Pre-trauma hair cortisol moderates adverse childhood experience and hypothalamic volume effects on stress symptoms after adult trauma

Abstract

Adverse childhood experience (ACE) has been associated with impairments of the hypothalamic-pituitary-adrenal (HPA) axis including reduced hair cortisol concentrations (HCC) and hypothalamus volumes in adults from previous trauma studies, but how both impairments contribute to post-traumatic stress symptoms (PTSS) after new trauma remains unclear. It is possible that a combination of reduced pre-trauma baseline levels of cortisol and early post-trauma hypothalamic dysregulation contributes to inadequate stress reactions to acute trauma. To test this hypothesis, 73 adult trauma survivors completed a hair sample collection immediately after trauma for pre-trauma HCC measures. The PTSD Checklist (PCL), Childhood Trauma Questionnaire (CTQ), and an MRI scan for hypothalamic and subunit volume measures were obtained within 2 weeks post-trauma. PTSS was further assessed using PCL 3 months later. The results indicate that interactions of pre-trauma HCC and early post-trauma volumes of the hypothalamus or its posterior subunits significantly affect PTSS severity at 2 weeks post-trauma (bilateral t = 2.502 ~ 3.920, p = 0.016 ~ 0.001) and re-experiencing symptom severity at 3 months post-trauma (left side only, t = 2.196 ~ 2.529, p = 0.017 ~ 0.037). CTQ scores interact with pre-trauma HCC and with both HCC and left hypothalamus volume to significantly affect re-experiencing symptom severity at 3 months post-trauma (p = 0.017 ~ 0.031). These findings suggest that combinations of pre- and early post-trauma conditions of the HPA axis affect acute PTSS and moderate an association between ACE and re-experiencing symptoms in subsequent months after trauma.

Introduction

Millions of Americans experience a traumatic event in their lifetime that leads to post-traumatic stress symptoms (PTSS) quickly after trauma. Some trauma survivors do not fully recover and subsequently develop post-traumatic stress disorder (PTSD), which significantly impacts public health [1–3]. Identifying risk factors and related mechanisms for the emergence of PTSS after acute trauma is needed for understanding and preventing PTSD development. Given the unpredictable nature of trauma, identifying pre- and/or peri-trauma factors that may contribute to PTSS is difficult.

Adverse childhood experiences (ACEs), which can involve physical or emotional abuse, sexual abuse, or physical or emotional neglect, occur in 7–60% of children [4]. ACEs have been linked to mental health problems later in life [5–7]. For example, 17–23% of young adults who experienced at least one type of ACEs were diagnosed with PTSD after adult trauma as compared to 10% of those without an ACE history [8]. ACEs can affect the development of brain structures, which in turn, may increase the risk of severe PTSS and developing PTSD after adult trauma [9, 10]. ACE-related stress impacts the hypothalamic-pituitary-adrenal (HPA) axis, a key component of stress regulation. The HPA axis contributes to stress responses and emotional and physiological homeostasis, in part, by regulating cortisol levels [11–16]. ACEs are associated with low cortisol levels in adults [17–19], but the findings remain inconsistent [20]. The hypothalamus, a diencephalon region between the thalamus and the midbrain that is composed of several subunits, plays a critical role in regulating cortisol release [15, 21, 22]. The hypothalamic posterior subunit includes the mammillary bodies (MBs) and tuberomamillary nucleus (TMN), both of which are particularly important in stress responses and memory of aversive life experiences [23–25]. Other subunits of the hypothalamus also play roles in the release of hormones to regulate the HPA axis, metabolism, reproduction, circadian rhythm, and other essential functions [26]. Disruption of hypothalamic structure and function leads to cortisol dysregulation and deficient stress responses [26, 27]. Our recent work suggests that ACEs are associated with bilaterally smaller hypothalamus volumes within 2 weeks after an adult trauma, and early post-trauma smaller right hypothalamus volume mediated an association between ACEs and PTSD symptom severity one year after trauma [28]. This raises a possibility that ACEs may impair hypothalamus development resulting in small volume and dysregulation of cortisol that, in turn, contribute to the emergence of PTSS and subsequent PTSD development after adult trauma. Given this possibility, ACEs may impact the HPA axis prior to an adult trauma. However, we did not find a significant relationship between hypothalamus volumes and PTSS at 2 weeks post trauma [28], suggesting other mechanisms may influence the association of hypothalamus volumes and PTSS after acute trauma.

