**6. Sex as a moderator**

#### **6.1 Social support**

Empirical evidence suggests that high levels of oxytocin may enhance the effect of social support on stress (Taylor et al., 2006), suggesting that the tend-and-befriend response may not only affect how females respond to stress, but may also affect how they benefit from social support in the aftermath of trauma. Not only do females seek, receive, and provide more support in stressful situations than males (Andrews et al., 2003; Tamres et al., 2002), there are also some indications that the positive effects of social support as well as the negative effects of negative social attention on PTSD may be more pronounced in females than in males (Ahern et al., 2004; Andrews et al., 2003; Gavrilovic et al., 2003). However, other studies have not found this (Farhood et al., 1993), and the relationship between sex, social support, and PTSD appears to be complex (Christiansen & Elklit, 2008). In order to test the hypothesis that sex moderates the impact of social support on PTSD, we conducted

activation of the HPA axis and the SNS corresponds to the male stress response proposed above. Similarly, the idea that the dissociation pathway should be linked to the PNS is in accordance with the female stress response. There thus appears to be some overlap between the findings concerning different pathways to PTSD in children and sex differences in the initial stress response in adults. This suggests, that even though the two studies on pathways in children did not report any sex differences in the anxiety/arousal and dissociation pathways, males may be more likely to follow an anxiety/arousal pathway and females a dissociation pathway to PTSD. This is in accordance with the ideas proposed by Perry et al. (1995), who argued that activation and sensitisation of the dissociative response pattern was more common in females than in males, whereas the opposite appeared to be the case for the hyperarousal pattern. Furthermore, although the avoidance pathway identified in the study by Kaplow et al. (2005) was not associated with a specific physiological system, it could be argued that avoidance is a form of fleeing behaviour associated with the fight-or-flight system (Taylor, 2006), and as this pathway was more pronounced in boys than in girls, males may also be more likely to follow an avoidance

If males and females follow different pathways to PTSD, we would expect to find sex differences in the relationship between certain risk factors and PTSD. Whereas research has mainly focused on mediation effects between sex and PTSD, the moderation hypothesis states that there may be sex differences in the relationship between different risk factors and symptomatology (Stone et al., 2010), suggesting that some risk factors may predict PTSD in females but not in males, whereas others may better predict PTSD in males. Furthermore, if such sex-specific pathways are related to the fight-or-flight versus tend-and-befriend response to stress, we would hypothesise that particularly behaviours related to either the tend-and-befriend or the fight-or-flight response to stress would show sex differences in their ability to predict PTSD. For example, physiological processes involved in the tend-andbefriend response, such as activation of the PNS and the role of oxytocin, could make emotion-focused coping, particularly support seeking, social support, and possibly dissociation more closely related to PTSD in females than in males. In contrast, arousal, anxiety, and possibly problem-focused and avoidant coping strategies could be hypothesised to be more important in the development or prevention of PTSD in males

Empirical evidence suggests that high levels of oxytocin may enhance the effect of social support on stress (Taylor et al., 2006), suggesting that the tend-and-befriend response may not only affect how females respond to stress, but may also affect how they benefit from social support in the aftermath of trauma. Not only do females seek, receive, and provide more support in stressful situations than males (Andrews et al., 2003; Tamres et al., 2002), there are also some indications that the positive effects of social support as well as the negative effects of negative social attention on PTSD may be more pronounced in females than in males (Ahern et al., 2004; Andrews et al., 2003; Gavrilovic et al., 2003). However, other studies have not found this (Farhood et al., 1993), and the relationship between sex, social support, and PTSD appears to be complex (Christiansen & Elklit, 2008). In order to test the hypothesis that sex moderates the impact of social support on PTSD, we conducted

pathway than females.

compared to females.

**6.1 Social support** 

**6. Sex as a moderator** 

a number of new analyses on previously published data based on two separate samples (see Christiansen & Elklit, 2008). One sample consisted of residents in an area, which had been almost completely destroyed by an industrial disaster (the explosion sample), the other consisted of high school students who had witnessed a stabbing incident, were a young girl was killed (the high school sample). In contrast to the original analysis, this time we used only the symptoms included in the DSM-IV to measure PTSD severity (i.e. the first 17 items of the Harvard Trauma Questionnaire; Mollica et al., 1992). As can be seen in Table 1, we found no significant sex differences in correlations between feeling let down by others and PTSD severity, and positive social support was unrelated to PTSD severity in both males and females in the explosion sample. However, we did find that positive social support was significantly more strongly correlated with PTSD severity in females compared to males in the high school sample. There thus appears to be some support for the hypothesis that sex moderates the relationship between social support and PTSD.


Pearson coefficients between risk factors and PTSD severity (combined score on the 17 first HTQ items) are shown separately for males and females from the explosion study and the high school stabbing incident.

a r values from the explosion study / r values from the stabbing incident study \* *<sup>p</sup>* < .05; \*\* p < .01; b *<sup>p</sup>* = .08; n.s.: not significant

Em. cop.: emotion-focused coping; rat. cop.: rational coping; Avoi. cop.: avoidance coping.

Table 1. Correlations between risk factors and PTSD severity for males and females

The impact of sex on the relationship between social support and PTSD is further complicated by the possibility that different types of support may have differential impact on PTSD symptomatology in males and females. For example, it is possible that talking about the traumatic event may be preferable for females, whereas males may appreciate practical assistance more, because such types of support are most compatible with the preferred coping strategies used by each sex. Furthermore, males as well as females depend mostly on female providers of support, and whereas females often have multiple support sources, adult males often cite their spouses as their only confidantes (Shumaker & Hill, 1991). This may be the reason why being unmarried serves as a risk factor for mortality in males to a greater extent than is the case for females (Shumaker & Hill, 1991). Thus, whereas females may be more vulnerable to lack of support and negative reactions from others more generally, males may be more dependent on the support of their spouse and thus be relatively more vulnerable to PTSD in case of separation or spousal bereavement. In accordance with this idea, a study on elderly bereaved has reported a lack of sex differences in PTSD severity (Elklit & O'Connor, 2005), which is in contrast to the otherwise wellestablished increased prevalence of PTSD among females.

