Posttraumatic stress disorder (PTSD) has been associated with altered regulation of the hypothalamic pituitary adrenal (HPA) axis, namely increased glucocorticoid (GC) sensitivity [1]. This has been shown by the increased suppressive response of cortisol to the dexamethasone suppression test, and the increased glucocorticoid receptor (GR) number. Accordingly, lower cortisol levels have been found in specific subgroups and time of day of assessment, such as morning cortisol (e.g. [2],). Time point measurements of cortisol in blood, saliva or urine, are all affected by pulsatile secretion, diurnal rhythm, and day to day variation [3]. Levels of scalp hair cortisol constitute a longterm measurement which is not subject to such variations.4 Scalp hair cortisol has been shown to be a marker of longterm endogenous cortisol levels [4]. One cm of hair corresponds to one month of hair growth [5]. Therefore, hair cortisol concentrations (HCC) can be used to retrospectively study cortisol exposure from recent months, depending on the length of the hair segment. HCC have been shown to be a reliable method of measuring cortisol exposure and for reflecting the systemic effects of cortisol [4,6].

In conditions of chronic stress, HCC have been shown to be increased (e.g. [7],). Conflicting results have been shown for subjects with PTSD. For example, Steudte et al. [8] found that shortly after a traumatic event (TE), HCC were elevated in PTSD subjects when compared to traumatised controls, whereas Luo et al. [9] found that immediately after the TE, HCC were higher in the traumatised group, either with, or without PTSD, when compared to the non-traumatised control group. In that study, more than six months after the TE, the PTSD group showed significantly lower HCC when compared to the non-PTSD traumatized group, however there were no differences when compared to the non-traumatised control group. A recent study did not show any associations between HCC and PTSD symptom severity, although a negative association with avoidance symptoms was observed [10]. Steudte et al. [11] found no differences between PTSD subjects and trauma exposed controls, but significantly lower HCC in these TE exposed groups compared to a non-traumatised control group. Interestingly, these authors found negative correlations between HCC and severity of intrusion symptoms and also with time since traumatisation. Another study of the effects of traumatisation on HCC (PTSD subjects were excluded) found significantly higher levels in traumatised individuals when compared to controls, which declined 3.5 years after the trauma [12]. Collectively, the previous studies all seem to show that HCC could be associated with the traumatised state of subjects, with specific PTSD phenotypes, and that it depends on the time that had elapsed since exposure to a TE and also the type of trauma [13]. Some studies seem to support the idea that PTSD could be associated with lower HCC in female [9,14] but not in male [15] samples. Lower HCC have also been associated with childhood adversities [16].

PTSD studies suggest an intergenerational transmission of neurobiological correlates of PTSD to offspring [17]. Children of parents suffering from PTSD showed a significantly increased risk of developing PTSD and also other psychopathology [18-20]. The neurobiological substrate for intergenerational transmission of PTSD to offspring seems to be related to HPA axis dysregulation, which is also found in their parents [21]. The intergenerational transmission could be the result of inherited genetic factors [22,23], or as a consequence of veterans’ war-related traumatic experiences and PTSD for the family, which are associated with interpersonal violence, affecting parenting [24-26]. The highly adverse family emotional state prevalent when there is a veteran father with PTSD in the family dysregulates the HPA axis of the offspring [17,27]. Enhanced cortisol suppression after low dose dexamethasone was found in adult Holocaust survivors offspring (HSO) with PTSD, when compared with HSO without PTSD, or offspring of parents not exposed to the Holocaust [28]. Lower plasma cortisol levels and lower 24 hour urinary cortisol excretion were also found to be associated with the offspring of Holocaust survivors with PTSD in other studies [20,29]. These intergenerational effects have been found to be specifically associated with PTSD in mothers [18,29]. Up until now, no studies have evaluated an intergenerational influence of PTSD with respect to HCC.

PTSD is comorbid with major depressive disorder (MDD) in 52% of cases [30]. PTSD and MDD have distinctive pathophysiologic mechanisms [31], particularly of opposing directions regarding GR function. MDD is associated with increased point measurements of cortisol, as well as increased HCC [32,33]. It is also known that childhood trauma is a vulnerability factor for both PTSD and MDD [20,34] and that it is associated with adult HPA alterations, including lower systemic cortisol, as reflected by time point, short-term, and hair measurements (e.g. [35,36],). Studies of cortisol in patients with PTSD should take MDD comorbidity into account.

The main objectives of this study were to evaluate HCC in veterans with, and without lifetime PTSD which was assessed 40 years after exposure, and also to investigate how these veterans are influenced by childhood adversities, war exposure, and MDD. Furthermore, associations of PTSD symptom subclasses with HCC were investigated. We also aimed to compare HCC in the offspring of veterans with lifetime PTSD versus the offspring of veterans without lifetime PTSD. Additionally, we studied whether offspring’s HCC were influenced by childhood adversities and their fathers’ war exposure and PTSD symptoms. We also studied the relation between depression and HCC in these specific study groups. In addition, we investigated whether veterans’ HCC were associated with their offspring’s HCC.