A recent study retrospectively assessed pre-trauma cortisol by measuring hair cortisol concentrations (HCC) collected immediately after adult trauma and found that lower HCC in the preceding three months predicted greater PTSS severity in the days after trauma [20]. Cortisol absorbed from blood into hair follicles becomes incorporated into hair shafts [29]. Growing hair takes approximately 7–10 days to reach the scalp surface [30], and on average, hair grows from the scalp one cm per month [31]. Thus, HCC from scalp hair samples taken immediately after trauma retrospectively reflect sustained cortisol levels over a pre-trauma period that is determined by the hair length from scalp. This provides an opportunity to assess how sustained HCC levels in 1–3 months preceding trauma interact with other factors to influence acute PTSS severity. HCC has been reported to associate with accumulated stress [19, 32]. HCC has been increasingly used to study cortisol levels because cortisol that is accumulated in hair is less sensitive to situational and circadian fluctuations [33] and consequently has advantages over single-time point cortisol measure from blood, saliva, or urine [34]. Correlations between HCC, PTSS severity in chronic PTSD patients, and other mental health symptoms have been reported [34–39]. A recent study reports that ACEs are associated with low HCC in adults [19], but the pilot study reports that pre-trauma HCC is not significantly associated with ACEs [20]. Therefore, pre-trauma baseline HCC may be associated with emergence of PTSS after acute trauma, but its contribution to the association between ACEs and PTSS needs further investigation.

Taken together, these findings raise the possibility that ACEs may impair hypothalamic development, resulting in smaller hypothalamus volumes and dysregulated cortisol. These alterations, combined with lower pre-trauma baseline HCC associated with ACEs and other stressful experiences, may contribute to inadequate stress responses to acute trauma in adulthood. Impairments in the MBs and TMN within the hypothalamic posterior subunit may influence stress-related memory and contribute to PTSS symptoms. To test this possibility, we collected hair samples immediately after the traumatic event to assess sustained HCC levels in the month(s) that preceded the trauma. Pre-trauma ACEs and PTSS symptoms at 2 weeks and 3 months post-trauma were assessed with standard questionnaires. A structural MRI scan was collected for FreeSurfer automated measures of hypothalamic total and subunit volumes within 2 weeks post-trauma. The goal was to assess whether/how ACEs, pre-trauma HCC, and early post-trauma hypothalamic volumes contribute to PTSS severity over the initial days to months after acute trauma in adulthood.

Materials and methods

Ethics approval and consent to participate

All methods were performed in accordance with the relevant guidelines and regulations. Approval has been obtained from The University of Toledo Institutional Review Board (Protocol 201575-UT). Informed consent was obtained from all participants.

Participants and recruitment

This study is an add-on component of an ongoing acute trauma MRI study in our lab [28]. Adult trauma survivors aged 18–60 years old were recruited from Emergency Departments (EDs) of local hospitals within 48 h after a life-threatening traumatic event. Traumatic events included motor vehicle collision (MVC), physical assault, sexual assault, or other trauma. Excluded were survivors who: 1) were severely injured or had contraindications for MRI scanning, e.g. pregnancy, ferrous-containing metals within the body, claustrophobia, 2) had been diagnosed with severe psychiatric, physical, or neurological problems or had a history of moderate or severe traumatic brain injury, or 3) were under the influence of alcohol or substances at the time of trauma. The PTSD Checklist for DSM-5 (PCL) survey was used to identify trauma survivors with high stress levels (PCL total score > 28) to ensure that subjects at high risk for developing PTSD were enrolled. Participants of the parent study were consented to this study if they could provide hair samples. Additional exclusion of subjects involved hair conditions, i.e., dyed or chemically treated hair, or scalp hair length less than 1 cm.

Using G*Power [40], we conducted a power analysis to determine whether the study was adequately powered for planned moderation regression analyses. Assuming a conventional a = 0.05 and desired power of 0.80, we targeted an effect size f ² = 0.35 for key predictors and interaction terms. In the context of multiple regressions with a potential predictor intercorrelation up to r ≈ 0.5, this analysis indicated a minimal sample of 36 subjects was required. A large number of subjects were recruited to ensure quality control in MRI imaging and HCC assays.

Hair sample collection and cortisol analysis

Trained research staff cut 3 cm scalp-end hair strands, or 1 cm hairs when 3 cm hairs were not feasible, as close to the scalp as possible from the posterior vertex of the head. Lengths of hair samples were recorded and adjusted as a covariant in statistical analysis. Approximately 20 mg of hair was collected, wrapped in aluminum foil, and stored in opaque containers at room temperature.