#### **6.2 Anxiety/arousal**

The apparent role of AVP in the acute male stress response points to the possibility that arousal may predict PTSD in males more strongly than in females, and that males may be

Such results suggest that dissociative reactions in the female stress response may coexist with affiliative behaviour, in which the role of oxytocin has repeatedly been demonstrated

Not only do females report higher levels of traumatic dissociation than do males, peritraumatic dissociation also appears to be a stronger risk factor for PTSD in females. An initial diagnosis of ASD, which is very much based on the presence of dissociation, has been reported to be a more accurate predictor of PTSD in females than in males, and it has been suggested that this can be explained by an increased prevalence and predictive power of persistent dissociation in females compared to males (Bryant & Harvey, 2003). Similarly, peritraumatic dissociation has been found to predict higher levels of PTSD in female but not male victims of both motor vehicle accidents (Fullerton et al., 2001) and an industrial disaster (the explosion sample; Christiansen & Elklit, 2008). However, in the latter study, new correlation analyses revealed that although dissociation appeared to be more strongly related to PTSD severity in females than in males (see Table 1), this difference was not significant. Furthermore, dissociation was reported to significantly predict PTSD severity in both males and females in a study of Palestinians (Punamäki et al., 2005) and in a study of adolescents who witnessed a stabbing incident (the high school sample; Christiansen & Elklit, 2008). However, in the latter study, further analyses shown in Table 1 revealed that although there were no significant sex differences after other variables had been controlled for, dissociation and PTSD severity were more strongly correlated in females than in males in a simple correlation analysis (r = .68 vs. r = .44; z = -2.96; p = .003; unpublished analyses). This suggests that there may be at least a partial moderation effect of gender on the

Banyard and Graham-Bermann (1993) have criticised that female coping, and particularly the efficacy of coping in females, have only been studied based on theories of what constitutes adaptive coping in males, and that this has often led to the conclusion that females are inferior copers compared to males. They further criticise coping studies of being almost exclusively conducted on male samples. Although the field of coping and stress research is no longer as deserving of this critique as it used to be, the possibility of sex differences in coping efficacy is still largely ignored in the coping literature and research. However, it is possible that different coping styles are effective for males and females in different situations. Both oxytocin and AVP have been linked to the long-term effects of stress on coping (Neumann et al., 2000), and these two hormones may thus be involved in

The tend-and-befriend response is in its nature related to coping through support seeking, which tends to be categorised as an emotion-focused coping strategy, although seeking practical support is often categorised as a problem-focused strategy. In contrast, Taylor (2006) has linked the use of avoidance coping strategies to the fight-or-flight response. Furthermore, whereas avoidance coping may be viewed as a form of fleeing behaviour, it could be argued that problem-focused or rational coping is a form of fighting behaviour. As a result, we hypothesise that emotion-focused coping, particularly support seeking may be more adaptive in females, whereas problem-focused and avoidant coping may be more strongly associated with PTSD in males. In accordance with this hypothesis, studies, which have evaluated the relationship between coping and psychopathology separately for males

(Taylor, 2006; Taylor et al., 2002).

relationship between dissociation and PTSD.

sex differences in the use of different coping strategies.

**6.4 Coping** 

more prone to follow the anxiety/arousal pathway to PTSD. As a result, we would expect that anxiety as well as arousal during or immediately following a traumatic event would be stronger predictors of PTSD in males than in females. One study has found that a preexisting anxiety disorder predicted PTSD in males, but not in females (Bromet et al., 1998), and a study on survivors of an industrial accident found that anxiety levels measured three months after the incident were significantly related to PTSD one year later in males, but not in females (the explosion study; Christiansen & Elklit, 2008). Furthermore, as can be seen in Table 1, new analyses revealed that there was a non-significant trend towards anxiety being more highly correlated with PTSD severity in males compared to females (r = .73 vs. r = .57; z = .73; p = .08; unpublished analyses). Finally, a study of PTSD in MVA victims found that the interaction between sex and prior anxiety disorder was not significant (Fullerton et al., 2001).

Arousal in the acute aftermath of trauma has primarily been studied in MVA victims. Two studies (Bryant et al., 2000; Shalev et al., 1998) have reported that increased heart rate, but not blood pressure, measured in the emergency department following an MVA was associated with later PTSD. In contrast, Blanchard et al. (2002) found that increased heart rate was associated with lower levels of PTSD. Interestingly, both studies, which found heart rate to be positively associated with later PTSD, were based on predominantly male samples, whereas Blanchard and colleagues based their results on a predominantly female sample. Furthermore, the relationship between lower heart rate and later PTSD was significant only in females (Blanchard et al. 2002). It is possible, that the lack of significance in males was due to the low number of male participants, but another possibility is that decreased heart rate measured in the emergency room was associated with later PTSD in females but not males because of a suppressed HPA and SNS response and increased rates of dissociation in females. It is thus possible that the results reflect the existence of different pathways leading to PTSD in males and females. Finally, one study examining the influence of HPA hormones on PTSD separately in males and females found that increased levels of epinephrine, norepinephrine, and cortisol were related to PTSD status in males, but not females, one month after an MVA (Hawk et al., 2000). Thus, there are some indications that anxiety and particularly physiological arousal may be more strongly related to PTSD in males than in females.

#### **6.3 Dissociation**

To the best of our knowledge, no study to date has examined a possible role of oxytocin in dissociative responses. However, there are some arguments for a possible role of oxytocin in dissociation. First, dissociation appears to be associated with a down-regulation of the HPAaxis (Olff et al., 2007) as well as a reduced heart rate, consistent with a down-regulation of autonomic reactivity (Griffin et al., 1997). This is consistent with the effects of oxytocin on the SNS and HPA-axis. Secondly, in the above mentioned study by Bryant and colleagues (2000) on MVA victims, acute stress disorder (ASD) was associated with a lower heart rate compared to subclinical ASD without dissociation, indicating that dissociative reactions may suppress SNS arousal, just as oxytocin appears to do. Finally, David and Lyons-Ruth (2005) reported that when faced with frightening maternal behaviour, female toddlers reacted with affiliative behaviour in combination with inhibition of all other behaviour, consistent with the freeze response. These findings are in line with the ideas of Perry and colleagues (1995), that dissociation in children is associated with help-seeking behaviour.

more prone to follow the anxiety/arousal pathway to PTSD. As a result, we would expect that anxiety as well as arousal during or immediately following a traumatic event would be stronger predictors of PTSD in males than in females. One study has found that a preexisting anxiety disorder predicted PTSD in males, but not in females (Bromet et al., 1998), and a study on survivors of an industrial accident found that anxiety levels measured three months after the incident were significantly related to PTSD one year later in males, but not in females (the explosion study; Christiansen & Elklit, 2008). Furthermore, as can be seen in Table 1, new analyses revealed that there was a non-significant trend towards anxiety being more highly correlated with PTSD severity in males compared to females (r = .73 vs. r = .57; z = .73; p = .08; unpublished analyses). Finally, a study of PTSD in MVA victims found that the interaction between sex and prior anxiety disorder was not significant (Fullerton et al.,

Arousal in the acute aftermath of trauma has primarily been studied in MVA victims. Two studies (Bryant et al., 2000; Shalev et al., 1998) have reported that increased heart rate, but not blood pressure, measured in the emergency department following an MVA was associated with later PTSD. In contrast, Blanchard et al. (2002) found that increased heart rate was associated with lower levels of PTSD. Interestingly, both studies, which found heart rate to be positively associated with later PTSD, were based on predominantly male samples, whereas Blanchard and colleagues based their results on a predominantly female sample. Furthermore, the relationship between lower heart rate and later PTSD was significant only in females (Blanchard et al. 2002). It is possible, that the lack of significance in males was due to the low number of male participants, but another possibility is that decreased heart rate measured in the emergency room was associated with later PTSD in females but not males because of a suppressed HPA and SNS response and increased rates of dissociation in females. It is thus possible that the results reflect the existence of different pathways leading to PTSD in males and females. Finally, one study examining the influence of HPA hormones on PTSD separately in males and females found that increased levels of epinephrine, norepinephrine, and cortisol were related to PTSD status in males, but not females, one month after an MVA (Hawk et al., 2000). Thus, there are some indications that anxiety and particularly physiological arousal may be more strongly related to PTSD in