Characteristics of the total veteran sample (n=59) included in the HCC analysis are described in Table 1. Groups were well-matched for socio-demographic and hair-related variables. No veterans had their hair treated, such as by dying, bleaching, permanent waving, or straightening. There were no differences between the offspring of veterans with and without PTSD concerning the use or not of hair treatments. Characteristics of the sample of veterans whose offspring was entered in the HCC analysis (n=43) were identical to those whose offspring did not participate, or were excluded (n=17). Table 2 shows the veterans’ offspring characteristics. Offspring’s groups according to lifetime PTSD diagnosis of their fathers were also well matched on socio-demographic and hair-related variables.

This study aimed to investigate the associations between war veterans’ lifetime PTSD and HCC and intergenerational effects on the respective offspring. To our knowledge, this is the first study to address the intergenerational effects of veterans’ PTSD on long-term cortisol levels, as measured by the hair analysis of the offspring.

In our sample, we did not find differences in HCC between male veterans with, and without lifetime PTSD. This result is in accordance with previous studies, which reported no association between HCC and PTSD [11], and also in studies specifically carried out with samples of mainly men [15,48]. Some studies have found HCC to be significantly lower in PTSD patients when compared with traumatised controls, but only in the case of female samples [9,14]. Furthermore, previous studies of war-related trauma showed no associations between PTSD and HCC either [48,49]. The fact that our study avoided inflating the discrimination between veterans with and without PTSD, by excluding those with intermediate CAPS scores, may have contributed to the absence of differences in HCC between veterans with, and without lifetime PTSD. Veterans with, and without lifetime PTSD did not show differences in HCC between those who have lifetime MDD, and those who do not. Grouping veterans according to having lifetime MDD, or not, also showed no difference in HCC between those with, or without lifetime PTSD. The grouping of veterans depending on lifetime MDD was based on the assumption that both combat PTSD and MDD each have distinctive pathophysiologic mechanisms [31], particularly of opposing directions regarding GR function. A previous meta-analytic study found no significant differences in daily cortisol output effect sizes for the TE, PTSD, and PTSD+MDD groups [2]. Pinna et al. [50] found increased salivary cortisol waking response in victims of intimate partner violence with MDD and also with both MDD and PTSD, but not with PTSD only, when compared to those with neither disorder. These authors conclude that comorbid depression could drive the relationship between PTSD and salivary cortisol response to waking, although the nature of the TE may also have had a strong influence. Since we measured HCC 40 years after war-related trauma, we may infer that the absence of differences in veterans’ HCC might be attributed to the effect of time since trauma, which has been largely reported to be associated with decreased HCC when compared to those who were not exposed [13]. In future studies, it would be important to include a non-traumatised group. There is also a possibility that the absence of differences in veterans’ HCC could have been influenced by treatments that veterans received over the years, either with psychotropic medications, or psychosocial interventions, although none had been in a structured psychotherapeutic intervention, such as the one described by Olff et al. [51] which was found to be associated with cortisol increase in PTSD patients responders when controlling for depressive symptoms.

We did not observe any associations between war exposure and HCC in the total sample of veterans. One previous metaanalytic study found no associations of cortisol levels with trauma exposure [52]. Recent studies have found lower plasma cortisol levels and specifically lower HCC to be associated with traumatic events [13,15]. We found a positive association between war exposure and HCC after excluding veterans with lifetime MDD, which remained after adjusting for psychotropic use. Different results might be accounted by different methodologies, assessment after a long time since traumatization, and to the fact that we have only studied HCC association with a specific type of trauma.

With regards to childhood maltreatment, veterans who reported more frequent physical abuse had lower HCC. This association remained after adjusting for psychotropic use. In previous studies, cortisol measurements in bodily fluids, which reflect time point, or short-term measurements, have also been found to be associated with childhood adversities, in particular, physical abuse (e.g. [20,35],). Only a few studies have investigated the association between childhood adversities and long-term systemic cortisol, as measured by hair analysis. Hinkelmann et al. [36] reported a negative association between childhood trauma and HCC, which was also independent of depression. A recent meta-analysis found a negative moderate effect size association between childhood maltreatment and HCC [16]. Similar to Steudte et al. [11], we also found no association between total childhood trauma scores and HCC, probably due to methodology limitations. However, several studies increasingly support the association between early life stress and altered HPA axis function mediated by gene x environment interactions through mechanisms of epigenetic alterations [53,54]. These epigenetic alterations have been associated with altered HCC and increased risk to develop PTSD [16,55]. The association of childhood adversities and HCC could constitute one mechanism that moderates the association between HCC and PTSD. Further studies, particularly longitudinal ones, are warranted to better understand these effects.

