3 %
No		27.7 %
Reacted with intense fear, helplessness or horror	134
Yes		67.9 %
No		32.1 %
Debriefing	138
Yes		44.2 %
No		55.8 %
Life events	102		1.4 (0.491)	1
0–1		60,1 %
2–10		39.9 %
Psychiatric medication	137
Yes		44.5 %
No		55.5 %
Amount of alcohol consumed weekly, number of units	136
0–14		81.6 %
More than 14		18.4 %

Preliminary analyses

With the purpose of identifying covariates for subsequent analyses, preliminary analyses were conducted between demographic variables and the ITQ, TSC-26, AS, CSS and MCMI-III scales (Table 2). No significant correlations were found between ITQ-scores and the demographic factors age (r = 0.02, p = .819) and number of children (r=−0.994, p = .996). A one-way ANOVA analysis indicated no significant relationship between parental (F₍₂,₁₃₂₎= 0.028, p=.867) or marital status (F₍₃,₁₃₄₎=0.335, p = .716) and the ITQ-total scores. We found no significant relationship between characteristics related to prior deployments: number of deployments (r = 0.06, p = .49), years since last deployment (r = 0.07, p = .43), experience of danger to life (F₍₂,₁₃₄₎=0.029, p = .87), injury (F₍₂,₁₃₃₎=1.72, p = .19), reactions of intense fear (F₍₂,₁₃₀₎=1.224, p = .271) or debriefing (F₍₂,₁₃₄₎=1.997, p = .16), and the ITQ-total score.

Table 2.

Descriptive statistics and reliability of subscales.

	Number of items	Mean (SD)	Median	Alpha	Range	Min.	Max.
ITQ	12	31.47 (8.78)	32	.85	38	10	48
PTSD	6	15.33 (5.09)	15	.80	22	2	24
DSO	6	16.13 (5.1)	17	.81	20	4	24
TSC-26	26	66.89 (13.1)	66	.90	64	40	104
TSC-26 Negative affectivity	10	26.68 (6.2)	27	.84	28	12	40
TSC-26 Somatization	11	29.56 (5.59)	29	.77	29	15	44
TSC-26 Dissociation	5	10.64 (3.41)	10	.78	15	5	20
AS	6	4.53 (4.86)	4	.62	27	0	27
CSS	7	28.57 (7.03)	28	.65	36	10	46

However, a significant correlation was found between the ITQ-total score and amount of recent major life events (r = 0.2, p = .044) and similarly with use of psychotropic medication (F₍₂,₁₃₅₎=7026, p = .009). A relationship was found between age and TSC-26 Dissociation, with 48–66 year olds reporting significantly higher levels of dissociation than the two younger age groups 22–37 and 38–47 (F_(3,132)=4.521, p = .013).

Furthermore, we found positive significant relationships between the CSS scale and getting injured or experiencing others getting injured during deployment (F₍₁,₁₂₇₎=9.965, p = .002), as well as between the CSS subscale and receiving debriefing (F₍₁,₁₂₈₎=4.790, p = .030). A negative significant relationship was found between the CSS subscale and weekly alcohol consumption (F₍₄,₁₂₄₎=2.562, p = .042).

Correlations between subscales and the ITQ

All subscales correlated significantly with the ITQ-total score (see Table 3). There was a significant negative correlation between CSS and the ITQ DSO-score alone (r=−0.306, p < .001), with low levels of perceived social support associated with high levels of posttraumatic severity. However, the ITQ PTDS-score alone did not correlate significantly with the CSS. AS showed an overall positive, significant relationship with the other subscales, with CSS as exception (see Table 3).

Table 3.

Correlations between scales, subscales, age, and deployment.

