Estradiol Levels are Differentially Associated with Pulse Wave Velocity in Trauma-exposed Premenopausal Women with and without PTSD

Abstract

Arterial stiffness is a well-known risk factor for cardiovascular disease. Although estradiol (E2) is known to be cardioprotective, the available data point to a growing cardiovascular disease risk in women before menopause due to post-traumatic stress disorder (PTSD). The present study aimed to investigate the effects of E2 on arterial compliance in trauma-exposed premenopausal women, with and without a clinical diagnosis PTSD. We hypothesized that E2 will be differentially associated with pulse wave velocity (PWV) in women with PTSD (PTSD+, n=45) and without PTSD (PTSD−, n=47). Estradiol and PWV were measured during two separate study visits. Serum E2 levels were measured via the quantitative sandwich enzyme-linked immunoassay technique (ELISA) and log-transformed due to non-normal distribution. Carotid to femoral applanation tonometry was used to measure PWV. Our analyses revealed an overall weak and non-significant correlation between E2 and PWV (r=−0.119, p=0.350). However, when examining each group, we found a negative association between E2 and PWV in PTSD− (r=−0.466, p=0.004). In contrast, we found an unexpected positive association between E2 levels and PWV in PTSD+ (r=0.360, p=0.037). Furthermore, a multiple linear regression revealed that E2 was predictive of PWV in PTSD− only, even after accounting for phase of the menstrual cycle, age, BMI, diastolic blood pressure and PTSD symptom severity (R²= 0.670, p=0.005). Interestingly, we also found lower levels of E2 in PTSD+ compared to PTSD− (1.4±0.4 vs 1.6±0.4 pg/mL, p=0.022). These findings suggest that PTSD may inhibit the protective effects of E2 on arterial compliance in women prior to menopause.

Graphical Abstract

NEW AND NOTEWORTHY

In trauma-exposed premenopausal women, we found that serum estradiol E2 was a predictor of PWV only in the absence of a PTSD diagnosis, even after accounting for phase of the menstrual cycle, age, BMI, diastolic blood pressure and PTSD symptom severity. Moreover, E2 levels were lower in women with PTSD compared to women without PTSD. We collected E2 and PWV during two separate visits and controlled for the phase of the menstrual cycle in our analyses.

INTRODUCTION

Cardiovascular diseases (CVD), such as coronary artery disease, are the leading cause of death in women in the United States, causing one in three deaths each year (1). One of the emerging risk factors for CVD is posttraumatic stress disorder (PTSD) (2, 3). A recent meta-analysis of nine prospective studies with 151,144 participants revealed that individuals with PTSD had a 61% higher rate of incident coronary heart disease (4). PTSD is a debilitating mental disorder that often develops after exposure to traumatic events. The Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for PTSD include the presence of specific symptoms for a minimum of one month: intrusion, avoidance, negative mood, and cognitive alterations (5). According to the National Center for PTSD, more than half of all women are exposed to at least one traumatic event in their lifetime, starting at a young age (i.e., prior to adulthood). Thus, age was a significant moderator of PTSD risk, with potentially traumatic events occurring earlier in life (6-8). Moreover, women have a two to three times higher risk of developing PTSD compared to men (6). Although women are typically thought to be protected from CVD before menopause, available data indicate a growing risk of CVD in this population by virtue of PTSD.

Low estradiol (E2) levels are associated with PTSD vulnerability and severity in women(9). Estradiol, a sex hormone that is primarily involved in reproductive health, has been implicated in cognitive and emotional processes (10). Therefore, lower levels of E2 in women might expose them to the double burden of PTSD and CVD development. Estradiol is known to have a protective effect on the vasculature of women via the production of nitric oxide (11, 12) which peripherally improves endothelial function and vascular stiffness (13-16). Arterial stiffness is a well-known risk factor for CVD that can be estimated non-invasively in humans (17)^,(18). Thus, low E2 states, such as menopause, are also associated with vascular dysfunction and accelerated development of CVD (19)^,(20). However, few studies have focused on the association between PTSD, E2, and vascular function in young premenopausal women.

