**1. Introduction**

Posttraumatic stress disorder (PTSD) diagnostic criteria have recently been reviewed in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [1]. PTSD diagnosis requires the experience of a traumatic event (TE) to bring on its development. The TE definition has also changed from DSM-IV to DSM-5 criteria, with the latter not requiring the person to feel intense fear, helplessness, or horror—as DSM-IV does. It has been estimated that 9.2% of those exposed to a TE will have PTSD [2]. The reason why only a minority of exposed people come to develop PTSD is the focus of much research. Vulnerability factors have been described and can be grouped into three clusters. Pre-traumatic vulnerability factors have been found, such as: arousal, negative affect, hostility, anger, lower cognitive abilities, psychopathology, prior trauma, poor family functioning, poverty, and

a family history of psychopathology. Among perceived peritraumatic vulnerability factors, life stress, emotional responses, and dissociation have all been found to predict PTSD development. Finally, it was found that lack of social support was the main posttrauma vulnerability factor for predicting PTSD development [3,4]. It has been found that peritraumatic factors are more predictive of posttraumatic growth, whilst pretrauma and personality-related variables are only predictive of PTSD [5]. Out of the personality variables, avoidant attachment significantly contributed to variance in PTSD risk.

Attachment is a construct which is related to the pattern of relationships established by a person with significant others. Attachment starts with the first relationships established between children and their caretakers, and is shaped by other relationships and events throughout the life cycle [6]. Attachment security influences the way a person copes with adversities and stress through positive mental representations of self and others [7]. On the other hand, insecure attachment orientations (anxiety and/or avoidance) predispose a person to mental disorders, due to the absence of a stable mental organization [7,8]. Attachment orientation predicts how adults react to stress and TEs [9,10]. Some studies sugges<sup>t</sup> that attachment patterns moderate the association between TEs and PTSD [11,12]. Several attachment-related TEs are characterized by high conditional risk for PTSD development, or cause high PTSD burden to society, such as the sudden unexpected death of a loved one, war, sexual violence, and witnessing atrocities [2,13–15]. Furthermore, trauma severity has also been found to moderate the association between attachment and PTSD [16]. O'Connor and Elklit [17] found a negative correlation between secure attachment and PTSD symptoms when studying a non-clinical population, which suggested that secure attachment protects against the development of PTSD. This was also found in the case of combat-related PTSD [18,19]. On the other hand, PTSD was shown to influence attachment insecurity [20,21]. Furthermore, attachment has been shown to be negatively influenced by maltreatment as a child [22]. Attachment orientations can be studied as pretraumatic vulnerability factors for PTSD development, but also as peritraumatic, and even as a consequence of the disorder itself.

Among the pretraumatic vulnerability factors, adverse childhood experiences (ACE) have been related to the development of PTSD, with an increase of risk after exposure to the index TE [23]. Childhood adversities can challenge secure attachment organizations and have enduring consequences on attachment orientations and psychopathology [24,25]. ACE have also been associated with combat-related PTSD, particularly in the case of physical neglect and multiple types of adversities [26].

Different TEs are related to different incidence rates of PTSD, the highest rates being related to interpersonal violence [2]. War-related TEs are well-known risk factors for the development of PTSD, as well as traumatic load [27,28]. Specific combat experiences are associated with different risks to develop PTSD [29,30]. Some of these war experiences, such as atrocities and killing, may constitute severe transgressions of combatants' deepest moral standards, and cause what Litz et al. [31] defined as 'moral injury'—the long-term negative consequences at psychological, behavioral, religious, emotional, biological, and social levels. The study of war-related PTSD demands that war experiences are well characterized. However, the results regarding the influence of attachment orientations in the relationship between TEs and PTSD development are conflicting [9,11].

Portugal was involved in a war conflict from 1961 to 1974 with Angola, Mozambique, and Portuguese Guinea (currently Guinea-Bissau), which were former colonies fighting for their independence. Most of the soldiers were deployed non-voluntarily for a period of 24 months of guerrilla war.

Portuguese veterans of the colonial wars have a high prevalence (39%) of probable PTSD [32]. In addition, the time lapse of 40 years after the end of these wars provides a long after-trauma period for PTSD to develop, including delayed onset—a subtype of PTSD which applies to those cases when symptoms only begin six months or more after exposure to the TE [33]. Nevertheless, only a few studies have researched this population. Furthermore, the number of older war veterans is rising and, as the majority retain their diagnosis of PTSD following evidence-based interventions, and one third drop out of treatment, it is therefore important to characterize this population further [34]. The aim of this study was to investigate some of the vulnerability factors for lifetime PTSD development, over a

period of 40 years after war-related TE, in a sample of Portuguese war veterans, especially focusing on ACE, attachment orientations, war experiences, and the experience of non-war-related TEs. In addition, we studied whether the association between war exposure and lifetime PTSD was confounded by or interacted with attachment orientations.