Offspring of veterans with lifetime PTSD did not show differences in HCC when compared to those of veterans without lifetime PTSD. Even though no other studies have addressed HCC in the offspring of parents with PTSD, lower plasma cortisol levels and lower 24 hour urinary cortisol excretion have been associated with the offspring of parents with PTSD (e.g. [29,56],). Accordingly, we did not find any association between veterans’ total lifetime PTSD symptoms, neither with total war exposure and respective offspring’s HCC. War exposure intensity has been associated with child behaviour problems [57,58], however no studies have investigated associated HPA axis alterations in the children of war veterans. Most research studying cortisol in the intergenerational transmission of neurobiological correlates of PTSD studied HSO, using plasma or urine. These studies found decreased cortisol levels in the HSO of parents with PTSD, and particularly of mothers, when compared to parents without PTSD [17,59]. Our study differs with respect to a different methodology, as we did not include mothers, studied a different population, and used HCC, which reflected long-term, rather than short-term cortisol levels [4]. In this study, offspring of veterans with lifetime PTSD that have lifetime MDD had significantly higher HCC when compared with those without lifetime MDD. This result is in accordance with the theory of depression relating to an altered HPA axis function in the opposite direction of PTSD [60]. In fact, increased HCC have been previously found in depressed patients [32]. Increased point measurements of cortisol, and enhanced cortisol stress reactivity have also been found in cases of depression [33,61].

Offspring without lifetime MDD showed a negative association between childhood physical abuse and HCC. As mentioned above, childhood adversities have previously been related with alterations of HPA axis functioning, particularly those with lower cortisol levels (e.g. [16,61,62],). Specifically, physical abuse has been found to be associated with a blunted salivary cortisol response to psychosocial stress in a non-clinical sample [62] and with lower cortisol plasma levels in personality disorder patients [35]. Hinkelmann et al. [36] and Suzuki et al. [61] found that the HPA axis alterations associated with childhood adversities did not depend on depression. In our study, we only found an association in the case of offspring without depression. This result is consistent with that of their fathers, who showed a negative association between childhood physical abuse and HCC, which did not depend on the presence of depression. The absence of an association between fathers’ and offspring’s HCC reflects the complexity of inheritance on the regulation of HPA axis functioning.

Some important aspects of this study need to be highlighted, such us: All the war veterans were evaluated with the gold standard instrument to diagnose PTSD, the CAPS; both veterans and their offspring were objectively and directly evaluated. Moreover, this study did not inflate the discrimination between veterans with, and without PTSD, as we did not exclude veterans with intermediate CAPS scores.

One limitation of this study is the cross-sectional nature of the design, which does not allow for drawing causal inferences from our findings. The small sample size warrants future studies to confirm our results, which should include larger samples. The fact that only men in the group of parents were studied does not allow for generalising the results for females. The same applies to the generalisation of TE, as the index trauma was only combat-related. It would be important to have a nottrauma-exposed group to observe the differences that could be attributed to exposure. Since PTSD was studied as a categorical variable, when we performed a sensitivity analysis no associations were found between HCC and CAPS total and symptom subscales scores, considering the DSM-IV threefactor model and the five-factor model, even after adjusting for lifetime MDD in the total group of veterans. This could also be attributed to a confounding effect. In the statistical analysis, we did not consider dependency by family. However, we estimated the dependency of HCC by family using interclass correlation and the value was zero, meaning that we can ignore the dependency by family for this analysis. We did not study the modifying role of the offspring’s mother, which is another limitation to the understanding of what conditions were really inherited from veterans. We cannot rule out that our results may be explained by a type I error, as we did not correct for multiple-testing, considering the resulting risk of increase of type II error in this study.

In summary, no differences in HCC were found between veterans with and without lifetime PTSD, neither between their respective offspring. This could be attributed to the sample of veterans being exclusively male. Veterans-with-PTSD’s offspring with lifetime MDD showed increased HCC, when compared to veterans-with-PTSD’s offspring without lifetime MDD. Veterans without lifetime MDD showed a positive association between total war exposure and witnessing casualties amongst the enemy, and HCC. Veterans reporting more childhood physical abuse had lower HCC, which was an effect that did not depend on the presence of lifetime MDD. No associations were found between offspring’s HCC and their fathers’ HCC, neither war exposure, nor total CAPS score. Offspring’s childhood physical abuse was negatively associated with HCC in those without lifetime MDD. Accordingly, we can conclude from this study that HCC do not seem to constitute a marker of intergenerational heritage of war-related PTSD, except in the case of offspring with lifetime MDD of veterans with PTSD. However, our data suggest that HCC is a marker of adult reported childhood physical abuse.