Our laboratory recently showed that young women with a clinical diagnosis of PTSD present with higher arterial stiffness and higher central blood pressure than trauma-exposed women without PTSD (21). To further understand the underlying mechanism of this arterial stiffness, we investigated in this paper the effects of E2 on arterial compliance in young trauma-exposed women with and without a PTSD diagnosis. We hypothesized that E2 levels would be negatively associated with arterial stiffness in women with PTSD, when controlling for factors such as phase of the menstrual cycle, age, body mass index, and PTSD symptom severity.

METHODS

Ethical Oversight:

This study was approved by the Institutional Review Board of the University of Minnesota. Informed consent was obtained from all participants, and all procedures and protocols were performed in accordance with the ethical standards of the 1964 Declaration of Helsinki and its amendments.

Study Sample:

Participants were recruited from the University of Minnesota Twin Cities (Minneapolis and St. Paul) campus, the surrounding community, women’s shelters, and through the University of Minnesota Medical Center, Fairview Hospital. Eligibility criteria, as previously described (22) included biological women aged 18-40 years who were premenopausal, had a history of traumatic events, and had the ability to provide informed consent. Exclusion criteria included trans women and men, pregnancy, breastfeeding, medical conditions such as hypertension, diabetes, heart disease, vascular disease, current illicit drug use, current substance abuse, excessive alcohol use (>2 drinks per day), hyperlipidemia, autonomic dysfunction, any serious systemic disease, psychiatric illness, severe traumatic brain injury, or the inability or unwillingness to abstain from nicotine use for at least 12 h prior to the study visits. The eligibility criteria were assessed via online screening and phone interviews. Trauma exposure was self-reported and stated in writing by the participants as part of our standardized measure for PTSD symptoms described in the Measures section. Although participants were asked to report the trauma that currently affects them the most, they were welcomed to list more than one traumatic event if applicable.

Experimental Design:

As previously described (22), we used a cross-sectional, observational study design in the present paper. After completing the online screening survey, participants who met the inclusion criteria were invited to the lab for the first two study visits. Both visits were conducted between 8:00 AM and 1:00 PM. During Visit 1, we obtained written informed consent, administered interviews and surveys, measured vital signs, and collected anthropomorphic data (height, weight, body mass index, and abdominal circumference). Three seated blood pressure (BP) and heart rate measurements were made using an automated digital blood pressure device (Omron, HEM-907XL, Omron Healthcare, Kyoto, Japan) with an appropriately sized cuff placed on the upper arm, with the arm resting at heart level after at least 5 min of quiet rest. The resting respiratory rate was measured between BP measurements by counting the number of breaths taken in one minute. Venous blood samples were collected to quantify the circulating levels of E2.

The participants returned for Visit 2, during which arterial stiffness was assessed. They were instructed to abstain from smoking, exercise, alcohol, caffeine, and over-the-counter medications affecting BP for 12 h before this visit. Participants were also asked to fast for at least 6h before the visit. Visit 2 was scheduled within 1-5 days after the onset of menstruation to ensure that arterial stiffness measurements were recorded during the early follicular phase of the menstrual cycle. Women with irregular menstrual cycles were asked to contact the laboratory at the onset of menstruation for their second visit. For each participant, a urine pregnancy test was performed to exclude pregnancy. During Visit 2, after a 10-minute rest and acclimation in the supine position, we performed pulse wave analysis and pulse wave velocity (PWV) measurement using Applanation Tonometry for the assessment of arterial stiffness (described below as well). The participants were compensated for their time spent in the study.

Measures:

PTSD checklist for DSM-5 (PCL5) with criterion A (traumatic event).

PCL5 is a 20-item self-reported questionnaire based on the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) symptoms of PTSD. As part of this standardized measure, participants were asked in criterion A to “briefly identify their worst traumatic event (if they felt comfortable doing so).” Next, participants reported how much they were bothered over the past month by symptoms related to that traumatic event, using a 5-point Likert scale (0 = ‘Not at all,’ 1 = ‘A little bit,’ 2 = ‘Moderately,’ 3 = ‘Quite a bit,’ 4 = ‘Extremely’). This questionnaire is one of the most widely used screening measures for PTSD, and the findings support PCL5 as a psychometrically sound measure in individuals at a high risk of exposure to trauma. The total PCL5 score ranges from 0 to 80. A PCL-5 cutoff score between 31-33 is indicative probable PTSD, 33-45 is considered mild, ≥ 45 ≤60 is classified as moderate, and >60 is classified as severe PTSD.

