Introduction
Sex differences in posttraumatic stress disorder (PTSD) have been documented widely. Across nations, time, study type, or diagnostic criteria, women have been reported to be at higher risk for PTSD than men (Ben-Ezra et al., Reference Ben-Ezra, Karatzias, Hyland, Brewin, Cloitre, Bisson and Shevlin2018; Frans, Rimmo, Aberg, & Fredrikson, Reference Frans, Rimmo, Aberg and Fredrikson2005; Kessler, Sonnega, Bromet, Hughes, & Nelson, Reference Kessler, Sonnega, Bromet, Hughes and Nelson1995; Otten et al., Reference Otten, Tibubos, Schomerus, Brähler, Binder, Kruse and Beutel2021). Among the general population, women are approximately twice as likely to develop PTSD compared to men (Goldstein et al., Reference Goldstein, Smith, Chou, Saha, Jung, Zhang and Grant2016; McCall-Hosenfeld, Mukherjee, & Lehman, Reference McCall-Hosenfeld, Mukherjee and Lehman2014; Seedat et al., Reference Seedat, Scott, Angermeyer, Berglund, Bromet, Brugha and Kessler2009), with the highest reported risk differences ranging up to six-folds greater odds in women (Seedat et al., Reference Seedat, Scott, Angermeyer, Berglund, Bromet, Brugha and Kessler2009). In addition, current research suggests women experience more chronic and severe PTSD symptoms than men (Carmassi et al., Reference Carmassi, Akiskal, Bessonov, Massimetti, Calderani, Stratta and Dell׳Osso2014; Carragher et al., Reference Carragher, Sunderland, Batterham, Calear, Elhai, Chapman and Mills2016; Haering et al., Reference Haering, Schulze, Geiling, Meyer, Klusmann, Schumacher and Engel2024b; Kessler et al., Reference Kessler, Sonnega, Bromet, Hughes and Nelson1995; Tolin & Foa, Reference Tolin and Foa2006). While female sex is sometimes considered a PTSD risk factor itself, settling at this point disregards the underlying mechanisms that drive these risk differences. Rather than oversimplifying this relationship to a maxim in which women fundamentally are at higher vulnerability to adverse trauma-related outcomes, a better understanding of which risk factors affect whom, when, and how will allow researchers to design interventions that tackle modifiable constructs such as cognitive, behavioral, or structural processes that are associated with sex.
Several advances have been made in explaining the prominent sex differences in PTSD outcomes. Sexual trauma, for instance, more commonly experienced by women than men (Tolin & Foa, Reference Tolin and Foa2006), has been found to be associated with a higher risk for PTSD compared to other trauma types (Kessler et al., Reference Kessler, Sonnega, Bromet, Hughes and Nelson1995; Perkonigg, Kessler, Storz, & Wittchen, Reference Perkonigg, Kessler, Storz and Wittchen2000). Yet, even when controlling for sexual trauma exposure, sex differences in PTSD prevalence and severity remain (Tolin & Foa, Reference Tolin and Foa2006), and sex differences in PTSD have also been found in samples of non-sexual trauma survivors, such as victims of motor vehicle accidents (Fullerton et al., Reference Fullerton, Ursano, Epstein, Crowley, Vance, Kao and Baum2001). In addition to characteristics of the traumatic event, sex differences in neurobiological processes, such as fear mechanisms (Dark et al., Reference Dark, Harnett, Hurst, Wheelock, Wood, Goodman and Knight2022; Ramikie & Ressler, Reference Ramikie and Ressler2018), as well as physiological (Lalonde et al., Reference Lalonde, Mekawi, Ethun, Beurel, Gould, Dhabhar and Michopoulos2021) and psychosocial risk factors, such as acute stress responses and appraisals (Olff, Reference Olff2017; Olff, Langeland, Draijer, & Gersons, Reference Olff, Langeland, Draijer and Gersons2007) have been examined. More recently, researchers also have started to study the impact of gender-related factors such as gender norms or gender role stress (Christiansen & Berke, Reference Christiansen and Berke2020), and now are slowly beginning to examine the interplay of sex- and gender-related factors (Christiansen, McCarthy, & Seeman, Reference Christiansen, McCarthy and Seeman2022). Tannenbaum, Ellis, Eyssel, Zou, and Schiebinger (Reference Tannenbaum, Ellis, Eyssel, Zou and Schiebinger2019) define sex as referring to biological attributes, and gender as referring to sociocultural factors, such as gender norms, gender identity, or gender relations. As the analyses included in this study are based on sex assigned at birth, we use the term sex in this text to refer to differences between women and men, while simultaneously acknowledging that both sex and gender are important predictors of health, mutually influence each other, and are often intertwined (Krieger, Reference Krieger2003).