To the best of our knowledge, no study to date has examined a possible role of oxytocin in dissociative responses. However, there are some arguments for a possible role of oxytocin in dissociation. First, dissociation appears to be associated with a down-regulation of the HPAaxis (Olff et al., 2007) as well as a reduced heart rate, consistent with a down-regulation of autonomic reactivity (Griffin et al., 1997). This is consistent with the effects of oxytocin on the SNS and HPA-axis. Secondly, in the above mentioned study by Bryant and colleagues (2000) on MVA victims, acute stress disorder (ASD) was associated with a lower heart rate compared to subclinical ASD without dissociation, indicating that dissociative reactions may suppress SNS arousal, just as oxytocin appears to do. Finally, David and Lyons-Ruth (2005) reported that when faced with frightening maternal behaviour, female toddlers reacted with affiliative behaviour in combination with inhibition of all other behaviour, consistent with the freeze response. These findings are in line with the ideas of Perry and colleagues (1995), that dissociation in children is associated with help-seeking behaviour.

2001).

males than in females.

**6.3 Dissociation** 

Such results suggest that dissociative reactions in the female stress response may coexist with affiliative behaviour, in which the role of oxytocin has repeatedly been demonstrated (Taylor, 2006; Taylor et al., 2002).

Not only do females report higher levels of traumatic dissociation than do males, peritraumatic dissociation also appears to be a stronger risk factor for PTSD in females. An initial diagnosis of ASD, which is very much based on the presence of dissociation, has been reported to be a more accurate predictor of PTSD in females than in males, and it has been suggested that this can be explained by an increased prevalence and predictive power of persistent dissociation in females compared to males (Bryant & Harvey, 2003). Similarly, peritraumatic dissociation has been found to predict higher levels of PTSD in female but not male victims of both motor vehicle accidents (Fullerton et al., 2001) and an industrial disaster (the explosion sample; Christiansen & Elklit, 2008). However, in the latter study, new correlation analyses revealed that although dissociation appeared to be more strongly related to PTSD severity in females than in males (see Table 1), this difference was not significant. Furthermore, dissociation was reported to significantly predict PTSD severity in both males and females in a study of Palestinians (Punamäki et al., 2005) and in a study of adolescents who witnessed a stabbing incident (the high school sample; Christiansen & Elklit, 2008). However, in the latter study, further analyses shown in Table 1 revealed that although there were no significant sex differences after other variables had been controlled for, dissociation and PTSD severity were more strongly correlated in females than in males in a simple correlation analysis (r = .68 vs. r = .44; z = -2.96; p = .003; unpublished analyses). This suggests that there may be at least a partial moderation effect of gender on the relationship between dissociation and PTSD.

#### **6.4 Coping**

Banyard and Graham-Bermann (1993) have criticised that female coping, and particularly the efficacy of coping in females, have only been studied based on theories of what constitutes adaptive coping in males, and that this has often led to the conclusion that females are inferior copers compared to males. They further criticise coping studies of being almost exclusively conducted on male samples. Although the field of coping and stress research is no longer as deserving of this critique as it used to be, the possibility of sex differences in coping efficacy is still largely ignored in the coping literature and research. However, it is possible that different coping styles are effective for males and females in different situations. Both oxytocin and AVP have been linked to the long-term effects of stress on coping (Neumann et al., 2000), and these two hormones may thus be involved in sex differences in the use of different coping strategies.

The tend-and-befriend response is in its nature related to coping through support seeking, which tends to be categorised as an emotion-focused coping strategy, although seeking practical support is often categorised as a problem-focused strategy. In contrast, Taylor (2006) has linked the use of avoidance coping strategies to the fight-or-flight response. Furthermore, whereas avoidance coping may be viewed as a form of fleeing behaviour, it could be argued that problem-focused or rational coping is a form of fighting behaviour. As a result, we hypothesise that emotion-focused coping, particularly support seeking may be more adaptive in females, whereas problem-focused and avoidant coping may be more strongly associated with PTSD in males. In accordance with this hypothesis, studies, which have evaluated the relationship between coping and psychopathology separately for males

trauma types (Elklit & Christiansen, in press), it is possible that this relationship holds true not only for number of traumatic events but also for severity of traumatic events. Many studies on the relationship between coping and PTSD examine coping across situations in general, instead of specifically in relation to the traumatic event. As the preference, and especially the efficacy, of different coping strategies are likely to differ according to the characteristics of the situation at hand, studies which assess coping in relation to the specific traumatic event may be best suited to illuminate the complex relationship between coping

To sum up, there is some support for the moderation hypothesis, although results are inconsistent. Studies indicate that peritraumatic physiological arousal may be predictive of later PTSD in males but not in females. There is also some support that peritraumatic and persistent dissociation are more predictive of later PTSD in females. Bryant et al. (2000) found that later PTSD status was best identified through either increased heart rate or a diagnosis of ASD, indicating that two distinct pathways may lead to PTSD. There is thus some support for the hypothesis that males may primarily follow an arousal pathway and females a dissociation pathway to PTSD, although these findings are only preliminary and more research is needed. Some support was also found for a moderation effect on social support and PTSD. Although the association between social support and PTSD was stronger in females than in males, consistent with the tend-and-befriend theory and the greater role of oxytocin in females, it is unknown how such findings relate to the proposed sex-specific pathways to PTSD. One possibility is that the positive effects of social support and the negative effects of dissociation are somehow combined in the dissociation pathway to PTSD, possibly because dissociation occurs when tend-and-befriend behaviour proves ineffective, as suggested by Perry et al. (1995). However, another possibility is that lack of social support may represent an independent pathway to PTSD. Furthermore, support for the moderation hypothesis did not appear to be as strong for anxiety as was the case for arousal. It has been suggested that physiological arousal associated with the fight-or-flight system is related to fear but not anxiety (Grillon, 2008). It is therefore possible that anxiety and physiological arousal may not be combined in a single pathway to PTSD, but may instead represent similar but independent pathways. Finally, there was not much support for the moderation hypothesis in relation to coping. It is thus possible, that if the avoidance pathway can be

Figure 1 shows a preliminary model of how sex differences in the acute response to trauma may affect the development of PTSD. It is hypothesised that exposure to a traumatic stressor results in activation of the SNS in both sexes. However, once the SNS has activated the HPA axis, sex differences in the stress response begin to show. In males, the SNS response is amplified by AVP, which is secreted as part of the HPA response and fuelled by testosterone. This sustained activation of the SNS and increased activity in the HPA axis prepares the body for fighting or fleeing. If the SNS and the HPA axis are not downregulated and brought back to homeostasis, this fight-or-flight response may persist and make up one or more pathways to PTSD. Such pathways are hypothesised to be mediated by physiological arousal and anxiety and possibly by avoidance coping, and problem-focused coping. In contrast, HPA activation in females is associated with the secretion of oxytocin, the effects of which are amplified by oestrogen. Oxytocin is assumed

and PTSD in males and females.