Estradiol quantification.

Blood samples were collected and centrifuged at 4 degrees Celsius, 3000 rpm for 15 minutes after clotting. The serum was then pipetted into cryovials and stored at −80 degrees Celsius until testing. Blood samples were tested at the Cytokine Reference Laboratory of the University of Minnesota. This is a CLIA’88 licensed facility (license #24D0931212). This core laboratory has extensive experience in sex hormone analysis. Serum samples were tested via immunoassays using ELISA (conventional and multiplex ELISA), fluorescence bead-based antibody systems (Luminex platform), and microfluidic (Ella) that allow multiple assays on a very small amount of material. Commercially available kits (catalog number ESD31-K01 for the ELISA kit) and reagents were used for this study. Samples were analyzed for total E2 using the ELISA platform and were tested in duplicate. The ELISA had a standard Curve Range of 20-3200 pg/mL, and a sensitivity of 10 pg/mL. Per assay protocols, extraction was not needed. No values were thrown out for being outside of the reference range. Values were interpolated from log-log or 4PL fitted standard curves.

Pulse Wave Velocity.

Arterial stiffness was quantified via applanation tonometry (SphygmoCor XCEL; Atcor Medical, Sydney, Australia) with the participant lying in the supine position. Pulse-wave velocity (PWV) measurements were performed noninvasively using a tonometer placed over the carotid artery and a BP cuff over the femoral artery. The distance from the carotid to femoral artery measurement sites was measured, and PWV was calculated as the quotient of the distance and the time delay between the carotid and femoral pressure waveforms (m/s). The SphygmoCor method uses the foot of the waveform as an onset point for calculating the time differences between the R wave and the pulse waveforms at each site (23).

Data analysis:

The Shapiro-Wilk test revealed that E2 values were not normally distributed so were transformed to log10 values before inclusion in all statistical analyses. We ran a descriptive analysis for the baseline characteristics of the overall sample and used an independent t-test to compare participants diagnosed with PTSD (PTSD+) and those without (PTSD−). We then ran bivariate Pearson product correlations for all variables of interest and potential covariates for subsequent linear regression analyses. The main variables of interest were E2 level and pulse wave velocity analysis to assess arterial stiffness. Age and BMI were planned for inclusion as covariates in linear regression analyses, given the consistent relationship between age and BMI and vascular function. We also planned to control for the phase of the menstrual cycle (luteal vs. follicular) during which E2 was measured (Visit 1). To minimize the likelihood of making a type-II error with multiple testing, bivariate Pearson product correlations were used to identify other variables to be included in the regression models to predict PWV along with E2. Next, a series of linear regression models were fitted to examine the effect of E2 on arterial stiffness in the PTSD+ versus the PTSD− group. All analyses were conducted using SPSS version 28, with a significance level of P < 0.05. Given that regression analyses are very sensitive to outliers, we removed bivariate outliers. A listwise deletion approach was used to handle the missing data in the multivariate linear regressions, resulting in a final analysis sample of 70 women (37 PTSD+ and 33 PTSD−) for the two prediction models.

RESULTS

Demographics and baseline characteristics.

Data collected during visit 1 are presented in Table 1 along with PWV and E2 levels. Of the 92 trauma-exposed women enrolled in our study, 45 were clinically diagnosed with (PTSD+) and 47 were not (PTSD−). As shown in Table 1, women with PTSD were older and had a higher BMI. Resting systolic blood pressure was comparable between the groups. However, diastolic blood pressure was higher in PTSD+ compared to PTSD−. As expected, PTSD symptoms severity (PCL5 score) was higher in PTSD+ compared to PTSD−. Consistent with our prior report (21), PWV was lower in PTSD− compared to PTSD+.