HCC was analyzed by applying a published procedure using an LC-MS/MS system consisting of a Shimadzu Nexera XR ultra-high performance liquid chromatograph (Shimadzu Scientific Instruments, Columbia, MD, USA) and AB Sciex QTRAP 6500 mass spectrometers equipped with an Electrospray Ionization (ESI) source (AB Sciex, Framingham, MA, USA) [41]. Hair segments were washed twice in 2.5 mL of isopropanol for three minutes at room temperature, then air-dried under a fume hood for a minimum of 12 h. To extract steroid, 7.5 mg of finely minced hair was incubated with 20 μL of internal standard and 1.8 mL of methanol for 18 h at room temperature, following which the extract was evaporated using a 50 °C nitrogen stream. The dried residue was resuspended in 225 μL of methanol/water (v/v = 50/50) and 100 μL of the solution was injected into the LC-MS/MS system for analysis [41]. Intra- and inter-assay coefficients of variation for HCC ranged between 3.7–8.8% [29].

ACE history evaluation

Consented subjects completed the Childhood Trauma Questionnaire (CTQ) to retrospectively evaluate ACE history. A 28-item CTQ survey was used to assess five types of childhood maltreatment that occurred up to the age of 18, including emotional and physical abuse, sexual abuse, and emotional and physical neglect [42]. Each item was evaluated on a 5-point Likert-type scale according to the frequency of occurrence (1 = never true, 5 = very often true). The total CTQ score was used in this study.

PTSS severity assessments

The PCL survey was used to assess stress at enrollment, and at 2 weeks and 3 months post trauma. PCL is a self-report 20-item survey that evaluates the trauma-related symptoms that led to the patient’s ED visit and recruitment. Symptoms included re-experiencing (PCL-B), avoidance (PCL-C), negative alterations in cognition and mood (PCL-D), and hyperarousal (PCL-E) symptom clusters. Each item was graded on a 5-point Likert scale of severity (0 = Not at all, 4 = Extremely). This yielded a total score and four subscores. The PCL has strong internal consistency (α = 0.94) and test-retest reliability (r = 0.82) [43].

Structural MRI acquisition and processing

Subjects were scanned within two weeks after trauma using a 3 T General Electric Signa HDx MRI scanner (GE Healthcare, Chicago, IL). A high-resolution T1-weighted sMRI was obtained using a previously validated high-resolution 3D FSPGR structural MRI protocol (TR = 7.836 ms, TE = 2.976 ms, FA = 9°, NEX = 1, field of view = 256 × 256 mm, matrix = 256 × 256, slice thickness = 1 mm, voxel dimensions = 1 × 1 × 1 mm³, 164 contiguous axial slices) [28]. A trained MRI team member checked sMRI image quality, and subjects with blurred sMRI images caused by head motion or other issues were excluded from analyses. An experienced radiologist screened sMRI images for clinical abnormalities.

sMRI images were processed using FreeSurfer programs (Version 7.2) (https://surfer.nmr.mgh.harvard.edu). The hypothalamus and its subunits, which include anterior-inferior, anterior-superior, tubular-inferior, tubular-superior, and posterior subunits on each side, were segmented and their volumes were measured using the fully automated segmentation and deep machine learning analysis in the FreeSurfer package. This package has high inter-rater reliability and intra-rater precision [22]. An inspector blinded to psychological assessments and other measures visually checked image segmentation. Intracranial volume (ICV) was measured.

Statistical analysis

Based on prior observations [44], extremely high outliers of HCC measures (3 standard deviations (SD) above the mean) were considered unreliable and excluded from analyses. The distribution of absolute HCC measure exhibited non-normality (Skewness = 2.7 and Kurtosis = 8.3, Supplementary Table 1). A one-sample Kolmogorov–Smirnov test indicated that log-transformed HCC data had a normal distribution (log HCC, p = 0.20). Consequently, log-transformed HCC data were used in the analyses.