**6.5 Evaluation of the moderation hypothesis** 

found in adults, it may not be moderated by sex.

and females, have generally found important sex differences (Hovanitz & Kozora, 1989). Although this has not been examined specifically for PTSD, there is some support that problem-focused coping may be more adaptive for males, whereas emotion-focused coping may be less maladaptive for females than has been found for males (Hovanitz & Kozora, 1989). Furthermore, emotion approach coping, which is a certain kind of emotion-focused coping involving active processing and expression of emotions, has been shown to decrease depression and increase life satisfaction in females, but to have the opposite effect on males (Folkman & Moskowitz, 2004). In contrast, problem-focused coping has been reported to be generally more adaptive for males (Tamres et al., 2002). Although coping was not included in the original analyses by Christiansen & Elklit (2008), new analyses based on data from the explosion study showed that emotion-focused and avoidant coping were associated with higher PTSD levels in both males and females, whereas rational coping was not associated with PTSD severity for either sex (see Table 1). *Z*-tests revealed no significant sex differences.

Thus, although there is some support for the hypothesis that different coping strategies may be adaptive for males and females in relation to symptomatology in general, this has not been documented in relation to PTSD. Furthermore, the studies mentioned above, which have found sex differences in the effectiveness of coping have generally focused on more specific coping strategies rather than broader coping styles, such as emotion- or problemfocused coping. This may suggest, that such coping styles include strategies, which are too heterogeneous for the detection of moderation effects. In particular, we find it problematic that emotion-focused coping is often operationalised to include both behaviours which would normally be considered positive, such as seeking social support, and behaviours which are more likely to represent symptomatology, such as rumination (Stanton, Danoff-Burg, Cameron, & Ellis, 1994). In fact, in the new analyses introduced above, in which we found emotion-focused coping to be significantly and positively related to PTSD severity in both males and females, the questionnaire used to assess coping (a revised version of the Coping Style Questionnaire; Elklit, 1996) did not include any items on support seeking behaviour. In order to shed more light on sex differences in the relationship between emotion-focused coping and PTSD, future studies should look more closely at which specific emotion-focused behaviours may be more adaptive for females than for males.

The question of whether sex moderates the effect of coping on symptomatology becomes even more complex when one considers that the efficacy of different coping styles may depend upon the characteristics of the situation in question. Araya et al. (2007) examined mental distress and quality of life in a sample of internally displaced Ethiopians and found that the relationship between coping and mental distress varied not only by sex but also by degree of trauma exposure. For males, emotion-focused coping was associated with higher levels of mental distress only when trauma exposure was high, whereas for females, emotion-focused coping was only significantly associated with higher levels of symptomatology when exposure was low (J. Chotai, personal communication; see also Araya et al., 2007). This suggests, that for women who have been repeatedly traumatised, emotion-focused coping may not be maladaptive, whereas that appears to be the case for men who have been exposed to multiple trauma. Although the study did not specifically examine PTSD symptoms, it is possible that a similar relationship between sex, degree of exposure, and emotion-focused coping exists for PTSD. Furthermore, because the doseresponse model of trauma may refer to both repeated exposure and exposure to more toxic

and females, have generally found important sex differences (Hovanitz & Kozora, 1989). Although this has not been examined specifically for PTSD, there is some support that problem-focused coping may be more adaptive for males, whereas emotion-focused coping may be less maladaptive for females than has been found for males (Hovanitz & Kozora, 1989). Furthermore, emotion approach coping, which is a certain kind of emotion-focused coping involving active processing and expression of emotions, has been shown to decrease depression and increase life satisfaction in females, but to have the opposite effect on males (Folkman & Moskowitz, 2004). In contrast, problem-focused coping has been reported to be generally more adaptive for males (Tamres et al., 2002). Although coping was not included in the original analyses by Christiansen & Elklit (2008), new analyses based on data from the explosion study showed that emotion-focused and avoidant coping were associated with higher PTSD levels in both males and females, whereas rational coping was not associated with PTSD severity for either sex (see Table 1). *Z*-tests revealed no significant sex

Thus, although there is some support for the hypothesis that different coping strategies may be adaptive for males and females in relation to symptomatology in general, this has not been documented in relation to PTSD. Furthermore, the studies mentioned above, which have found sex differences in the effectiveness of coping have generally focused on more specific coping strategies rather than broader coping styles, such as emotion- or problemfocused coping. This may suggest, that such coping styles include strategies, which are too heterogeneous for the detection of moderation effects. In particular, we find it problematic that emotion-focused coping is often operationalised to include both behaviours which would normally be considered positive, such as seeking social support, and behaviours which are more likely to represent symptomatology, such as rumination (Stanton, Danoff-Burg, Cameron, & Ellis, 1994). In fact, in the new analyses introduced above, in which we found emotion-focused coping to be significantly and positively related to PTSD severity in both males and females, the questionnaire used to assess coping (a revised version of the Coping Style Questionnaire; Elklit, 1996) did not include any items on support seeking behaviour. In order to shed more light on sex differences in the relationship between emotion-focused coping and PTSD, future studies should look more closely at which specific emotion-focused behaviours may be more adaptive for females than for males. The question of whether sex moderates the effect of coping on symptomatology becomes even more complex when one considers that the efficacy of different coping styles may depend upon the characteristics of the situation in question. Araya et al. (2007) examined mental distress and quality of life in a sample of internally displaced Ethiopians and found that the relationship between coping and mental distress varied not only by sex but also by degree of trauma exposure. For males, emotion-focused coping was associated with higher levels of mental distress only when trauma exposure was high, whereas for females, emotion-focused coping was only significantly associated with higher levels of symptomatology when exposure was low (J. Chotai, personal communication; see also Araya et al., 2007). This suggests, that for women who have been repeatedly traumatised, emotion-focused coping may not be maladaptive, whereas that appears to be the case for men who have been exposed to multiple trauma. Although the study did not specifically examine PTSD symptoms, it is possible that a similar relationship between sex, degree of exposure, and emotion-focused coping exists for PTSD. Furthermore, because the doseresponse model of trauma may refer to both repeated exposure and exposure to more toxic

differences.

trauma types (Elklit & Christiansen, in press), it is possible that this relationship holds true not only for number of traumatic events but also for severity of traumatic events. Many studies on the relationship between coping and PTSD examine coping across situations in general, instead of specifically in relation to the traumatic event. As the preference, and especially the efficacy, of different coping strategies are likely to differ according to the characteristics of the situation at hand, studies which assess coping in relation to the specific traumatic event may be best suited to illuminate the complex relationship between coping and PTSD in males and females.