Despite recent progress, reviews on sex and/or gender differences in PTSD risk factors unanimously call for an improved consideration of sex and gender in trauma research, as the full picture of what accounts for the disparities in PTSD outcomes still remains unclear (Christiansen et al., Reference Christiansen, McCarthy and Seeman2022; Olff & Langeland, Reference Olff and Langeland2022; Ramikie & Ressler, Reference Ramikie and Ressler2018). This gap in knowledge is not surprising, however, given various methodological challenges in the examination of sex differences in PTSD risk factors: (1) First, risk factors should ideally be examined in a prospective manner. Yet, as it is difficult to foresee traumatic events, designing prospective PTSD studies is challenging. (2) Second, much of the prospective research on PTSD risk factors has been conducted in predominantly male samples, such as soldiers, police officers or firefighters (Eraly et al., Reference Eraly, Nievergelt, Maihofer, Barkauskas, Biswas, Agorastos and Baker2014; Sopp, Michael, Lass-Hennemann, Haim-Nachum, & Lommen, Reference Sopp, Michael, Lass-Hennemann, Haim-Nachum and Lommen2021; Sørensen, Olesen, Midtgaard, & Willert, Reference Sørensen, Olesen, Midtgaard and Willert2022). However, even in more naturalistic samples, such as prospective emergency department studies, men are studied twice as much as women (Haering et al., Reference Haering, Schulze, Geiling, Meyer, Klusmann, Schumacher and Engel2024b). The lack of female representation becomes even more apparent in psychobiological PTSD research, where only 2% of the research has been conducted in females (Olff, Reference Olff2017), as it is feared that the ‘messy’ hormonal variability associated with the female menstrual cycle might confound study results (Bale & Epperson, Reference Bale and Epperson2017; Beery & Zucker, Reference Beery and Zucker2011). While recent research has shown that this dogma is inaccurate (Levy et al., Reference Levy, Hunter, Lin, Robinson, Gillis, Conlin and Datta2023; Wiseman, Reference Wiseman2023), the underrepresentation of females in prospective trauma research makes it hard to yield optimal conditions for the analysis of sex differences (Rechlin, Splinter, Hodges, Albert, & Galea, Reference Rechlin, Splinter, Hodges, Albert and Galea2022). (3) Third, even when males and females are included in studies, sex is infrequently used as a discovery variable to examine potentially different mechanisms in women and men (Haering et al., Reference Haering, Schulze, Geiling, Meyer, Klusmann, Schumacher and Engel2024b; Rechlin et al., Reference Rechlin, Splinter, Hodges, Albert and Galea2022). (4) Even if researchers are willing to conduct discovery analyses, they are faced with a lack of best practice examples on how to examine sex differences in risk factors of mental disorders. Other than in pre-clinical studies, studies on PTSD etiology usually have a limited sample size, experimental manipulation and control is not possible, and different statistical approaches come with variable advantages and disadvantages. This situation has, amongst others, led to a substantial proportion of reported sex differences not supported by sufficient statistical evidence (Garcia-Sifuentes & Maney, Reference Garcia-Sifuentes and Maney2021). (5) Given ongoing controversies on publication bias (Ferguson & Heene, Reference Ferguson and Heene2012) and hesitancy to publish null results, it is unclear if individual studies that published (statistically supported) sex differences in PTSD risk factors represent replicable insights in underlying mechanisms, or whether they are in fact just statistical artifacts. Given these shortcomings, it is still largely unclear which risk factors do or do not have sex-dependent effects in women and men.
We aimed to systematically explore sex differences in PTSD risk factors in a sample of 2924 participants (62% women) enrolled in the AURORA (Advancing Understanding of RecOvery afteR traumA) study, a prospective multisite longitudinal study of the onset and course of adverse posttraumatic neuropsychiatric sequalae. We pre-registered a sex-sensitive framework to systematically explore sex differences in PTSD risk factors with the AURORA consortium (Haering, Stevens, & Powers, Reference Haering, Stevens and Powers2022). The focus of the current study was to determine if a sex-dependent vulnerability to PTSD risk factors might contribute to women's higher PTSD severity 3-months post-acute trauma, i.e. to examine whether sex moderates the association between a risk factor and PTSD in a way that the strength, significance and/or direction of the association differs between women and men. We selected our predictors of interest based on a literature review of sex differences in PTSD predictors (Christiansen, Reference Christiansen, Martin, Preedy and Patel2016). Predictors that are assumed to be associated with PTSD differently for men and women, and were assessed in the AURORA study, were included in this analysis. Specifically, preexisting anxiety symptoms, prior trauma exposure, and peritraumatic distress as well as lower socioeconomic status, being member of a marginalized group, being unmarried or being unemployed have been summarized as risk factors more strongly associated with PTSD in men. On the other hand, preexisting depression symptoms, anxiety sensitivity, neuroticism, peritraumatic life threat and dissociation, as well as lack of social support and acute stress disorder have been summarized as risk factors more strongly associated with PTSD in women. Finally, mixed findings have been reported regarding the interaction of sex and age (Christiansen, Reference Christiansen, Martin, Preedy and Patel2016). In addition to these factors, we further included lifetime sexual assault exposure into our investigation of sex-dependent vulnerabilities. Although sex differences in sexual assault exposure are well-established, the impact of lifetime sexual assault exposure as a risk factor for posttraumatic dysfunction following a new trauma exposure has been less explored. Yet, recent evidence has highlighted the role of prior sexual assault exposure on subsequent trauma exposure (Rowland et al., Reference Rowland, Roeckner, Ely, Lebois, van Rooij, Bruce and Harnett2023). Thus, we determined, whether the vulnerability to the aforementioned 16 pre-, peri-, and post-traumatic PTSD predictors differed between acutely traumatized women and men.