	2	3	4	5	6	7	8	9	10	11	12
1 ITQ total	.861^⁎⁎	.862^⁎⁎	.761^⁎⁎	.682^⁎⁎	.620*	.645^⁎⁎	.346^⁎⁎	−0.233^⁎⁎	.020	−0.060	.070
2 PTSD		.485^⁎⁎	.606^⁎⁎	.437^⁎⁎	.542^⁎⁎	.620^⁎⁎	.266^⁎⁎	−0.097	.035	−0.043	.063
3 DSO			.701^⁎⁎	.731^⁎⁎	.529^⁎⁎	.492^⁎⁎	.331^⁎⁎	−0.306^⁎⁎	.001	−0.060	.059
4 TSC-26 total				.882^⁎⁎	.887^⁎⁎	.775^⁎⁎	.480^⁎⁎	−0.201*	.030	−0.047	.075
5 Negative Affectivity					.630^⁎⁎	.526^⁎⁎	.440^⁎⁎	−0.221*	−0.032	−0.060	−0.011
6 Somatization						.614^⁎⁎	.391^⁎⁎	−0.143	.023	−0.056	.059
7 Dissocitation							.404^⁎⁎	−0.149	.133	.006	.188
8 AS								−0.112	−0.189*	−0.085	−0.015
9 CSS									.046	.049	−0.142
10 Age										.307^⁎⁎	.514^⁎⁎
11 Number of deployments											−0.225^⁎⁎
12 Time since last deployment

Note:.

⁎

p<.05;.

⁎⁎

p<.01.

A bivariate correlation showed differences in the relationship between the MCMI-III subscales and ITQ PTSD- and DSO-scores respectively. The DSO-score showed more significant, positive correlations with the MCMI-III subscales than the PTSD-score, correlating significantly with all Axis I and Axis II subscales except the Narcissistic and Compulsive personality disorders. The PTSD-score correlated significantly with six of the Axis II subscales (Avoidant, Negativistic, Self-defeating, Schizotypal, Borderline and Paranoid), as well as five of the Axis I subscales (Anxiety, Somatoform, Dysthymia, Thought Disorder and Major Depression; see Table 4). It is remarkable that most of the correlations are much stronger for the DSO than for the PTSD, many of the twice as strong or more. This could be an indication of a chronification of disorders that aligns well with the CPTSD diagnosis.

Table 4.

Bivariate associations (Pearsons Correlations) MCMI-III subscales and ITQ-PTSD and -DSO scores.

	PTSD	DSO
1 - Schizoid	.053	.405^⁎⁎
2A - Avoidant	.217*	.539^⁎⁎
2B - Depressive (Melancholic)	.161	.512^⁎⁎
3 - Dependent	.179	.352^⁎⁎
4 - Histrionic	−0.137	−0.395^⁎⁎
5 - Narcissistic	.005	−0.149
6A - Antisocial	.035	.255^⁎⁎
6B - Sadistic	.145	.214*
7 - Compulsive	.095	−0.142
8A - Negativistic	.207*	.395^⁎⁎
8B - Self-defeating	.275^⁎⁎	.557^⁎⁎
S - Schizotypal	.267^⁎⁎	.516^⁎⁎
C - Borderline	.248^⁎⁎	.567^⁎⁎
P - Paranoid	.262^⁎⁎	.385^⁎⁎
A - Anxiety	.588^⁎⁎	.448^⁎⁎
H - Somatoform	.277^⁎⁎	.474^⁎⁎
N - Bipolar (Manic)	.063	.200*
D - Dysthymia	.247^⁎⁎	.637^⁎⁎
B - Alcohol Dependence	.070	.211*
T - Drug Dependence	.164	.292^⁎⁎
SS - Thought Disorder	.247^⁎⁎	.582^⁎⁎
CC - Major Depression	.339^⁎⁎	.629^⁎⁎
PP - Delusional Disorder	.174	.313^⁎⁎

Note:.

⁎

p<.05,.

⁎⁎

p<.01.

Regression analysis

Models of hierarchical multiple regression analyses were conducted to assess the ability of the independent measures (TSC-26, AS, CSS, and MCMI-III) to predict posttraumatic severity (ITQ PTSD- and DSO-scores respectively), when controlling for demographic factors that showed significant relationship with the ITQ-total score in preliminary analyses (recent major life event and the use of psychotropic medication). Only significant correlations with an r-value greater than 0.5 for the ITQ-DSO and 0.25 for the ITQ-PTSD from preliminary analyses have been included (see Tables 5 and 6).

Table 5.

Model I Hierarchical regression analysis, ITQ-PTSD scores.