Table 1:

Baseline characteristics

	PTSD− (n=47)		PTSD+ (n=45)
	Mean ± SD	Min-Max	Mean ± SD	Min-Max	T-test p-value
Age (years)	25 ± 6	18 – 40	29 ± 8	19 – 49	0.002
BMI (kg/m2)	24 ± 4	17 – 35	29± 6	13 – 47	<0.001
SBP (mmHg)	105 ± 9	89 – 127	108 ± 11	88 - 137	0.116
DBP (mmHg)	66 ± 8	48 - 80	71 ± 9	52 – 90	0.005
Pulse Pressure (mmHg)	39 ± 7	28 – 58	37 ± 7	23 – 58	0.170
Heart Rate (beats/min)	74 ± 10	52 – 95	74± 14	55 – 120	0.091
PWV (m/s)	5.3 ± 0.6	4.2 – 6.6	6.1 ± 0.8	4.3 –7.7	<0.001
* E2 (pg/mL, log10)	1.6 ± 0.4	.9 – 2.7	1.4 ± 0.4	0.4 – 2.5	0.022
PCL5 (a.u.)	27 ± 15	2 – 62	41 ± 17	3 – 38	<0.001
	Frequency		Frequency		Chi-square p-value
PTSD Severity Category
Borderline	32		14		0.001
Mild	10		12
Moderate	4		13
Severe	1		6
Contraceptives
Yes	20		26		0.175
No	27		19
* Menstrual Cycle Phase
Follicular	15		22		0.204
Luteal	18		15

BMI, Body Mass Index; DBP, Average Diastolic Blood Pressure; SBP, Average Systolic Blood Pressure; PWV= Pulse Wave Velocity; E2, Estradiol. PCL5, Posttraumatic Stress Disorder (PTSD) Checklist for DSM5. Borderline=None or probable PTSD. *Sample size for E2 and Menstrual Cycle Phase: PTSD−, n= 33; PTSD+, n= 37.

Serum E2 levels.

The concentration of serum E2 levels (log¹⁰, pg/mL) between the groups is shown in Table 1 (mean) and depicted in Figure 1 (boxplot). E2 levels were significantly lower in PTSD+ compared to PTSD−.

Figure 1:

Serum E2 Levels in 33 women without (PTSD−) and 37 women with PTSD (PTSD+). An independent t-test was used to compare the two groups. The PTSD+ group had significantly lower levels of serum E2 serum compared to PTSD−.

Association of E2 and PWV.

To test our hypothesis that PWV is a function of E2 level, we ran a bivariate correlation analysis. Figure 2 shows the scatterplots of the relationships between PWV (m/s) and serum E2 levels (pg/mL). Panel A includes all participants with available PWV and E2 level data (n= 70), with PTSD+ in blue (n=37) and PTSD− in green (n=33). Panel B includes only PTSD− participants and Panel C includes PTSD+ participants. PWV is negatively correlated with E2 levels in PTSD− women. PWV increased with decreasing E2 level, as expected. However, this relationship was not observed in PTSD+ women. Instead, we observed a positive association between PWV and E2. Next, we ran predictive models for both groups (PTSD− and PTSD+). Table 2 shows the results of two separate multiple linear regression models conducted to investigate the effect of E2 level on PWV, while controlling for age, BMI, average diastolic blood pressure, PCL5 score, and menstrual phase (luteal vs. follicular) for PTSD− and PTSD+. These covariates were chosen based on the differences between the two groups at the baseline as well as known fluctuations in E2 across the two phases of the menstrual cycle. The predictive model for PTSD− accounted for 67% of the variance in PWV (R²= 0.670 P= .005), and the model for PTSD+ accounted for 41% of the variance in PWV (R²= 0.413 P= .058). In the PTSD− model, E2 (β= −0.461, P= 0.020) significantly contributed to predicting PWV after accounting for the other variables. In contrast, in the PTSD+ group, E2 level was not a significant predictor in the model (β=0.559, P= 0.087). Notably, in the PTSD− group, age (β= 0.045, P= 0.022), DBP (β= 0.045, P= 0.002), and PCL5 (β= 0.015, P= 0.023) were also significant predictors.