Methods
Participants and procedure
Data from the n = 2924 AURORA participants (n = 1124 men and n = 1818 women) were collected from September 2017 through June 2021. Detailed information on participant characteristics is given in Table 1. An overview of risk factors by sex and by trauma type (mvc v. non-mvc trauma) is presented in eTable 1 and eTable 2 in the Supplement.
Table 1. Demographic and trauma characteristics
Note: Data available for a99.6%, b99.7%, c88.2%, d87.7%, e83.3% of the sample, respectively; fEvent exposing participant and at least several other individuals to traumatic stress, not covered by other categories (e.g. plane crash, natural disaster).
The AURORA study procedures are described in detail elsewhere (McLean et al., Reference McLean, Ressler, Koenen, Neylan, Germine, Jovanovic and Kessler2020). In brief, AURORA participants were adults who had experienced a traumatic event within the past 72 h and were evaluated in one of 29 emergency departments (ED) across the United States. Participants were 18–75 years old, able to speak and read English, able to comprehend the enrollment protocol, and possessed a smartphone and e-mail address. Participants were excluded from data collection if they were or became pregnant or incarcerated. In addition, we excluded one participant with missing information on sex assigned at birth. Participants completed assessments in the ED (baseline), and at scheduled follow-ups, during which information on psychological symptoms, physical health, and functioning was assessed. Data for the analyses in this report were collected via self-report at baseline, 2 weeks, 8 weeks, and 3 months post-trauma. All participants provided written informed consent and were compensated for their study participation at each follow-up. The project was approved by each participating institutional review board. The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
Measures
We assessed 16 PTSD risk factors spanning pre-, peri-, and acute posttraumatic predictors. Pre-traumatic predictors included age, race–ethnicity, marital status, education, income, employment status, pre-trauma depression symptoms, pre-trauma anxiety symptoms, neuroticism and trauma load and lifetime exposure to sexual assault. Peri-traumatic risk factors included peritraumatic distress and perceived life threat; and acute post-traumatic predictors included social support, acute stress disorder, and acute dissociative symptoms. Sex assigned at birth was used as stratification variable. Detailed information on all measures is presented in Supplement 1.
Outcome
The outcome evaluated was self-reported PTSD severity at the 3-month follow-up, as assessed by the PTSD Symptom Checklist for DSM-5 (PCL-5). The PCL-5 is a 20 item self-report questionnaire that assesses the presence and severity of various posttraumatic stress symptoms (Weathers et al., Reference Weathers, Litz, Keane, Palmieri, Marx and Schnurr2013). Participants rated the severity of each symptom on a scale of 0 (not at all) to 4 (extremely), and items were summed to create a total severity score.
Statistical analysis
All analyses were conducted in R version 4.2.1 (R Core Team, 2022). The code for all analyses can be found in the OSF repository: https://osf.io/tkncz/. Missing values for constructs that were assessed longitudinally in the AURORA study (depressive symptoms, anxiety symptoms, acute dissociative symptoms, and social support) were imputed using multiple imputation with predictive mean matching via the aregImpute function of the Hmisc package (Harrell, Reference Harrell2022), including the respective longitudinal assessments of each construct as auxiliary variables. In line with current recommendations for imputing data when moderation analyses are planned, missing values were imputed separately for males and females (Heymans & Eekhout, Reference Heymans and Eekhout2019). Student's t tests and χ2 tests were performed to analyze sex differences in continuous and categorical variables, respectively.
To examine a sex-dependent vulnerability to the selected risk factors, we first calculated sex-disaggregated associations between each predictor and 3-month PTSD severity. As suggested by Christiansen, Olff, and Elklit (Reference Christiansen, Olff and Elklit2014) the equality of coefficients was tested using Fisher's z tests and Zou's confidence intervals, applying the package cocor (Diedenhofen & Musch, Reference Diedenhofen and Musch2015). Adapted from Jun et al. (Reference Jun, Nirenberg, Weinberger, Sharma, Pujadas, Cordon-Cardo and Huang2021) we then performed multivariable regression models controlling for participant demographics, baseline mental health, and life-time sexual assault exposure, and conducted subgroup analyses by including an interaction term between sex and the subgroup variable to the multivariable model. This procedure was done for each subgroup variable to identify significant effect differences between women and men. Finally, to assess the robustness of results across trauma types, we performed sensitivity analyses with the motor vehicle collision subgroup only (n = 2194, 74.6% of the full sample). Categorical variables were dummy coded, continuous variables were standardized for model estimations. Statistical significance was evaluated using 0.05-level two-sided tests. Most of the risk factors were positively associated with each other (see eTable 3). In spite of this inter-correlation, examination of the variance inflation factors (VIF) provided no indication of multicollinearity of the 16 predictors (all VIF < 3).