#### **6.5 Evaluation of the moderation hypothesis**

To sum up, there is some support for the moderation hypothesis, although results are inconsistent. Studies indicate that peritraumatic physiological arousal may be predictive of later PTSD in males but not in females. There is also some support that peritraumatic and persistent dissociation are more predictive of later PTSD in females. Bryant et al. (2000) found that later PTSD status was best identified through either increased heart rate or a diagnosis of ASD, indicating that two distinct pathways may lead to PTSD. There is thus some support for the hypothesis that males may primarily follow an arousal pathway and females a dissociation pathway to PTSD, although these findings are only preliminary and more research is needed. Some support was also found for a moderation effect on social support and PTSD. Although the association between social support and PTSD was stronger in females than in males, consistent with the tend-and-befriend theory and the greater role of oxytocin in females, it is unknown how such findings relate to the proposed sex-specific pathways to PTSD. One possibility is that the positive effects of social support and the negative effects of dissociation are somehow combined in the dissociation pathway to PTSD, possibly because dissociation occurs when tend-and-befriend behaviour proves ineffective, as suggested by Perry et al. (1995). However, another possibility is that lack of social support may represent an independent pathway to PTSD. Furthermore, support for the moderation hypothesis did not appear to be as strong for anxiety as was the case for arousal. It has been suggested that physiological arousal associated with the fight-or-flight system is related to fear but not anxiety (Grillon, 2008). It is therefore possible that anxiety and physiological arousal may not be combined in a single pathway to PTSD, but may instead represent similar but independent pathways. Finally, there was not much support for the moderation hypothesis in relation to coping. It is thus possible, that if the avoidance pathway can be found in adults, it may not be moderated by sex.

Figure 1 shows a preliminary model of how sex differences in the acute response to trauma may affect the development of PTSD. It is hypothesised that exposure to a traumatic stressor results in activation of the SNS in both sexes. However, once the SNS has activated the HPA axis, sex differences in the stress response begin to show. In males, the SNS response is amplified by AVP, which is secreted as part of the HPA response and fuelled by testosterone. This sustained activation of the SNS and increased activity in the HPA axis prepares the body for fighting or fleeing. If the SNS and the HPA axis are not downregulated and brought back to homeostasis, this fight-or-flight response may persist and make up one or more pathways to PTSD. Such pathways are hypothesised to be mediated by physiological arousal and anxiety and possibly by avoidance coping, and problem-focused coping. In contrast, HPA activation in females is associated with the secretion of oxytocin, the effects of which are amplified by oestrogen. Oxytocin is assumed

associated with the later development of PTSD (e.g. Ozer et al., 2003). Furthermore, because oxytocin appears to play a greater role in the female stress response as well as in other aspects of female behaviour, females are likely to be more vulnerable to dysregulation of the oxytocin system than males. Similarly, it is not the tend-and-befriend response in and of itself that leads to higher rates of PTSD in females. Rather, it has been suggested that PTSD develops when the tend-and-befriend response is compromised (Olff et al., 2007). This idea is consistent with studies suggesting that social support serves as a protective factor to a greater extent in females than what appears to be the case in males. Thus, it is possible that females rely on the tend-and-befriend response as protection against stressors, but when it is not possible to seek protection among others, as is often the case with interpersonal assault, the same physiological processes underlying the tend-and-befriend response may lead to dissociation as an alternative protection against the stressor. Thus, the females stress response may be more vulnerable to outside influences (e.g. a perpetrator who does not respond to the pleas of the victim or significant others who fail to understand the victim's distress) than the male stress response. This may be part of the reason why females appear

Finally, it should be mentioned that moderation effects have been reported in relation to other risk factors not reported here. For example, depression, negative affectivity/ neuroticism, and younger age have sometimes been reported to be related to PTSD in females more than in males, whereas prior traumatic exposure has been reported to predict PTSD in males only (for an overview see Christiansen & Elklit, 2008). Sex differences in such risk factors do not appear to fit easily in the model introduced above. In fact, depression has been linked to increased SNS activity (Veith et al., 1994) as well as HPA hyper-reactivity (Pariante & Lightman, 2008). In the proposed model, PTSD in females is associated with a suppression of SNS activity and hypo-reactivity in the HPA axis, and it is thus unclear why depression should be particularly related to PTSD in females. More research on sex differences in the physiological correlates of PTSD as well as depression is needed in order to understand these apparently contradictory results. Furthermore, based on data from the NCS, Bromet, Sonnega, and Kessler (1998) reported that more risk factors predicted PTSD in female than male trauma survivors. Similarly, King and colleagues (1999) reported that the relative importance of pre-, peri-, and post-traumatic variables in predicting PTSD differed for male and female Vietnam veterans. These findings suggest not only that the relationship between different risk factors and the presence or severity of PTSD may be different for males and females, but also that we have identified more of the risk factors involved in the development of PTSD in females than in males. Therefore, in order to better understand and prevent the development of PTSD, more studies need to examine risk factors on PTSD

Even though gender based theories may account for mediation effects of sex in PTSD, such as why females use more emotion-focused coping strategies, we do not believe that genderspecific socialisation and differences in the daily stressors that males and females are exposed to are likely to account for the moderation effects of sex in PTSD, which have been discussed here. Such moderation effects are more likely to arise from sex differences in the physiological response to stress and its sequalae. For example, isolation has been reported to serve as a stressor in female but not male rats. Similarly being with other rats has been found to reduce the adverse effects of chronic foot shock stress in female rats, while increasing such effects in males (Westenbroek et al, 2003). This suggests that sex differences

to be more vulnerable to the negative sequalae of trauma.

separately in males and females.

to suppress SNS and HPA activity and instead be associated with activation of the PNS. This triggers the tend-and-befriend response. Although the role played by dissociation in the tend-and-befriend response is unknown, it appears that both are associated with a suppression of SNS activity, and dissociation seems to often co-occur with affiliative behaviour. Whether sensitisation of the tend-and-befriend response and dissociative reactions make up separate or combined pathways to PTSD is unknown. However, such pathways may be mediated by risk factors such as social support, dissociation, and possibly emotion-focused coping. Testosterone and oestrogen levels are not hypothesised to be affected by trauma exposure in the model. Rather, they should be viewed as ways in which sex may influence the initial stress response in males and females. Even though there are obvious differences in the distribution of sex hormones, both oestrogen and testosterone are present and may influence the stress response in both sexes. Thus, although the tend-andbefriend response is hypothesised to dominate in females and the fight-or-flight response is hypothesised to dominate in males, both response patterns and the associated pathways to PTSD may exist in both sexes.