Steps	Variable	β	Std. E	t	Sig.	F (df)	Sig.	R^2	Adj. R^2	ΔR^2
1.						7.012 (2,134)	.001	.095	.081	.095
	Psychiatric medication	.218	.839	2.648	.009
	Recent major life events	.206	.210	2.504	.013
2.						18.392 (4,112)	<0.001	.405	.383	.254
	Psychiatric medication	.191	.767	2.530	.013
	Recent major life events	.160	.188	2.095	.039
	A - Anxiety	.535	.095	6.168	<0.001
	H - Somatoform	−0.027	..137	−0.319	.750
3.						10.577 (7,105)	<0.001	.414	.374	.008
	Psychiatric medication	.168	.800	2.140	.035
	Recent major life events	.145	.196	1.820	.072
	A - Anxiety	.566	.112	5.535	<0.001
	H - Somatoform	−0.013	.145	−0.148	.883
	8B - Self-defeating	.022	.102	.192	.848
	S - Schizotypal	−0.147	.123	−1.075	.285
	P - Paranoid	.114	.103	1.036	.302
4.						10.363 (12,92)	<0.001	.575	.519	.166
	Psychiatric medication	.095	.746	1.290	.200
	Recent major life events	.017	.191	.219	.827
	A - Anxiety	.467	.109	4.573	<0.001
	H - Somatoform	−0.101	.145	−1.104	.273
	8B - Self-defeating	−0.048	.105	−0.408	.684
	S - Schizotypal	−0.362	.119	−2.723	.008
	P - Paranoid	.154	.100	1.432	.156
	TSC-26 Negative Affectivity	.164	.096	1.434	.155
	TSC-26 Somatization	.183	.103	1.584	.117
	TSC-26 Dissociation	.339	.148	3.409	<0.001
	CSS	.005	.053	.070	.945
	AS	−0.018	.088	−0.211	.834

Table 6.

Model II Hierarchical regression analysis, ITQ-DSO scores.

Steps	Variable	Std. B	Std. E	t	Sig.	F (df)	Sig.	R^2	Adj. R^2	ΔR^2
1.						3.650 (2,134)	.029	.052	.038	.052
	Psychiatric medication	.144	.861	1.713	.089
	Recent major life events	.168	.215	1.996	.048
2.						21.571 (5,107)	<0.001	.502	.479	.432
	Psychiatric medication	.137	.698	1.990	.049
	Recent major life events	.100	.169	1.451	.150
	D - Dysthymia	.300	.147	2.572	.011
	SS - Thought Disorder	.233	.109	2.426	.017
	CC - Major Depression	.206	.131	1.674	.097
3.						12.542 (9,102)	<0.001	.551	.508	.050
	Psychiatric medication	.140	.697	2.034	.045
	Recent major life events	.153	.172	2.173	.032
	D - Dysthymia	.179	.170	1.327	.187
	SS - Thought Disorder	.155	.150	1.170	.245
	CC - Major Depression	.153	.131	1.246	.216
	2A - Avoidant	.346	.114	2.730	.007
	8B - Self-defeating	−0.134	.128	−0.940	.349
	S - Schizotypal	−0.139	.131	−0.958	.340
	C - Borderline	.211	.111	1.769	.080
4.						11.913 (14,89)	<0.001	.668	.611	.114
	Psychiatric medication	.079	.672	1.193	.236
	Recent major life events	.114	.173	1.679	.097
	D - Dysthymia	.180	.163	1.371	.174
	SS - Thought Disorder	.122	.149	.933	.354
	CC - Major Depression	.078	.130	.635	.527
	2A - Avoidant	.278	.111	2.270	.026
	8B - Self-defeating	−0.203	.127	−1.418	.160
	S - Schizotypal	−0.198	.126	−1.416	.160
	C - Borderline	.181	.112	1.530	.130
	TSC-26 Negative Affectivity	.393	.087	3.791	<0.001
	TSC-26 Somatization	.014	.092	.141	.888
	TSC-26 Dissociation	.092	.141	.990	.325
	CSS	−0.037	.050	−0.525	.601
	AS	.023	.078	.304	.762

Model I (PTSD)

In the first step psychotropic medication usage and recent major life events were included, accounting for 8,1 % of the variance in perceived PTSD symptoms. In the second step, two Axis I subscales, Anxiety and Somatoform, was entered, explaining 38.3 % of the variance. In the third step, three Axis II subscales, Self-defeating, Schizotypal and Paranoid, were added, explaining 37.4 % of the variance. Finally, in the fourth step, the three TSC-26 subscales (Negative Affectivity, Dissociation and Somatization) were entered along with AS and CSS, explaining 51.9 % of the variance in perceived PTSD symptoms. All four steps significantly contributed to the model. Psychotropic medication usage had a significant individual effect in the first three steps but not in the fourth. Recent major life events had a significant individual effect in first two steps, but not in the subsequent steps. Anxiety demonstrated a significant individual effect in the second, third and fourth step, while the Schizotypal subscale showed a significant negative individual effect only in the fourth step. Additionally, the TSC-26 Dissociation subscale had a significant individual effect in the fourth step.