Results
Women and men differed significantly in PTSD severity at three months post trauma, with women scoring higher on the PCL-5 than men (M (s.d.)f = 26.7 (18.9); M (s.d.)m = 22.3 (19.3); p < 0.001). PTSD severity also differed by sex in the subgroup of participants with motor vehicle collisions (MVC; M (s.d.)f_mvc = 26.6 (18.8); M (s.d.)m_mvc = 22.1 (19.4); p < 0.001), which accounted for 74.6% of index trauma types in this sample. Sexual assault as index trauma was reported by less than 1% of all participants. As shown in Fig. 1, all 16 risk factors revealed statistically significant univariable associations with 3-month PTSD severity in men. In women, these associations were present for all predictors except for age and minority status. Statistical comparison of the male and female correlations demonstrated significant sex-dependent differences in the association of PTSD severity and five risk factors: acute dissociation (difference d, [95% CI] = −0.09 [−0.14 to −0.03], p = 0.002), peritraumatic distress (d = −0.08 [−0.15 to −0.01], p = 0.018), pre-traumatic anxiety symptoms (d = −0.07 [−0.13 to −0.01], p = 0.027), the participant-reported chance of dying during the index event (d = −0.07 [−0.14 to 0.00], p = 0.044), and acute stress disorder (d = −0.05 [−0.09 to −0.00], p = 0.039). As depicted in Fig. 1, the univariable correlations for all these six risk factors with PTSD severity were stronger in men than in women.
Figure 1. Sex-disaggregated associations with 3-month PTSD severity. The forest plot depicts the univariable associations of each predictor with PTSD severity at 3-months post-trauma disaggregated by sex. The correlations are depicted in blue for men and in red for women. The equality of coefficients was tested using Fisher's z tests.
In a multi-variable regression model adjusting for participant demographics, baseline mental health, and lifetime sexual assault exposure, female sex remained an independent predictor of PTSD severity at 3-months (β [s.e.] = 0.13 [0.04], p < 0.001, see eTable 4 in the Supplement). Using this model as baseline model, we next performed subgroup analyses to identify significant interactions between sex and each of our predictors of interest (Fig. 2): Analyses identified a statistically significant interaction effect for pre-traumatic anxiety symptoms (β = −0.11 [0.04], p = 0.005) and acute dissociative symptoms (β = −0.10 [0.04], p = 0.003). Figure 3 depicts the sex-by-risk factor interaction effects under the specified model: the graphs show that women and men diverge at the lower end of the risk factor spectrum, with women showing more PTSD symptoms even at lower levels of anxiety and dissociation. Due to men's higher vulnerability to pre-traumatic anxiety and acute dissociative symptoms, PTSD levels of women and men converge at the upper end of the risk factor spectrum. For the remaining predictors, no significant sex differences in the association with PTSD were detected. Our subgroup analyses identified main effects (but no interaction effects) for acute stress disorder (β = 0.56 [0.03], p < 0.001), anxiety sensitivity (β = 0.44 [0.03], p < 0.001), neuroticism (β = 0.38, [0.03], p < 0.001), lifetime sexual assault (β = 0.29 [0.10], p = 0.005), pre-traumatic depressive symptoms (β = 0.26 [0.04], p < 0.001), peritraumatic distress (β = 0.24 [0.03], p < 0.001), chance of dying (β = 0.20 [0.04], p < 0.001), minority status (β = 0.17 [0.07], p = 0.008), family income < 19 k (β = 0.14 [0.07], p = 0.040), social support (β = −0.13, [0.03], p < 0.001), age (β = 0.09 [0.04], p = 0.006), and trauma load (β = 0.07 [0.04], p = 0.014).
Figure 2. Subgroup analysis for female sex as a predictor of PTSD severity at 3 months. The forest plot depicts the parameter estimate and 95% confidence interval associated with female sex (v. male sex as the reference group) within the subgroup specified in a multivariable regression model, including an interaction term between the subgroup variable and sex, adjusting for demographics, baseline mental health and lifetime sexual assault. Continuous variables were standardized, categorical variables were dummy coded.
Figure 3. Interaction effects of sex and (a) pre-traumatic anxiety and (b) dissociation at week two post-trauma. The scatter plots depict the interaction effect of sex and (a) pre-traumatic anxiety symptoms as well as (b) dissociative symptoms at two weeks post-trauma, controlling for demographics, baseline mental health, and lifetime sexual assault. All variables were standardized to an overall sample mean of 0 and a standard deviation of 1.