Fig. 1. A model of potential sex differences in initial stress response and pathways to PTSD

On a final note, the hypothesised involvement of oxytocin and social support in the development of PTSD in females should not be taken to suggest that the release of oxytocin and support seeking behaviour are maladaptive. In fact, higher oxytocin levels have been associated with faster HPA recovery in females after an acute stress laboratory challenge (Olff et al., 2007). Thus, it is most likely that it is a dysregulated oxytocin response which is related to PTSD in females, possibly because the downregulation of the SNS associated with the secretion of oxytocin may trigger dissociative reactions, which appear to be directly

to suppress SNS and HPA activity and instead be associated with activation of the PNS. This triggers the tend-and-befriend response. Although the role played by dissociation in the tend-and-befriend response is unknown, it appears that both are associated with a suppression of SNS activity, and dissociation seems to often co-occur with affiliative behaviour. Whether sensitisation of the tend-and-befriend response and dissociative reactions make up separate or combined pathways to PTSD is unknown. However, such pathways may be mediated by risk factors such as social support, dissociation, and possibly emotion-focused coping. Testosterone and oestrogen levels are not hypothesised to be affected by trauma exposure in the model. Rather, they should be viewed as ways in which sex may influence the initial stress response in males and females. Even though there are obvious differences in the distribution of sex hormones, both oestrogen and testosterone are present and may influence the stress response in both sexes. Thus, although the tend-andbefriend response is hypothesised to dominate in females and the fight-or-flight response is hypothesised to dominate in males, both response patterns and the associated pathways to

Fig. 1. A model of potential sex differences in initial stress response and pathways to PTSD On a final note, the hypothesised involvement of oxytocin and social support in the development of PTSD in females should not be taken to suggest that the release of oxytocin and support seeking behaviour are maladaptive. In fact, higher oxytocin levels have been associated with faster HPA recovery in females after an acute stress laboratory challenge (Olff et al., 2007). Thus, it is most likely that it is a dysregulated oxytocin response which is related to PTSD in females, possibly because the downregulation of the SNS associated with the secretion of oxytocin may trigger dissociative reactions, which appear to be directly

PTSD may exist in both sexes.

associated with the later development of PTSD (e.g. Ozer et al., 2003). Furthermore, because oxytocin appears to play a greater role in the female stress response as well as in other aspects of female behaviour, females are likely to be more vulnerable to dysregulation of the oxytocin system than males. Similarly, it is not the tend-and-befriend response in and of itself that leads to higher rates of PTSD in females. Rather, it has been suggested that PTSD develops when the tend-and-befriend response is compromised (Olff et al., 2007). This idea is consistent with studies suggesting that social support serves as a protective factor to a greater extent in females than what appears to be the case in males. Thus, it is possible that females rely on the tend-and-befriend response as protection against stressors, but when it is not possible to seek protection among others, as is often the case with interpersonal assault, the same physiological processes underlying the tend-and-befriend response may lead to dissociation as an alternative protection against the stressor. Thus, the females stress response may be more vulnerable to outside influences (e.g. a perpetrator who does not respond to the pleas of the victim or significant others who fail to understand the victim's distress) than the male stress response. This may be part of the reason why females appear to be more vulnerable to the negative sequalae of trauma.

Finally, it should be mentioned that moderation effects have been reported in relation to other risk factors not reported here. For example, depression, negative affectivity/ neuroticism, and younger age have sometimes been reported to be related to PTSD in females more than in males, whereas prior traumatic exposure has been reported to predict PTSD in males only (for an overview see Christiansen & Elklit, 2008). Sex differences in such risk factors do not appear to fit easily in the model introduced above. In fact, depression has been linked to increased SNS activity (Veith et al., 1994) as well as HPA hyper-reactivity (Pariante & Lightman, 2008). In the proposed model, PTSD in females is associated with a suppression of SNS activity and hypo-reactivity in the HPA axis, and it is thus unclear why depression should be particularly related to PTSD in females. More research on sex differences in the physiological correlates of PTSD as well as depression is needed in order to understand these apparently contradictory results. Furthermore, based on data from the NCS, Bromet, Sonnega, and Kessler (1998) reported that more risk factors predicted PTSD in female than male trauma survivors. Similarly, King and colleagues (1999) reported that the relative importance of pre-, peri-, and post-traumatic variables in predicting PTSD differed for male and female Vietnam veterans. These findings suggest not only that the relationship between different risk factors and the presence or severity of PTSD may be different for males and females, but also that we have identified more of the risk factors involved in the development of PTSD in females than in males. Therefore, in order to better understand and prevent the development of PTSD, more studies need to examine risk factors on PTSD separately in males and females.

Even though gender based theories may account for mediation effects of sex in PTSD, such as why females use more emotion-focused coping strategies, we do not believe that genderspecific socialisation and differences in the daily stressors that males and females are exposed to are likely to account for the moderation effects of sex in PTSD, which have been discussed here. Such moderation effects are more likely to arise from sex differences in the physiological response to stress and its sequalae. For example, isolation has been reported to serve as a stressor in female but not male rats. Similarly being with other rats has been found to reduce the adverse effects of chronic foot shock stress in female rats, while increasing such effects in males (Westenbroek et al, 2003). This suggests that sex differences

One study has found that males and females seeking treatment for PTSD related to interpersonal assault expressed comparable levels of PTSD severity, depression, guilt, and trait anger (Galovski et al., 2011). However, the same study found that males reported higher levels of state anger and females reported more health-related complaints. Unfortunately, symptoms of dissociation and physiological arousal were not included in this study. The findings suggest that there are minor differences in relation to features associated with PTSD in males and females, but that PTSD levels are comparable. In accordance with this view, it has been reported that males and females with PTSD show equal levels of functional impairment (Schonfeld et al., 1997). However, despite presenting with PTSD symptomatology very similar to that in females, the increased levels of state anger in males suffering from PTSD may result in specific problems. If males respond with anger when being reminded of the trauma, they may be more likely than females to push others away. PTSD constitutes a burden to the friends and family of the person suffering from PTSD and may lead to decreases in support availability over time (Jakupcak et al., 2006; Kaniasty & Norris, 2008). Although both males and females with PTSD may experience a decrease in social support as their symptoms persist, this phenomenon has been particularly reported for male combat veterans with PTSD (Kaniasty & Norris, 2008). In contrast, an increase in health-related complaints is less likely to result in the withdrawal of support and may in fact cause females to seek more support from both health professionals and informal support sources. Longitudinal studies are needed to illuminate how the relationship between social support and PTSD develops over time, and whether such developments differ between males and females. For example, it is possible that social support is a better predictor of PTSD symptomatology in females, but that PTSD leads to

In accordance with the relative similarity in symptomatology found in males and females with PTSD, studies of comorbidity generally find more similarities than differences. One study on treatment seeking patients with PTSD reported that males and females presented with fairly comparable clinical profiles overall, but that males with PTSD experienced more comorbid disorders than did females, particularly substance abuse and antisocial personality disorder (APD; Zlotnick et al., 2001). De Bellis & Keshavan (2003) have demonstrated that males show evidence of more adverse neurobiological consequences of childhood trauma than do females, and that PTSD related to childhood maltreatment appears to be more strongly associated with conduct disorder and criminal offences in males than in females. It is therefore possible that the increased prevalence of APD in males compared to females with PTSD may be particularly apparent in relation to childhood trauma. However, we have no knowledge of any studies examining this possibility. Research on sex differences in comorbidity is scarce as most comorbidity research has been