Model II (CPTSD)

In this model, the first step included use of psychotropic medication usage and recent major life event, explaining 3.8 % of the variance in perceived DSO-symptoms. In the second step, three Axis I subscales - Dysthymia, Thought Disorder, and Major Depression - were added, accounting for 47.9 % of the variance. In the third step, four Axis II subscales: Avoidant, Self-defeating, Schizotypal, and Borderline were included, explaining 50,8 % of the variance. In the fourth step, the three TSC-26 subscales (Negative Affectivity, Somatization and Dissociation) were entered along with AS and CSS, explaining now 61.1 % of the variance in perceived DSO-symptoms. All steps significantly contributed to the model. Recent major life events had significant individual effects in the first and third steps, while psychotropic medication usage had significant individual effects in the second and third steps. Dysthymia and Thought Disorder both demonstrated significant individual effects in the second step. The Avoidant subscale had a significant individual effect in the third and fourth steps, while the TSC-26 Negative Affectivity subscale showed a significant individual effect in the fourth step.

Discussion

The present study concerns a large group of Danish veterans in treatment. They are middle-aged, have several deployments and a mean of 14 years since the last one. The majority have been exposed to dangerous situations, had been injured or watched others being injured. Almost half take psychotropic medicine. This and recent life events were associated with trauma symptoms while many of the demographic and stressor variables surprisingly did not correlate with the ITQ. This might be due to the length of time and that many of the stressors have become personal history and transformed into present psychological traits, attitudes and behavior.

Social support was positively associated with injuries and debriefing, suggesting that these events had triggered valuable support from the army. Social support was negatively associated with alcohol consumption that can be seen as a way of self-medication and surprisingly, also with DSO symptoms. The latter may be an example of what Kaniasty & Norris³¹ call social deterioration: if survivors do not get better after a trauma, family and friends will tend to withdraw and so does the survivor eventually to avoid ‘social debts’ that he can never pay back. Aggression was strongly associated with all the trauma scales, indicating anger and emotional dysregulation as concurrent conditions.

Our second objective was to explore the relationship between PTSD and CPTSD, and clinical syndromes as well as personality pathology. Our findings suggest that CPTSD has a stronger relationship with both clinical syndromes and personality pathology than PTSD when assessed using the ITQ and MCMI-III. The fact that DSO scores correlated significantly with almost all MCMI-III clinical syndromes and pathological personality patterns suggest that there may be an underlying structure of psychopathology (the “p-factor”; Caspi & Moffitt³³) involved, which could make some people more vulnerable to develop chronic posttraumatic disorders.

The analyses show that perceived PTSD symptoms have a strong association with anxiety, as well as dissociative and schizotypal symptoms. This is expected, as both anxiety symptoms and dissociative states are common symptomatic presentations of PTSD, according to the ICD-11.⁴ Additionally, anxiety is established as a common comorbidity in relation to PTSD.²⁶ A review of the association between dissociation and PTSD²⁷ found a moderate relation between dissociative symptoms and trauma exposure and severity. Dissociation symptoms were found to rise sharply immediately after trauma exposure, then gradually decline for most, but stay high for some.²⁷ Furthermore, schizotypal symptoms appear to have a protective function in relation to perceived PTSD symptoms. This is surprising, since extant literature has linked experienced trauma to elevated levels of schizotypal symptoms.²⁸ A possible explanation for the negative association in this study, could be that cognitive and perceptual distortions, such as magical thinking and ideas of reference associated with schizotypal traits, might influence how PTSD symptoms are interpreted and reported, thereby affecting the accuracy of self-reported symptoms. In addition to this, schizotypal personality disorder is characterized by constricted or inappropriate affect, which could dull the subjective experience of emotional distress associated with PTSD.²⁹