Sensitivity analyses for the motor vehicle collision (MVC) subgroup showed that sex differences in univariable comparisons remained robust for twelve of the 16 risk factors assessed. Compared to the main analyses, sex differences were no longer found for pre-traumatic anxiety symptoms and participant-reported chance of dying during the index event (see eTable 5). However, among the MVC subgroup significant sex differences were observed for the associations of 3-month PTSD severity with trauma load (d = 0.11 [0.01–0.20], p = 0.027) and lifetime sexual assault exposure (d = 0.14 [0.04–0.24], p = 0.005). In contrast to previous results among the full sample, analyses in the MVC subgroup thus revealed two risk factors for which associations with 3-month PTSD severity were stronger in female than male MVC-exposed individuals, and both reflected exposure to prior stressors, whereas the major risk factors favoring men reflected internal trait-like factors.
In MVC subgroup analyses of the multivariable models, sex remained a robust moderator of acute dissociative symptoms (β = −0.11 [0.04], p = 0.007). Moreover, an interaction of sex with anxiety sensitivity was observed (β = −0.09 [0.05], p = 0.044), albeit with the 95% confidence interval nearly including zero. Detailed results of the multivariable MVC subgroup analyses are presented in eTable 6.
Discussion
The current study systematically examined sex differences in the vulnerability to PTSD risk factors to better understand which processes drive the prominent sex disparities in PTSD outcomes. In univariable analyses we found sex-dependent associations with 3-month PTSD severity in five of 16 risk factors, which previously had been hypothesized to show different associations with PTSD for women and men: pre-traumatic anxiety symptoms, peritraumatic distress, perceived chance of dying during the trauma, acute dissociative symptoms, and acute stress disorder symptoms showed stronger associations with PTSD severity at 3 months in men than in women. Two of these predictors, pre-traumatic anxiety and acute dissociative symptoms, also showed sex-dependent effects in multivariable models with interaction terms, indicating men might be more vulnerable to these risk factors than women, even when controlling for demographics and baseline mental health. Our main analyses did not indicate any risk factor to which women were more vulnerable, which was surprising given the selection of predictors based on prior research (Christiansen, Reference Christiansen, Martin, Preedy and Patel2016). Subgroup analyses of an MVC-only sample suggested trauma type-conditional sex vulnerabilities for trauma load and lifetime sexual assault exposure. Finally, a number of known PTSD risk factors showed similar predictive value for females and males, both in the main and subgroup analysis, including age, income below poverty line, racial/ethnic minority status, social support, trauma severity, and a variety of symptoms and personality factors.
Our study suggests how underlying processes may contribute differentially toward PTSD severity in men and in women. The results of our main analyses highlight two risk factors that consistently showed a stronger impact on men compared to women. First, pre-existing anxiety symptoms were more strongly related to PTSD severity in men than in women. This finding is in line with previous research, suggesting traumatized men are more vulnerable to pre-existing anxiety than women (Bromet, Sonnega, & Kessler, Reference Bromet, Sonnega and Kessler1998; Christiansen & Elklit, Reference Christiansen and Elklit2008, Reference Christiansen, Elklit and Ovuga2012). Second, acute dissociation also showed stronger associations with PTSD severity in men than in women. Given previous evidence on sex-dependent effects of dissociation (Bryant & Harvey, Reference Bryant and Harvey2003; Christiansen & Elklit, Reference Christiansen and Elklit2008; Fullerton et al., Reference Fullerton, Ursano, Epstein, Crowley, Vance, Kao and Baum2001), which suggested a stronger negative impact in women, this finding is unexpected. However, among prior studies, peritraumatic dissociative experiences rather than acute dissociative responses were examined (Fullerton et al., Reference Fullerton, Ursano, Epstein, Crowley, Vance, Kao and Baum2001), and reporting of sex differences was not always supported by sufficient statistical evidence (Bryant & Harvey, Reference Bryant and Harvey2003; Christiansen & Elklit, Reference Christiansen and Elklit2008; Garcia-Sifuentes & Maney, Reference Garcia-Sifuentes and Maney2021). Interpreting our results, it might be that men with high levels of dissociative experience are at greater risk for PTSD symptoms than women, as greater mental health stigma among men amplifies the negative impact of maladaptive cognitive beliefs. Previous evidence has shown a link between negative appraisals of initial PTSD symptoms (such as dissociative experiences) with later PTSD outcomes (Brown, Wood, Carter, & Kannis-Dymand, Reference Brown, Wood, Carter and Kannis-Dymand2022; Kannis-Dymand, Carter, Lane, & Innes, Reference Kannis-Dymand, Carter, Lane and Innes2019). Such negative appraisals include for instance interpretations of PTSD symptoms as meaning ‘I am going crazy’ (Steil & Ehlers, Reference Steil and Ehlers2000). Given that men seem to hold greater stigmatizing beliefs about mental health symptoms (Bradbury, Reference Bradbury2020; Chandra & Minkovitz, Reference Chandra and Minkovitz2006), it might be that seemingly ‘abnormal’ experiences like dissociative experiences are more strongly linked to PTSD severity in men than in women, as men might have greater negative interpretations about them, which in turn affects PTSD development and maintenance.