Associations between pre-trauma CTQ score and HCC, trauma type, and post-trauma hypothalamic volumes and PCL scores were tested using multiple linear regression analysis. Age, sex, ICV, and hair sample length were used as covariates. Pre-trauma HCC moderating effects on the associations between other pre-/post- trauma factors and PTSS severity were of interest. Moderation analyses were performed to determine whether pre-trauma HCC moderated the associations between CTQ scores or hypothalamic volumes at 2 weeks post-trauma and PCL scores assessed at 2 weeks and 3 months post-trauma. We further explored whether pre-trauma HCC moderated associations between hypothalamic posterior subunit volumes at 2 weeks post-trauma and PCL scores at 2 weeks and 3 months post-trauma. In addition, complex moderation models were applied to test potential moderating effects of dual moderators, HCC and hypothalamus volumes, on associations between CTQ scores and PCL scores. Johnson-Neyman approaches were used to identify specific HCC levels at which the conditional effects of CTQ scores or hypothalamic volumes on PCL scores became statistically significant [45]. Age, sex, ICV, trauma type, and hair segment length were included as covariates in analyses.

Moderation results were illustrated with graphs to indicate conditional effects of CTQ scores or hypothalamus volume on PCL scores across a range of HCC values with upper and lower level 95% confidence limits (ULCL, LLCL). For dual moderator moderation results, we used a pick-a-point technique at the mean and ± 1 SD of both HCC and hypothalamus volume to visually illustrate conditional effects of CTQ on PCL scores [46].

Statistical analyses were conducted using SPSS 29 (IBM Corp., Armonk, N.Y., USA) and PROCESS v4.2 for SPSS. Data are reported as mean ± SD. FDR corrected p < 0.05 was considered significant in regression analyses and p < 0.05 was considered significant in the hypothesis-based moderation analyses.

Results

Sample

Seventy-three acute trauma survivors were recruited in this study. Sample demographic and summary information are indicated in Table 1. Ten subjects lacked HCC data due to unreliable measures, and twelve subjects lacked hypothalamic volume data due to subject withdrawal or poor image quality. Forty-four subjects provided PCL scores at three months post-trauma. Of these, twenty-two were diagnosed with PTSD using the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5) interview. CTQ scores were not significantly associated with pre-trauma HCC. PCL scores at 3 months, but not at 2 weeks, post trauma were significantly positively associated with CTQ scores (β = 0.333, standard error (SE) = 0.107, t = 3.123, FDR corrected p = 0.04). PCL scores at both time points were not significantly associated with HCC, left or right hypothalamus volumes, or trauma type in linear regression analysis. PCL subscores and hypothalamic subunit volumes were reported in Supplementary Table 1.

Table 1.

Demographic and summary information.

	mean ± SD (min. - max.) or N (%)
Age (year)	31 ± 11 (18 -55)
Sex (F/M)	57 (78%)/16 (22%)
Race (N)
White	44 (60.3%)
African American	24 (32.9%)
Native American	1 (1.4%)
Asian American	1 (1.4%)
Other	3 (4%)
Trauma type
MVC	45 (62.5%)
Physical Assault	18 (25%)
Sexual Assault	8 (11.1%)
Other	2 (1.4%)
Pre-trauma measures
CTQ score	57 ± 25 (25 - 121)
HCC (pg/mg)	9.05 ± 11.7 (0.31 - 57.78)
log HCC	0.69 ± 0.50 (−0.51 - 1.76)
Post-trauma measures
PCL score
at enrollment	47 ± 15 (31 - 75)
at 2 weeks post-trauma	50 ± 14 (0 - 76)
at 3 months post-trauma	38 ± 15 (3 - 72)
re-experiencing subscore	10 ± 4 (0 - 20)
Hypothalamus volume (mm3)
Left total volume	424 ± 41 (341 - 532)
Left posterior subunit	118 ± 15 (90 - 152)
Right total volume	409 ± 39 (329 - 496)
Right posterior subunit	117 ± 16 (77 - 155)
ICV (mm3)	1434621 ± 158266 (973810 - 1811060)

MVC motor vehicle collision, CTQ childhood trauma questionnaire, HCC hair cortisol concentration, PCL PTSD Checklist, ICV intracranial volume.