Although the studies mentioned above suggest that males and females display similar symptoms of post-traumatic stress, such findings have been based on participants already diagnosed with PTSD. It is possible that males tend to exhibit posttraumatic symptomatology other than the re-experiencing, avoidance, and arousal symptoms captured by the PTSD diagnosis, and that this can account for why there are more females than males with PTSD. As mentioned above, PTSD is more often associated with APD in males than in females. However, this sex difference in APD is also found in the general population (Cale & Lilienfeld, 2002). As APD has been repeatedly reported to be related to childhood trauma (Lobbestael & Arntz, 2010), it may be a common way for post-traumatic

greater deterioration of social support in males.

conducted on male samples or has not looked at sex differences.

in the beneficial effects of social support cannot fully be accounted for by gender differences. Instead, such results are consistent with an evolutionary pressure towards a tend-andbefriend response in females. However, whereas evolutionary-based sex differences in the initial stress response may best account for any consistent sex differences in different pathways to PTSD, the results reviewed here suggest that although there may be sex differences in the importance of the different risk factors, none of them appeared to be uniquely associated with either males or females. Thus, the different risk factors are likely to work together in very complex ways, and whether any particular individual follows a specific pathway to PTSD is likely to depend upon a number of different factors, including biology, contextual factors, and gender role. We believe that the role played by sex in the development of PTSD is one of mediated moderation, with sex influencing both the distribution of risk and protective factors and the relationship between such factors and the development and severity of PTSD.

#### **7. Two disorders or one? Sex differences in the expression of traumatisation**

When the PTSD diagnosis was introduced into the DSM-III (American Psychiatric Association, 1980), the different syndromes created to describe the symptoms reported by male war veterans and female survivors of rape and domestic violence were combined. However, if males and females follow different pathways to PTSD, which are mediated by different risk factors, it is possible that there are sex differences in the symptomatology of PTSD. It may therefore be time to reconsider, whether the post-traumatic symptomatology experienced by males and females is best captured by the present PTSD diagnosis, or whether we should go back to diagnosing trauma victims according to sex-specific syndromes. However, this question may prove very difficult to answer. Whereas numerous studies have examined sex differences in exposure to trauma and in the prevalence and severity of PTSD, sex differences in the expression of PTSD have been less extensively studied.

Preliminary support has been reported for a dissociative subtype of PTSD, which is associated with diminished physiological reactivity, and which may be more common in females than in males (Griffin et al., 1997; Olff et al., 2007). Furthermore, there is some support for the hypothesis that males and females with PTSD differ on measures of arousal and HPA disturbances. Although HPA activity is greater among healthy females than healthy males in response to acute stressors, studies have shown that male MVA victims with PTSD appear to have greater HPA arousal than females, although both sexes still display lower overall arousal levels than healthy controls (Paris et al., 2010). Based on these findings, it would be easy to assume, that males will score higher on symptoms of arousal than females. However, as mentioned earlier females experience more PTSD symptoms across all three symptom clusters than do males. The reason for this may be that arousal in PTSD research is often used to cover a wide range of increases in physiological activity. However, certain types of arousal may be associated with PTSD, whereas others may not. For example, PTSD in males has been reported to be related to elevated heart rate but not blood pressure (Bryant et al., 2000; Shalev et al., 1998), and it has been suggested that this is due predominantly to adrenergic rather than noradrenergic activation. Therefore, when considering the specific role of physiological arousal in male compared to female PTSD, it is important to note that the increased acute arousal, which may be associated with increased PTSD severity at later stages, may not be the same type of arousal as that included in the DSM-IV PTSD diagnosis (Hawk et al., 2000).

in the beneficial effects of social support cannot fully be accounted for by gender differences. Instead, such results are consistent with an evolutionary pressure towards a tend-andbefriend response in females. However, whereas evolutionary-based sex differences in the initial stress response may best account for any consistent sex differences in different pathways to PTSD, the results reviewed here suggest that although there may be sex differences in the importance of the different risk factors, none of them appeared to be uniquely associated with either males or females. Thus, the different risk factors are likely to work together in very complex ways, and whether any particular individual follows a specific pathway to PTSD is likely to depend upon a number of different factors, including biology, contextual factors, and gender role. We believe that the role played by sex in the development of PTSD is one of mediated moderation, with sex influencing both the distribution of risk and protective factors and the relationship between such factors and the

**7. Two disorders or one? Sex differences in the expression of traumatisation**  When the PTSD diagnosis was introduced into the DSM-III (American Psychiatric Association, 1980), the different syndromes created to describe the symptoms reported by male war veterans and female survivors of rape and domestic violence were combined. However, if males and females follow different pathways to PTSD, which are mediated by different risk factors, it is possible that there are sex differences in the symptomatology of PTSD. It may therefore be time to reconsider, whether the post-traumatic symptomatology experienced by males and females is best captured by the present PTSD diagnosis, or whether we should go back to diagnosing trauma victims according to sex-specific syndromes. However, this question may prove very difficult to answer. Whereas numerous studies have examined sex differences in exposure to trauma and in the prevalence and severity of PTSD,

Preliminary support has been reported for a dissociative subtype of PTSD, which is associated with diminished physiological reactivity, and which may be more common in females than in males (Griffin et al., 1997; Olff et al., 2007). Furthermore, there is some support for the hypothesis that males and females with PTSD differ on measures of arousal and HPA disturbances. Although HPA activity is greater among healthy females than healthy males in response to acute stressors, studies have shown that male MVA victims with PTSD appear to have greater HPA arousal than females, although both sexes still display lower overall arousal levels than healthy controls (Paris et al., 2010). Based on these findings, it would be easy to assume, that males will score higher on symptoms of arousal than females. However, as mentioned earlier females experience more PTSD symptoms across all three symptom clusters than do males. The reason for this may be that arousal in PTSD research is often used to cover a wide range of increases in physiological activity. However, certain types of arousal may be associated with PTSD, whereas others may not. For example, PTSD in males has been reported to be related to elevated heart rate but not blood pressure (Bryant et al., 2000; Shalev et al., 1998), and it has been suggested that this is due predominantly to adrenergic rather than noradrenergic activation. Therefore, when considering the specific role of physiological arousal in male compared to female PTSD, it is important to note that the increased acute arousal, which may be associated with increased PTSD severity at later stages, may not be the same type of arousal as that included in the

sex differences in the expression of PTSD have been less extensively studied.

development and severity of PTSD.

DSM-IV PTSD diagnosis (Hawk et al., 2000).

One study has found that males and females seeking treatment for PTSD related to interpersonal assault expressed comparable levels of PTSD severity, depression, guilt, and trait anger (Galovski et al., 2011). However, the same study found that males reported higher levels of state anger and females reported more health-related complaints. Unfortunately, symptoms of dissociation and physiological arousal were not included in this study. The findings suggest that there are minor differences in relation to features associated with PTSD in males and females, but that PTSD levels are comparable. In accordance with this view, it has been reported that males and females with PTSD show equal levels of functional impairment (Schonfeld et al., 1997). However, despite presenting with PTSD symptomatology very similar to that in females, the increased levels of state anger in males suffering from PTSD may result in specific problems. If males respond with anger when being reminded of the trauma, they may be more likely than females to push others away. PTSD constitutes a burden to the friends and family of the person suffering from PTSD and may lead to decreases in support availability over time (Jakupcak et al., 2006; Kaniasty & Norris, 2008). Although both males and females with PTSD may experience a decrease in social support as their symptoms persist, this phenomenon has been particularly reported for male combat veterans with PTSD (Kaniasty & Norris, 2008). In contrast, an increase in health-related complaints is less likely to result in the withdrawal of support and may in fact cause females to seek more support from both health professionals and informal support sources. Longitudinal studies are needed to illuminate how the relationship between social support and PTSD develops over time, and whether such developments differ between males and females. For example, it is possible that social support is a better predictor of PTSD symptomatology in females, but that PTSD leads to greater deterioration of social support in males.