Of note, within our study the majority of risk factors showed tendencies toward stronger negative impacts in men, i.e. a male vulnerability. Despite women's higher PTSD risk, these findings warrant the need for a closer examination of PTSD risk factors from a men-centered perspective. For men, our results suggest possible mechanisms that might be targeted when designing sex-sensitive preventive interventions, especially since men benefit less from trauma-focused treatment interventions for manifested PTSD. Exploring effective prevention options that integrate our full knowledge on sex-dependent mechanisms is needed.
Although our main analyses did not identify any risk factors to which women were more vulnerable than men, sensitivity analyses point toward trauma type-conditional effects: In sub-group analyses of the MVC sample, trauma load and lifetime sexual assault exposure entered in as sex-dependent risk factors, showing greater univariable associations with 3-month PTSD severity in women than in men. Interestingly, only MVC-specific analyses revealed risk factors to which women were more vulnerable to than men. Our subgroup analyses of the MVC sample thus suggest that for MVC-exposed individuals, men might be more vulnerable to internal factors, whereas women might be more vulnerable to external, exposure-based factors. One reason for the differences in the main v. subgroup analyses might be the fact that after an MVC, individuals are generally less likely to develop PTSD symptoms compared to other trauma types (e.g. interpersonal trauma; Kessler et al. (Reference Kessler, Aguilar-Gaxiola, Alonso, Benjet, Bromet, Cardoso and Koenen2017)). Thus, predisposing factors such as exposure-based predictors for women (i.e. lifetime sexual assault exposure, trauma load), might carry more weight in accidental traumas than in assault-based traumas, where individuals are at a higher risk for PTSD symptom development and variables such as perpetration characteristics may become more salient. Clearly, our findings once more are a testament to the intricate interplay of sex and trauma type which needs to be better understood in order to realize the full potential of targeted interventions.
Importantly, even though women in our main sample were less affected by the negative impacts of pre-existing anxiety or acute dissociation symptoms, they largely still experienced greater PTSD symptoms than men. Our results highlight the need to explore further aspects that contribute to women's higher PTSD severity, including different constructs, pathways, and mechanisms. For instance, understanding how sex interacts with cognitive factors might provide modifiable targets for early interventions. Furthermore, greater knowledge about hormonal (Ney, Gogos, Ken Hsu, & Felmingham, Reference Ney, Gogos, Ken Hsu and Felmingham2019) or genetic (Yu et al., Reference Yu, Chen, Pan, Kurz, Datner, Hendry and Linnstaedt2018) factors can increase our knowledge of sex-based contributors to PTSD. Finally, independent of vulnerability differences, i.e. differences in the strength of an association between a risk factor and PTSD severity, differences in risk factor prevalence/severity among women and men can further contribute to sex differences in PTSD severity. Thus, greater exposure to certain risk factors (Christiansen & Hansen, Reference Christiansen and Hansen2015) might help to explain women's higher PTSD severity in the present sample. A systematic overview of risk pathways between sex and PTSD as well as a complementary analysis is presented in Haering et al. (Reference Haering, Seligowski, Linnstaedt, Michopoulos, House, Beaudoin and Stevens2024c). A clear distinction between various forms of sex-related effects will not only help to better understand sex differences in mental health outcomes, it will also help to create a more rigorous and replicable mental health science.
Several limitations need to be taken into consideration when looking at our findings. Possible selection bias in who gets treated in the ED after trauma may impact generalizability of results. While EDs visits present a unique opportunity for providing traumatized individuals with secondary interventions, it is unclear, to what extent our findings may be transferable to other study contexts. It is furthermore not clear, whether results are generalizable across cultures, as data for this study was collected in the US only. Future research with international (Haering et al., Reference Haering, Kooistra, Bourey, Chimed-Ochir, Doubková, Hoeboer and De Haan2024a; Young & Chan, Reference Young, Chan, Safdar and Kosakowska-Berezecka2015) as well as intersectional (Bryant-Davis, Reference Bryant-Davis2019; Crenshaw, Reference Crenshaw and Fineman2013; Seng, Lopez, Sperlich, Hamama, & Reed Meldrum, Reference Seng, Lopez, Sperlich, Hamama and Reed Meldrum2012) perspectives can help to further improve our understanding of health disparities in trauma-related health outcomes and beyond. Furthermore, the majority of participants included in this study were individuals who experienced an MVC. Future research should aim to specifically target large and diverse samples beyond MVC-exposed individuals to disentangle how sex and trauma characteristics interact in the case of physical and sexual assault as well as other trauma types. Another limitation includes the assessment of PTSD symptoms via self-report. Clinician-based interviews have been shown to increase diagnostic accuracy through better comprehension of symptoms or increased awareness (Kramer, Whiteman, Petri, Spitzer, & Weathers, Reference Kramer, Whiteman, Petri, Spitzer and Weathers2023). Finally, this study is limited to the analysis of sex differences. To fully disentangle why women are more vulnerable to PTSD development, future studies should take gender-related aspects as well as the interplay of sex and gender into account. These limitations notwithstanding, to the best of our knowledge, this study was the first to examine how the effects of a comprehensive set of risk factors, that have previously been suggested to show sex-differential associations with PTSD development, differs between men and women. The strengths of our study include its prospective design, controlled timing of assessments following trauma, as well as size and diversity of the study sample. Moreover, all statistical code for the analyses is shared open-access, and we encourage researchers to perform replication analyses to examine the generalizability of our findings in further study cohorts.