PTSS in 2 weeks and 3 months after trauma was influenced by pre-trauma HCC and early post-trauma hypothalamic volumes

Interactions between pre-trauma HCC and left or right hypothalamic volumes at 2 weeks post-trauma on PCL scores at 2 weeks post-trauma were significant (left: β = 0.338, SE = 0.086, t = 3.920, p < 0.001; right: β = 0.255, SE = 0.083, t = 3.080, p = 0.004), suggesting that pre-trauma HCC moderated an association between left or right hypothalamic volumes and PCL scores at 2 weeks post-trauma. Johnson-Neyman analysis further revealed that left hypothalamus volumes and PCL scores at 2 weeks post-trauma were significantly positively correlated when pre-trauma log HCC was > 0.697 (HCC > 4.98 pg/mg), as seen in 45% of survivors; whereas, a significant negative correlation occurred between these measures when log HCC was < −0.087 (HCC < 0.82 pg/mg), as seen in 8% of survivors (Fig. 1a). Similarly, right hypothalamus volumes and PCL scores were significantly positively correlated when pre-trauma log HCC was > 1.114 (HCC > 13.0 pg/mg), as seen in 20% of participants, but were significantly negatively correlated when pre-trauma log HCC was < −0.161 (HCC < 0.69 pg/mg), as seen in 8% of participants (Fig. 1b). The main effects of pre-trauma HCC on PCL scores were significant in either the left or right hypothalamus model, but neither left nor right hypothalamus volumes were associated with PCL scores (Supplementary Table 2). Taken together, the above results suggest that associations between left and right hypothalamus volumes and PCL scores at 2 weeks post trauma varied from significantly positive, insignificant, and significantly negative depending on respective high to low levels of pre-trauma HCC.

Fig. 1. Simple moderation analyses showed the conditional effects of hypothalamus volumes on PTSS severity at a series of HCC values with upper and lower level 95% confidence limits (ULCL, LLCL).

Left (a) and right (b) hypothalamus volumes effects on PCL scores at 2 weeks post-trauma. (c) Left hypothalamus volumes effects on PCL re-experiencing subscores at 3 months post-trauma.

In contrast to the above findings for PCL scores at post-trauma 2 weeks, interactions between pre-trauma HCC and left or right hypothalamus volumes at 2 weeks post-trauma on subsequent PCL scores at 3 months post trauma were not significant. However, the interaction between pre-trauma HCC and left hypothalamic volumes at 2 weeks post-trauma was significant on PCL re-experiencing subscores at 3 months after trauma (β = 0.106, SE = 0.042, t = 2.529, p = 0.017), suggesting that pre-trauma HCC moderated an association between early left hypothalamic volumes and subsequent PCL re-experiencing subscores at 3 months post-trauma. Johnson-Neyman analysis revealed that when pre-trauma log HCC was > 1.179 (HCC > 15.10 pg/mg), as seen in 16% of subjects, left hypothalamus volumes at 2 weeks post-trauma was significantly positively associated with PCL re-experiencing subscores at 3 months post-trauma, whereas, when pre-trauma log HCC was < −0.311 (HCC < 0.49 pg/mg) these measures were significantly negatively correlated, as seen in 5% of subjects (Fig. 1c). The main effect of pre-trauma HCC, but not left hypothalamus volume, on PCL re-experiencing subscores at 3 months post trauma was significant in this moderation model (Supplementary Table 2). Interactions between HCC and hypothalamus volumes on other PCL subscores at 3 months post trauma were not significant.

Hypothalamic posterior subunit volumes

We further tested the above moderation analysis on the volumes of the hypothalamic posterior subunits on both sides. Interactions between pre-trauma HCC and either left or right hypothalamic posterior subunit volumes at 2 weeks post-trauma on PCL scores at 2 weeks post trauma were significant (left: β = 0.750, SE = 0.243, t = 3.088, p = 0.004; right: β = 0.548, SE = 0.219, t = 2.502, p = 0.016), suggesting that pre-trauma HCC moderated an association between either left or right hypothalamic posterior subunit volumes and PCL scores at 2 weeks post-trauma. The left hypothalamic posterior subunit volume was significantly positively associated with PCL scores at 2 weeks post-trauma when pre-trauma log HCC was > 0.619 (HCC > 4.16 pg/mg), as seen in 54.9% of subjects (Fig. 2a). Similarly, the right hypothalamic posterior subunit volume was positively associated with PCL scores at 2 weeks post-trauma when pre-trauma log HCC was > 0.966 (HCC > 9.25 pg/mg), as seen in 23.5% of subjects (Fig. 2b). The main effect of HCC was significant for PCL scores two weeks after trauma. Neither left nor right posterior subunit volumes were associated with PCL scores at 2 weeks post-trauma (Supplementary Table 2).