In accordance with the relative similarity in symptomatology found in males and females with PTSD, studies of comorbidity generally find more similarities than differences. One study on treatment seeking patients with PTSD reported that males and females presented with fairly comparable clinical profiles overall, but that males with PTSD experienced more comorbid disorders than did females, particularly substance abuse and antisocial personality disorder (APD; Zlotnick et al., 2001). De Bellis & Keshavan (2003) have demonstrated that males show evidence of more adverse neurobiological consequences of childhood trauma than do females, and that PTSD related to childhood maltreatment appears to be more strongly associated with conduct disorder and criminal offences in males than in females. It is therefore possible that the increased prevalence of APD in males compared to females with PTSD may be particularly apparent in relation to childhood trauma. However, we have no knowledge of any studies examining this possibility. Research on sex differences in comorbidity is scarce as most comorbidity research has been conducted on male samples or has not looked at sex differences.

Although the studies mentioned above suggest that males and females display similar symptoms of post-traumatic stress, such findings have been based on participants already diagnosed with PTSD. It is possible that males tend to exhibit posttraumatic symptomatology other than the re-experiencing, avoidance, and arousal symptoms captured by the PTSD diagnosis, and that this can account for why there are more females than males with PTSD. As mentioned above, PTSD is more often associated with APD in males than in females. However, this sex difference in APD is also found in the general population (Cale & Lilienfeld, 2002). As APD has been repeatedly reported to be related to childhood trauma (Lobbestael & Arntz, 2010), it may be a common way for post-traumatic

which aims to reduce levels of dissociation and increase levels of social support. In contrast, a therapeutic approach, which aims to reduce physiological arousal and dampen anxiety may be more beneficial to males. In accordance with these ideas, a recent pilot study has indicated that propranolol, a beta-blocker known to reduce heart rate and blood pressure, may decrease PTSD severity in males but increase PTSD severity in females (Nugent et al., 2010). Whereas there is some support for the hypothesis that emotion-focused coping is more beneficial in females than in males, it is possible that males as well as females will benefit more from a therapeutic approach, which aims to strengthen the coping strategies, which they do not automatically use. The idea is, that once PTSD has developed, it appears that the preferred coping strategies have proved ineffective. Some support has been reported for this idea in grief counselling, as problem-focused counselling was found to be more effective in females compared to males, whereas an emotion-focused intervention form appeared to be more effective in males compared to females 11 months postbereavement. However, such sex differences may not be transferable to less chronic

Finally, an important implication of the influence of multiple factors on HPA reactivity is the possibility that the disturbed HPA reactivity found in patients with PTSD can be treated through either pharmacological, psychotherapeutic, or social interventions. The efficacy of different treatment strategies may therefore not be dependent on what caused the HPA disturbances in the first place. Unfortunately, sex differences in treatment outcome are grossly understudied. Many PTSD treatment effect studies are based on sex-specific trauma samples, such as war veterans and sexual assault victims. Furthermore, of the few studies, which do include both-sex samples, few examine the impact of sex on treatment efficacy. As a result, little is known about whether the same treatments are equally effective (or noneffective) in males and females. Blain, Galovski, and Robinson (2010) reviewed the literature on sex differences in response to treatment and found only nine randomised controlled trials, which assessed sex differences in primary PTSD treatment outcome. Most of these studies were based on samples, which were too small to detect minor sex differences. Furthermore, 6/9 studies were based on samples exposed to a variety of trauma types, and not a single study examined sex differences in treatment outcome following interpersonal violence. Results from randomised as well as non-randomised clinical trials were mixed regarding the impact of sex on treatment outcome. The majority of studies reported no sex differences, but others found that females may respond better to trauma focused therapy than males, and that males may be more likely to drop out of treatment. However, more

than anything this review highlights the need for more research in the area.

As stated earlier, we believe that the research on sex differences in PTSD is still in its childhood. Consistent sex differences still remain to be documented following certain types of trauma (e.g. sexual assault, combat exposure). Furthermore, the degree to which sex differences in risk factors associated with PTSD may account for the increased vulnerability of PTSD in females is still unknown and deserves more attention. However, future research should move beyond simply focusing on establishing and explaining sex differences in

Sex differences have been reported in the initial response to threat, but the degree to which the tend-and-befriend response is dominant in females in response to specifically traumatic

**9. Future research on sex differences in PTSD** 

exposure and PTSD.

populations or to PTSD treatment.

symptomatology to be expressed in males independently of the PTSD diagnosis. Although irritability and aggressive outbursts are not uncommon in people suffering from PTSD, particularly in males, the remaining symptoms are mostly internalising. It has been reported that females are more likely to suffer from internalising disorders such as anxiety disorders and depression, whereas males are more likely to suffer from externalising disorders, such as substance abuse or conduct disorder in the general population (Kessler et al., 1995) and following traumatic exposure (Pratchett et al., 2010). Similarly, although results have been mixed, increases in substance abuse have been registered primarily in male trauma survivors, whereas increases in somatisation, anxiety, and suicide attempts have been documented primarily in female trauma victims (Tolin & Foa, 2008). Furthermore, PTSD in females tends to be more chronic than in males, and females are more prone to develop complex PTSD. The concept of complex PTSD was specifically developed to capture posttraumatic symptoms other than the ones included in the PTSD diagnosis following prolonged and chronic trauma (Cloitre et al., 2002). Complex PTSD covers symptoms similar to those associated with borderline personality disorder (e.g. impulsivity, aggression, self-destructive behaviour, dissociation, and interpersonal problems) as well as feelings of alienation and trust issues (Cloitre et al., 2002).

In sum, there appear to be some sex differences on measures of physiological arousal, state anger, dissociation, somatisation, and personality disorders. Such differences are likely to be caused by a combination of gender differences in how males and females express symptomatology and sex differences in how trauma affects brain development. However, it appears that even though females are at a higher risk of developing PTSD than males, once males do develop PTSD their overall symptomatology does not appear to differ much from that found in females. This suggests that although males and females appear to differ in their initial trauma response and may follow different pathways to PTSD, there is at present no support for the idea that the symptomatology of male and female trauma victims is best captured by more sex-specific syndromes. However, more research focusing on sex differences once PTSD has developed is needed in order to draw any conclusions in relation to this question, especially as both males and females also display some sex-specific symptomatology (e.g. complex PTSD, antisocial personality disorder), which is not fully captured by the PTSD diagnosis.