In summary, our study highlights the need for more sex- and gender-sensitive examinations of PTSD risk factors. Our results point toward the need to consider other pathways and mechanisms to explain women's higher PTSD risk. Our study also indicates mechanisms to which men might be particularly vulnerable. Continuing our efforts to disentangle sex differences in PTSD risk factors can help to more accurately understand how underlying mechanisms contribute toward an individual's PTSD risk and may lead to the development of refined targeted preventive interventions.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291724000941.
Acknowledgements
The authors would like to thank Aaron Williams, Xiaodi Yao, and the other members of the UNC Institute for Trauma Recovery for their efforts and aide in this research. We would also like to thank research staff at McLean Hospital, Emory University, Temple University, and Wayne State University for their efforts and aide. The investigators wish to thank the trauma survivors participating in the AURORA Study. Their time and effort during a challenging period of their lives make our efforts to improve recovery for future trauma survivors possible. This project was supported by NIMH under U01MH110925, the US Army MRMC, One Mind, and The Mayday Fund. The content is solely responsibility of the authors and does not necessarily represent the official views of any of the funders. Verily Life Sciences and Mindstrong Health provided some of the hardware and software used to perform study assessments. The Many Brains Project provided software for neurocognitive assessments. Data and/or research tools used in the preparation of this manuscript were obtained from the National Institute of Mental Health (NIMH) Data Archive (NDA). NDA is a collaborative informatics system created by the National Institutes of Health to provide a national resource to support and accelerate research in mental health. Dataset identifier(s): NIMH Data Archive Digital Object Identifier (DOI) 10.15154/qfkt-cg42. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIH or of the Submitters submitting original data to NDA.
Author contributions
Conceptualization: Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen, Stephanie Haering, Abigail Powers, Jennifer S. Stevens, Antonia Seligowski, Sarah Linnstaedt, Vasiliki Michopoulos; Methodology: Stacey L. House, Francesca L. Beaudoin, Xinming An, Jennifer S. Stevens, Thomas C. Neylan, Gari D. Clifford, Sarah D. Linnstaedt, Laura T. Germine, Scott L. Rauch, Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen; Formal analysis: Stephanie Haering, Abigail Powers, Jennifer S. Stevens; Investigation: Stacey L. House, Francesca L. Beaudoin, Xinming An, Jennifer S. Stevens, Thomas C. Neylan, Gari D. Clifford, Sarah D. Linnstaedt, Laura T. Germine, Scott L. Rauch, John P. Haran, Alan B. Storrow, Paul I. Musey Jr., Phyllis L. Hendry, Sophia Sheikh, Christopher W. Jones, Brittany E. Punches, Robert A. Swor, Nina T. Gentile, Lauren A. Hudak, Jose L. Pascual, Mark J. Seamon, Claire Pearson, David A. Peak, Roland C. Merchant, Robert M. Domeier, Niels K. Rathlev, Brian J. O'Neil, Leon D. Sanchez, Steven E. Bruce, Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen; Resources: Stacey L. House, Francesca L. Beaudoin, Xinming An, Jennifer S. Stevens, Thomas C. Neylan, Gari D. Clifford, Sarah D. Linnstaedt, Laura T. Germine, Scott L. Rauch, John P. Haran, Alan B. Storrow, Paul I. Musey Jr., Phyllis L. Hendry, Sophia Sheikh, Christopher W. Jones, Brittany E. Punches, Robert A. Swor, Nina T. Gentile, Lauren A. Hudak, Jose L. Pascual, Mark J. Seamon, Claire Pearson, David A. Peak, Roland C. Merchant, Robert M. Domeier, Niels K. Rathlev, Brian J. O'Neil, Leon D. Sanchez, Steven E. Bruce, Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen; Data Curation: Stacey L. House, Francesca L. Beaudoin, Xinming An, Jennifer S. Stevens, Thomas C. Neylan, Gari D. Clifford, Sarah D. Linnstaedt, Laura T. Germine, Scott L. Rauch, Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen; Writing – Original Draft: Stephanie Haering, Abigail Powers, Jennifer S. Stevens; Writing – Review & Editing: Antonia Seligowski, Sarah Linnstaedt, Vasiliki Michopoulos, Stacey L. House, Francesca L. Beaudoin, Xinming An, Thomas C. Neylan; Gari D. Clifford; Laura T. Germine; Scott L. Rauch; John P. Haran, Alan B. Storrow, Christopher Lewandowski, Paul I. Musey Jr., Phyllis L. Hendry, Sophia Sheikh, Christopher W. Jones, Brittany E. Punches, Robert A. Swor, Nina T. Gentile, Lauren A. Hudak, Jose L. Pascual, Mark J. Seamon, Claire Pearson, David A. Peak, Roland C. Merchant, Robert M. Domeier, Niels K. Rathlev, Brian J. O'Neil, Leon D. Sanchez, Steven E. Bruce; Steven E. Harte Samuel A. McLean, Ronald C. Kessler; Karestan C. Koenen; Jennifer S. Stevens, Abigail Powers Visualization: Stephanie Haering, Abigail Powers, Jennifer S. Stevens; Supervision: Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen, Stephanie Haering, Jennifer S. Stevens, Abigail Powers; Project Administration: Stacey L. House, Francesca L. Beaudoin, Jennifer S. Stevens, John P. Haran, Alan B. Storrow, Paul I. Musey Jr., Phyllis L. Hendry, Sophia Sheikh, Christopher W. Jones, Brittany E. Punches, Robert A. Swor, Nina T. Gentile, Lauren A. Hudak, Jose L. Pascual, Mark J. Seamon, Claire Pearson, David A. Peak, Roland C. Merchant, Robert M. Domeier, Niels K. Rathlev, Brian J. O'Neil, Leon D. Sanchez, Steven E. Bruce, Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen; Funding Acquisition: Samuel A. McLean, Ronald C. Kessler, Karestan C. Koenen. All authors revised the paper critically for important intellectual context and approved the final manuscript.