Fig. 2. Simple moderation analyses showed the conditional effects of hypothalamic posterior subunit volumes on PTSS severity at a series of HCC values with upper and lower level 95% confidence limits (ULCL, LLCL).

Left (a) and right (b) hypothalamic posterior subunit volumes effects on PCL scores at 2 weeks post-trauma. (c) Left hypothalamic posterior subunit volumes effects on PCL re-experiencing subscores at 3 months post-trauma.

Interactions between pre-trauma HCC and left or right hypothalamic posterior subunit volumes at 2 weeks post-trauma on the PCL scores at 3 months after trauma were not significant. However, interaction of pre-trauma HCC and left hypothalamic posterior subunit volumes on subsequent PCL re-experiencing subscores at 3 months post-trauma was significant (β = 0.263, SE = 0.120, t = 2.196, p = 0.037), suggesting that pre-trauma HCC moderated an association between early post-trauma left hypothalamic posterior subunit volumes and subsequent PCL re-experiencing subscores at 3 months post-trauma. When pre-trauma log HCC was < 0.096 (HCC < 1.25 pg/mg) as seen in 13.5% subjects, left hypothalamic posterior subunit volumes at 2 weeks post-trauma were significantly negatively associated with PCL re-experiencing subscores at 3 months after trauma (Fig. 2c). The main effects of either HCC or left hypothalamic posterior subunit volumes on PCL re-experiencing subscores at 3 months post trauma were significant (Supplementary Table 2).

Associations between PTSS in the months after trauma and ACE were influenced by pre-trauma HCC and hypothalamic volumes at 2 weeks post-trauma

Interactions between pre-trauma HCC and CTQ scores on PCL scores at either 2 weeks or 3 months post-trauma were not significant. However, the PCL re-experiencing subscores at 3 months post-trauma was significantly affected by this interaction (β = −0.166, SE = 0.065, t = −2.543, p = 0.017), suggesting that pre-trauma HCC moderated an association between CTQ scores and PCL re-experiencing subscores at 3 months post-trauma. Johnson-Neyman analysis indicated that when pre-trauma log HCC was < 0.660 (HCC < 4.57 pg/mg), as seen in 46% of subjects, CTQ scores were significantly positively associated with the PCL re-experiencing subscores at 3 months post-trauma (Fig. 3). The main effects of pre-trauma CTQ and HCC were significant in this model (Supplementary Table 2).

Fig. 3.

Simple moderation analysis showed the conditional effect of CTQ scores on PCL re-experiencing subscores at 3 months post-trauma at series of pre-trauma HCC values with upper and lower level 95% confidence limits (ULCL, LLCL).

Next, we included interactions between CTQ scores and pre-trauma HCC, as well as between CTQ scores and left or right hypothalamus volumes, in a complex moderation model to examine whether the dual moderators, pre-trauma HCC and early post-trauma hypothalamus volumes, affected the association between CTQ scores and PCL re-experiencing subscores at 3 months post-trauma. The joint moderating effect was significant in the left (R² = 0.164, F(2, 26) = 3.976, p = 0.031), but not the right (R² = 0.136, F (2, 26) = 2.972, p = 0.069), hypothalamus model. For the left hypothalamus model, the interaction between pre-trauma CTQ and HCC was significant (β = −0.172, SE = 0.065, t = −2.664, p = 0.013), whereas the interaction between CTQ and the left hypothalamus volume was not significant (t = 1.014, p = 0.320). Johnson–Neyman analysis indicated that, regardless of the left hypothalamus volume, a positive association between CTQ scores and the PCL re-experiencing subscores at 3 months post-trauma was significant when pre-trauma HCC was low (i.e., 1 SD below the mean). In contrast, this association was not significant when pre-trauma HCC was high (i.e., 1 SD above the mean). When pre-trauma HCC was around the mean level, the correlation between CTQ and PCL scores was affected by the left hypothalamus volume. A significant positive correlation was seen when the left hypothalamus volume was at or above the mean, whereas CTQ and PCL scores were not associated when the left hypothalamus volume was small (i.e., 1 SD below the mean) (Fig. 4). Further analysis for a three-way interaction involving CTQ scores, pre-trauma HCC, and early post-trauma left hypothalamus volume on subsequent PCL re-experiencing subscores at 3 months post-trauma was not significant.

Fig. 4. Dual moderators’ moderation was visually illustrated using the pick-a-point technique with the mean and mean± 1 SD of moderator.