Figure 2:

Association between log10 serum E2 levels (pg/mL) and pulse wave velocity (PWV, m/s) in 70 premenopausal, healthy trauma-exposed women (age range 18-49 years; panel A). Pearson product correlation coefficient was used to assess the relationship between serum E2 and PWV. After grouping our participants into the two groups based on PTSD diagnosis (PTSD−, n=33; panel B and PTSD+, n= 37; panel C), a strong correlation was observed between E2 and PWV in our PTSD− group.

Table 2:

Predictive value of Estradiol levels on PWV

	PTSD−					PTSD+
	Estimates	95% C.I.	t-statistic	Pr(>∣t∣)		Estimates	95% C.I.	t-statistic	Pr(>∣t∣)
Constant	1.013	−1.704 – 3.729	0.795	0.220		2.362	0.075 – 4.650	2.148	0.044
E2 Logged (pg/mL)	−0.461	−0.894 – −0.027	−2.263	0.020*		0.559	−0.088 – 1.207	1.796	0.087
Age (years)	0.045	0.002 – 0.088	2.215	0.022*		0.035	−0.005 – 0.075	1.809	0.085
BMI (kg/m2)	0.014	−0.054 – 0.082	0.441	0.333		0.020	−0.033 – 0.072	0.777	0.446
DBP (mmHg)	0.045	0.017 – 0.073	3.426	0.002*		0.021	−0.008 – 0.050	1.504	0.147
PCL5 (a.u.)	0.015	0.000 – 0.031	2.171	0.023*		0.000	−0.018 – 0.019	0.034	0.974
Menstrual Cycle	0.111	−0.293 – 0.515	0.583	0.284		−0.198	−0.731 – 0.335	−0.773	0.448
	R2 = 0.670, P = 0.005; n = 33					R2 = 0.413, P = 0.058; n = 37

E2, Estradiol; BMI, Body Mass Index; DBP, Diastolic Blood Pressure; PCL5, Posttraumatic Stress Disorder Checklist for DSM5; PWV, Pulse Wave Velocity; Menstrual Cycle, Early Follicular and Luteal Phase. * Indicates significant predictors.

DISCUSSION

This study aimed to investigate the effects of E2 on arterial compliance in premenopausal trauma-exposed women with and without a PTSD diagnosis. We collected E2 and PWV during two separate study visits and controlled for the phase of the menstrual cycle during which E2 was measured in our analyses. Our main finding is that in premenopausal women exposed to trauma, a diagnosis of PTSD negatively affects the relationship between serum E2 levels and arterial compliance, even after accounting for age, BMI, blood pressure, PTSD symptom severity and phase of the menstrual cycle. Specifically, there was a negative correlation between PWV and E2 levels in the PTSD− group. However, in the PTSD+ group, we found an unexpected positive correlation between PWV and E2 levels. These findings suggest that in premenopausal women, PTSD may disrupt the normal physiological link between E2 and vascular function, where higher E2 is known to be associated with lower arterial stiffness. This finding underscores the protective role of E2 in the development of cardiovascular disease.

Before the onset of menopause, women are thought to have a lower risk of CVD than males of the same age, possibly because of the influence of sex hormones. However, after menopause, the risk of CVD in women surpasses that in males (24-26). The available literature points to a positive impact of estrogen on vascular function. Sex hormones such as E2 are known to have protective effects against vascular (endothelial function and arterial compliance) protective effects in women (27, 28). Similarly, our current findings support a negative correlation between E2 and PWV in trauma-exposed women without PTSD, suggesting that as serum E2 levels decreased, arterial compliance decreased as well. On the other hand, the positive association in women diagnosed with PTSD suggest that this disorder might be blunting the protective effects of E2 on the vasculature by altering its sensitivity, thereby increasing the risk of CVD in women prior to the onset of menopause. Thus, PWV is increasing in PTSD+ women in the midst of high circulating E2 levels. Estrogen has both short- and long-term effects on the blood-vessel wall (29). In premenopausal women, 17β-estradiol produced by the ovaries is the chief circulating estrogen. After menopause, serum E2 concentrations decrease to values comparable to or lower than those found in men of similar age(29). In normal blood vessels, the endothelium releases nitric oxide in response to a variety of stimuli, causing vasodilation (30). In blood vessels with dysfunctional endothelium, the release of nitric oxide is reduced, thereby these stimuli cause the contraction of smooth muscle and paradoxical vasoconstriction (30). In our present analyses, E2 was a significant predictor of PWV in our control group (PTSD−) even after accounting for known modulators of vascular function, such as age, BMI, blood pressure, and menstrual cycle, which is expected in premenopausal women. However, in the PTSD+ group, E2 was not a significant predictor of PWV. Available data suggest that estrogen influences the bioavailability of endothelial-derived nitric oxide and, through nitric oxide-mediated increases in cyclic guanosine monophosphate, causes relaxation of vascular smooth muscle cells (29). However, the beneficial effects of E2 in PTSD extend beyond the vasculature.