Funding statement
Advancing Understanding of Recovery After Trauma (AURORA) is supported by grant U01MH110925 from the NIMH, the US Army Medical Research and Material Command, the One Mind Foundation, and TheMayday Fund. Verily Life Sciences and Mindstrong Health provided some of the hardware and software used to perform study assessments. Stephanie Haering's time is supported by a PhD scholarship of ‘Stiftung der Deutschen Wirtschaft’. The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Competing interests
– Dr Neylan has received research support from NIH, VA, and Rainwater Charitable Foundation, and consulting income from Jazz Pharmaceuticals.
– In the last three years Dr Clifford has received research funding from the NSF, NIH and LifeBell AI, and unrestricted donations from AliveCor Inc, Amazon Research, the Center for Discovery, the Gates Foundation, Google, the Gordon and Betty Moore Foundation, MathWorks, Microsoft Research, Nextsense Inc, One Mind Foundation, the Rett Research Foundation, and Samsung Research. Dr Clifford has financial interest in AliveCor Inc and Nextsense Inc. He also is the CTO of MindChild Medical and CSO of LifeBell AI and has ownership in both companies. These relationships are unconnected to the current work.
– Dr Germine receives funding from the National Institute of Mental Health (R01 MH121617) and is on the board of the Many Brains Project. Dr Germine's family also has equity in Intelerad Medical Systems, Inc.
– Dr Rauch reports grants from NIH during the conduct of the study; personal fees from SOBP (Society of Biological Psychiatry) paid role as secretary, other from Oxford University Press royalties, other from APP (American Psychiatric Publishing Inc.) royalties, other from VA (Veterans Administration) per diem for oversight committee, and other from Community Psychiatry/Mindpath Health paid board service, including equity outside the submitted work; other from National Association of Behavioral Healthcare for paid Board service; other from Springer Publishing royalties; and Leadership roles on Board or Council for SOBP, ADAA (Anxiety and Depression Association of America), and NNDC (National Network of Depression Centers).
– Dr Jones has no competing interests related to this work, though he has been an investigator on studies funded by AstraZeneca, Vapotherm, Abbott, and Ophirex.
– Dr Harte has no competing interest related to this work, though in the last three years he has received research funding from Aptinyx and Arbor Medical Innovations, and consulting payments from Indiana University and Memorial Sloan Kettering Cancer Center.
– Dr McLean served as a consultant for Walter Reed Army Institute for Research and for Arbor Medical Innovations.
– In the past 3 years, Dr Kessler was a consultant for Cambridge Health Alliance, Canandaigua VA Medical Center, Holmusk, Partners Healthcare, Inc., RallyPoint Networks, Inc., and Sage Therapeutics. He has stock options in Cerebral Inc., Mirah, PYM, and Roga Sciences.
– Dr Koenen's research has been supported by the Robert Wood Johnson Foundation, the Kaiser Family Foundation, the Harvard Center on the Developing Child, Stanley Center for Psychiatric Research at the Broad Institute of MIT and Harvard, the National Institutes of Health, One Mind, the Anonymous Foundation, and Cohen Veterans Bioscience. She has been a paid consultant for Baker Hostetler, Discovery Vitality, and the Department of Justice. She has been a paid external reviewer for the Chan Zuckerberg Foundation, the University of Cape Town, and Capita Ireland. She has had paid speaking engagements in the last three years with the American Psychological Association, European Central Bank. Sigmund Freud University – Milan, Cambridge Health Alliance, and Coverys. She receives royalties from Guilford Press and Oxford University Press.
Open access