When pre-trauma HCC was lower, CTQ scores were positively associated with greater PCL re-experiencing subscores at 3 months post-trauma regardless of hypothalamus volumes, and when pre-trauma HCC was at the mean level, this association was significant only when the left hypothalamus volumes were at mean volume or larger.

Additionally, interactions between pre-trauma HCC and CTQ scores were not significant for other PCL subscores at either post-trauma assessment. The dual moderation models for PCL scores and for all other PCL subscores at both time points likewise yielded no significant moderating results.

Discussion

The present study examined influences of pre- and early post- trauma HPA axis conditions on PTSS at 2 weeks and 3 months after adulthood trauma, and contributions of ACE history to these influences. Interaction effects suggest that associations between bilateral hypothalamus volumes and PTSS severity at 2 weeks post trauma vary from negative to positive depending on levels of HCC within 3 months pre-trauma. Similar effects of pre-trauma HCC levels were seen in an association between left hypothalamus volumes at 2 weeks post trauma and severity of PTSS re-experiencing symptoms 3 months after trauma. In addition, moderating effects of pre-trauma HCC were also seen on associations between hypothalamic posterior subunit volumes and PTSS severity at post-trauma 2 weeks and 3 months. Further findings indicated that ACE and PTSS re-experiencing symptom severity at 3 months post trauma were positively associated when pre-trauma HCC was at and below the mean level, but not at high levels. This HCC effect was not influenced when left hypothalamus volumes are at and above the mean level; however, when left hypothalamus volumes were small, ACE and PTSS re-experiencing severity at 3 months post-trauma were significantly associated only when HCC is also low. Overall, the present findings suggest that interactions between multiple factors contribute to PTSS in the initial days to months after adult trauma, including pre-trauma ACE and HCC, and early post-trauma hypothalamic structural conditions.

Pre-trauma HCC level and early post-trauma hypothalamic volumes affect PTSS severity after acute trauma

The observed interactions between pre-trauma HCC and early post-trauma hypothalamic volumes on PTSS severity over weeks to months after trauma raise the possibility that conditions of the HPA axis in terms of HCC levels and hypothalamic structure around the time of trauma may distinguish trauma survivors who are differently vulnerable to PTSS. For example, the HPA axis controls cortisol release due to stress [11, 12, 15], and subjects with high pre-trauma HCC may have been experiencing higher stress in the months before trauma. If the subjects also have larger hypothalamus volumes, they may have been better able to adjust HPA axis response, such as release of more cortisol, to stress from the new trauma. However, it is possible that potentially very high levels of cortisol in these subjects with high pre-trauma HCC and a large hypothalamus may lead to excessive responses to acute trauma and induce severe PTSS [13, 15, 16]. This over-production of cortisol could eventually disrupt hypothalamic regulation and lead to a cortisol deficit and hypocortisolism [15, 34].

We found interactions involving the posterior subunits of the left and right hypothalamus. The posterior subunit includes the mammillary bodies and tuberomammillary nucleus [22]. Mammillary bodies are elements of the Papez circuit that contribute to emotional memory formation, consolation, and retrieval. Mammillary bodies transmit information to the cingulate cortex to assist in emotion perception and memory retrieval [23, 24]. The tuberomammillary nucleus contains histaminergic neurons that contribute to learning, memory, and cortisol release [25]. Given the role of cortisol modulation on fear learning and memory [47, 48], subjects with high pre-trauma HCC and a large hypothalamus, especially the posterior subunit, may have altered fear learning and emotional memory formation and retrieval. This could lead to severe PTSS in the initial weeks post-trauma. This may also have contributed to the positive association between early hypothalamus volumes and PTSS re-experiencing symptoms at 3 months post trauma.

Other subjects had low pre-trauma HCC. This raises the possibility that these subjects either experienced a low stress load or impaired cortisol regulation before the trauma. If subjects had smaller bilateral hypothalamus volumes at 2 weeks post-trauma, they appeared to have more severe PTSS, possibly due to altered HPA axis function. Chronic stress can disrupt hypothalamic structure and cause hypocortisolism that would be consistent with these findings [12, 15, 21]. Subjects who had low pre-trauma HCC and small bilateral hypothalamus volumes in the initial weeks after trauma may have experienced impaired responses to the new trauma that could be consistent with the negative associations between post-trauma hypothalamus volumes and acute PTSS. Interestingly, subjects with low HCC also exhibited negative correlations between early post-trauma left hypothalamus and its posterior subunit volumes and PTSS re-experiencing symptoms 3 months after trauma. This raises a possibility that structural conditions in the left hypothalamus, especially its posterior subunit, during initial post-trauma weeks may have contributed to development of subsequent PTSS re-experiencing symptoms months later.