The symptoms of Avoidant Personality Disorder had a significant positive association with perceived DSO symptoms. Avoidant personality disorder is characterized by the avoidance of situations and interactions, a lack of friendships, and avoidance of intimacy.³⁰ This fear of contact could be an expression of the relational disturbances that are associated with CPTSD. Also, people suffering from chronic post traumatic states might experience a decrease in social support.³¹ During the time following the impact of a traumatic experience, overt suffering can be legitimized, but particularly among people with PTSD-symptoms that persist over time, losses in social support might occur. People suffering from chronic posttraumatic states sometimes experience diminished interest in interpersonal activities, feelings of detachment or estrangement from others, angry outbursts, potential for aggression and social skills deficits, which might be a part of the explanation.³¹ Experiencing chronic states of PTSD may make it hard to appreciate the support efforts of others fully, and the complicated and conflicted interpersonal dynamics they face might create overburdening demands on their social networks, causing them to withdraw from reliable acquaintances.³¹ These perspectives are in accordance with our findings that CSS is significantly negatively correlated with DSO symptoms, but not significantly correlated with PTSD symptoms, suggesting that a lack of social network may be of particular interest when trying to understand why some veterans develop chronic posttraumatic suffering.

Negative Affectivity showed the strongest effect on perceived DSO symptoms, overruling the effect of the related variable Dysthymia. It has been suggested that a negative worldview might influence the severity of trauma-related symptoms, potentially hindering post-traumatic recovery,³² and that negative affectivity and social withdrawal are interconnected, affecting PTSD symptom development.¹⁹ Furthermore, negative emotionality has been suggested as a core component of the p-factor, representing a diffuse unpleasant affective state, underlying many different psychiatric disorders.³³ This supports the aforementioned idea that the p-factor could have explanatory value in relation to CPTSD.

The links between DSO-symptoms, personality pathology, and Negative Affectivity respectively might inspire future research and inform trauma-related treatment. By broadening our understanding of these dynamics, we can contribute to more effective interventions and support systems, ultimately improving long-term outcomes for individuals living with the enduring consequences of trauma.

Limitations

Our study does not establish causal relationships between CPTSD and personality disorders. Therefore, it is still up for discussion whether the veterans suffering from chronic psychopathological states have had hidden predispositions or whether the personality patterns can be seen as coping strategies caused by war experience. Research has shown that the majority of Danish soldiers receiving a psychiatric diagnose or psychotropic medication post-deployment has no experience of this kind from before deployment,³⁴ pointing to a “healthy warrior-effect”. Yet, the greatest risk factor for diagnoses and the use of psychotropic medication is having prior experiences in this area before deployment,³² suggesting that some soldiers may be predisposed. This underscores the importance of further research on the subject. Two of the applied scales, the CSS and the AS had reliability coefficients that were below 0.70. This could be one reason that the two scales do not emerge as predictors in the final step of the analysis of PTSD and CPTSD. The AS counts episodes of explicit violent behavior and transforms them into a 7-point scale where a simpler measure without this transformation may provide a more robust indicator of experienced violent behavior. In the correlation analyses, the two scales functioned as expected, with high aggression correlating positively and social support negatively with trauma measures. So, the interpretation of the results should take into account the low reliability of the two scales.

Conclusions and perspectives

Therapists and clinicians can use the present research to inform treatment and screening processes with awareness of the role of a possible p-factor, negative affectivity, social avoidance, and lack of social support in CPTSD. Screening for personality pathology at the start of treatment could be crucial for identifying veterans who will require a more complex and prolonged therapeutic approach.

Our results suggest that for veterans with a CPTSD profile, a therapy focused exclusively on trauma (e.g., prolonged exposure) may be insufficient if the disturbances in self-organization are not directly addressed. Treatment for these veterans should be sequential or integrated, combining trauma-focused interventions with specific modules for affect regulation, restructuring of negative self-concept, and interpersonal skills training—core areas in Borderline and Avoidant personality pathology.

Funding

None.

Ethical approval

RIO at SDU (#16.510).

Acknowledgements

We want to thank the Department for Trauma and Torture Survivors (ATT), Middelfart, Denmark for their contribution with collection of data. Also, thanks to the reviewers for their constructive criticism.