There is prior evidence of more aversive affective experiences, such as stronger intrusive memories, greater fear responses, and impaired fear inhibition in low-estradiol states in women (9, 31, 32). Glover et al. revealed that lower circulating levels of E2 in women with PTSD may be linked to impaired physiological reactions to fear-potentiated startle stimuli (32). Similarly, Reider et al. reported an association between lower E2 levels and higher trauma-related symptoms, with trauma-exposed women showing an abnormal pattern of stress response to a trauma reminder (33). Furthermore, recent data by Sartin-Tarm et al. support the moderating role of E2 in the relationship between PTSD severity and arousal response habituation (i.e., from fear conditioning to extinction training), such that high E2 protects against the negative effect of severe PTSD symptoms on fear habituation (34). In line with these studies, we report in the current paper that trauma-exposed premenopausal women with a clinical diagnosis of PTSD had lower serum levels of E2 than those without a diagnosis of PTSD. Thus, in addition to the dissociation between E2 and arterial compliance, our participants also showed a reduction in circulating E2 levels associated with PTSD diagnosis. This additional finding supports the hypothesis that PTSD may increase CVD risk in this population by inhibiting or decreasing the protection that E2 affords women before menopause. Importantly, these findings underscore the potential of E2 levels to serve as a predictive biomarker for PTSD, CVD, and related conditions.

Our study had some limitations. The final sample size with available data in each group was smaller than the original sample size because not all participants had available blood samples for E2 quantification. However, despite the small sample size, we have sufficient power for mechanistic studies in humans to draw preliminary conclusions. Second, we did not ask for doctors' notes or official diagnosis letters for participants with PTSD and relied solely on self-reporting. However, based on the PTSD symptom severity data collected, women with PTSD had a higher score. Furthermore, the current study did not assess the trauma load or timeline of trauma (i.e., time elapsed since trauma exposure). Future research should investigate whether the timeline of trauma and exposure to different types of traumatic events are differentially related to vascular markers of CVD risk. Finally, E2 was measured during visit 1, which contrary to visit 2, was not standardized according to the menstrual cycle. However, we did control for the phase of the menstrual cycle when E2 was measured in our regression analyses.

Our manuscript has some strengths we would like to highlight. First, we controlled for the phase of the menstrual cycle. Second, all our participants were women who had experienced trauma; thus, the lack of association between E2 and PWV in the PTSD+ group cannot be merely attributed to trauma exposure.

In summary, the findings of this study have important clinical implications. The negative correlation between E2 levels and arterial stiffness in our PTSD participants seems to imply a protective association between E2 levels and the onset of arterial stiffness. However, the lack of a significant association in our PTSD+ group indicates the need for vascular health intervention resources. Our data show that in this sample of premenopausal women exposed to trauma, a PTSD diagnosis may disrupt the physiological link between E2 and vascular function (arterial compliance), further highlighting the protective role of E2 against cardiovascular disease and slowing arterial stiffness. Further investigations as to why this protective association is lost once one’s trauma meets a PTSD diagnosis need to be done, considering these extreme implications for otherwise healthy women who are diagnosed. Additionally, quantifying cortisol levels could help to evaluate the contribution of overall stress to these measures. In PTSD, elevated cortisol levels and heightened sympathoadrenal axis activity are key mechanisms that can significantly disrupt sleep quality, leading to insomnia, and contribute to increased CVD risk by causing vascular dysfunction and increasing blood pressure over time (22, 35, 36).

DATA AVAILABILITY

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.