Previous findings on the contributions of cortisol to acute stress responses to trauma are inconsistent and based on group data [35, 36, 38]. The current findings raise the interesting possibility that these inconsistencies may, to some degree, reflect group data that do not resolve and distinguish subjects with heterogeneous pre-trauma HCC and early post-trauma hypothalamic structural conditions. In this respect, subjects with a large hypothalamus who had high pre-trauma cortisol levels may have had an over-production of cortisol in response to the new trauma that increased early PTSS. On the other hand, subjects with low pre-trauma HCC and a smaller hypothalamus may have impaired HPA responses to the new trauma, contributing to PTSS and later re-experiencing symptoms. Thus, PTSS emerged in part due to specific factors that operated in the months before and initial weeks after the new trauma.

Pre-trauma HCC level and early post-trauma hypothalamus volumes contributed to associations between ACE and PTSS re-experiencing symptoms at 3 months after adult trauma

The present findings indicated there was a positive correlation between the severity of ACE history and PTSS at 3 months after adult trauma. This is consistent with previous works which suggest ACE as a risk factor for severe PTSS and subsequent PTSD [2, 5, 49]. Animal studies suggest early-life stress can cause hypothalamic cell loss and genetic changes that lead to stress control disruption [50, 51]. Low cortisol levels have been reported in human adults with an ACE history [17, 19, 52]. The present findings indicate there was a positive correlation between ACE severity and PTSS re-experiencing symptoms at 3 months post-trauma when pre-trauma HCC was low, but not when HCC was high. This suggests that low pre-trauma cortisol levels contributed to this association. In addition, this association occurred when both pre-trauma HCC and left hypothalamus volumes were low. This raises the possibility that hypothalamic structural deficit and related cortisol dysregulation contributed to this association. However, a weaker association was shown when pre-trauma HCC was at mean levels and post-trauma hypothalamus volumes were at or above mean volume levels. This suggests that involvement of different factors contribute to ACE effects on post-trauma responses. For example, ACE can impair development of brain regions involved in attention, memory, emotion, and sensory processing, which, in turn, lead to cognitive dysfunctions that may contribute to PTSS after adult trauma [9, 53–55].

Limitations

This study has several limitations. (1) No trauma-free control subjects were studied. (2) The sample is small. (3) Adult life experiences and stress during the pre-trauma HCC assessment period are not addressed. (4) Ages at which ACE occurred, and ACE duration are not assessed. (5) The study does not address mental disorders that can be comorbid factors for PTSS.

Conclusion

This study used survivors of acute adult trauma to longitudinally assess whether/how pre- and early post-trauma factors contribute to PTSS. PTSS severity over the early weeks to months after trauma was affected by interactions that involved pre-trauma ACE, pre-trauma HCC, and early post-trauma hypothalamic and hypothalamic posterior subunit volumes. The results suggest that PTSS in the initial weeks and months after adult trauma emerges from influences of both pre-trauma and early post-trauma conditions of the HPA axis. Furthermore, ACE may impair the HPA axis resulting in reduced pre-trauma HCC and early post-trauma hypothalamus volume, which increases risks for developing re-experiencing symptoms after trauma in adulthood. The findings support the view that PTSS is shaped by complex interactions among multiple factors present before and after trauma.

Author contributions

HX conceived and designed the study; conducted recruitment, hair sample collection, and statistical analyses; and drafted manuscript and revision. LD interpreted the results and drafted and revised the manuscript. RMH contributed to results interpretation and manuscript drafting. CHS conducted data collection and analyses. WG performed hair cortisol measurements. JTW contributed to study design and manuscript and revision editing. REM contributed to the conception and funding support. XW conceived and designed the study, acquired funding, supervised data collection and analyses, and contributed to manuscript revision.

Fundings

This work was supported by NIMH R01MH110483 and a departmental research incentive fund.

Data availability

The data supported by NIMH R01MH110483 are available through the NIMH Data Archive (NDA 2541). The data supported by the departmental research incentive fund are not publicly available; however, they may be obtained from the corresponding author upon reasonable request.

Competing interests

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest

Supplementary information

The online version contains supplementary material available at 10.1038/s41398-026-03901-1.