This was a prospective study approved by the Institutional Review Board (IRB) at the University of Mississippi Medical Center in Jackson, MS. Participants were recruited from August 2018 to June 2020 from outpatient clinics affiliated with the University. Written informed consent was obtained from all study participants who met the inclusion/exclusion criteria. Women between 18 and 40 years of age were eligible if they were between 20 and 26 weeks pregnant at recruitment with a single fetus free from fetal anomalies. Participants were excluded if they had a history of stillbirth, fetal death or loss of a child, history of alcohol or drug abuse, HIV/AIDS, history of preterm birth, or history of a high-risk pregnancy (i.e., preeclampsia; PreE). Also to ensure that we were only examining women with new cases of anxiety/depression a history or current diagnosis of anxiety/depression and/or a psychosis was also an exclusion factor. All of these factors have been reported to contribute to the development of PMAD so they were excluded for this initial study. In the current analysis women who met or exceeded the cut-off on either the PASS or EPDS were categorized as PMAD.

As outlined in the Figure 1, women completed two study visits (recruitment and 10 weeks post-recruitment) and a series of validated questionnaires were administered to assess stress (Perceived Stress Scale; PSS), depression (Edinburgh Postnatal Depression Scale; EPDS) and anxiety (Perinatal Anxiety Screening Scale; PASS) were completed.

Women who exceeded the cut-off scores were provided feedback on their study results and offered additional mental health resources outside of the study.

Participants also completed a survey of sociodemographic characteristics (i.e., employment and relationship status, education level). To evaluate socioeconomic status, we utilized the distressed community index (DCI). The DCI measures economic wellbeing at a zip code level by assessing location based disparities and evaluating the relationship between them and health outcomes, demographics and educational level (13). Communities are scored and divided into five tiers based on these distressed levels: prosperous (<20.0), comfortable (20.1–40.0), mid-tier (40.1–60.0), at risk (60.1–80.0) and distressed (> 80.0).

Health data was extracted from electronic medical records, and all data were stored in REDCap (14). The following electronic health data was collected: residential zip code, medical history including details on obstetric characteristics; body mass index (BMI), labor and delivery details, and final hypertensive status (e.g., PreE, gestational hypertension). Intrapartum complications such as fever, postpartum hemorrhage, and shoulder dystocia were also collected. From infant records the following was collected: birth weight and information pertaining to presence of infant complications at birth.

After all women had completed the study, participants were grouped based on the hypertensive status at the time of discharge from labor and delivery. Women were categorized as normotensive pregnancy (n = 39) or hypertensive pregnancy (n = 26). Women were considered to have a hypertensive pregnancy if they had chronic hypertension (cHTN), gestational hypertension (gHTN), PreE or super-imposed PreE (siPreE).

At each study visit blood was collected by a study nurse or a member of the facility laboratory. Samples were collected in EDTA (BD, United States) and serum collection tubes and allowed to clot. Tubes were centrifuged at 2,000 RPM at 4°C for 20 min, aliquoted and frozen at −20°C. Maternal plasma levels of specific cytokines were measured using commercially available Human Premixed Multi-Analyte Magnetic Luminex Assay (R&D systems, United States) and assayed in duplicate according to the manufacturers recommendations. Multiplex panels were designed for the following proteins and cytokines, all of which are associated with perinatal anxiety and depression or hypertensive pregnancy (15–17): soluble fms-like tyrosine kinase-1 (sFlt-1), soluble endoglin (sEng) and placental growth factor (PlGF). All Luminex assays were read and analyzed via the Luminex MAGPIX analyzer. Interleukin-17 (IL-17, Boster Bio, CA, United States) and tumor necrosis factor-alpha (TNF-α; RayBiotech, GA, United States) were measured via individual ELISAs according to the manufacturer’s protocol.

Summary statistics were presented as means (standard deviations) for continuous variables and using frequencies and percentages for categorical variables. Correlation among continuous variables was estimated using Pearson’s correlation coefficient. Chi-square test, Fisher’s exact tests, T-test were used to compare participant characteristics between HDP and Normotensive categories. Similar statistical tests were used to compare demographic characteristics of women among PMAD and non-PMAD categories. Multivariable adjusted logistic regression models were used to assess associations between HDP, maternal age, maternal BMI and race. Data for EPDS, PASS and PSS scores were presented at Visit 1 and Visit 2 for all women as well as by HDP and Normotensive categories. All statistical analyses were performed with SAS version 9.4 (SAS Institute Inc., Cary, NC, United States). A two-sided p < 0.05 was considered statistically significant.

The baseline sociodemographic characteristics of study participants participating in the study are shown in Table 1. Women who had a HDP were significantly older (p = 0.006) and had a higher BMI (41.9 ± 13.5 kg/m², p = 0.0002) relative to women with normotensive pregnancies. There was a significant difference in the racial distribution between women experiencing HDP and normotensive pregnancies (p = 0.0007). As shown in Table 1, black women comprised 84.5% of HDP while white women comprised 51.2% of normotensive pregnancies. Multiple logistic regression showed that women who were older (p = 0.02), had a higher BMI (p = 0.01) or who were of black race (p = 0.01) had significant odds of having a HDP (Supplementary Table 1).

There was not a significant difference in self-report There was a significant difference in relationship status, as women who had a HDP were less likely to be married and more likely to be single compared to women with a normotensive pregnancy (p = 0.01). However, no differences were found between the groups in relation to education level (p = 0.61), current employment status (p = 0.17) or their DCI (p = 0.76).

Women with a HDP gave birth significantly earlier (35.59 ± 3.17 weeks) than normotensive women (38.04 ± 2.15 weeks, p = 0.0004; Table 2). This corresponded to a higher incidence of preterm birth (p = 0.0001) and a significantly lower infant birth weight among women with HDP (p = 0.009; Table 2). Despite the higher rate of preterm delivery there were no differences in mode of delivery (p = 0.12), maternal labor complications (p = 0.34) or infant complications (p = 0.25) between the two groups of women.

There was not a significant difference in self-reported depression, anxiety or stress between HDP and normotensive women at either visit 1 or visit 2. Women were also ranked as low depression vs. probable depression based on their EPDS value. Although 16.9% of women were at risk for probable depression, no significant differences were observed between normotensive and HDP pregnancies when this relationship was further examined (p = 0.99, Table 3). Results from the PASS was used to evaluate minimal anxiety vs. moderate/severe anxiety among women with normotensive and HDP pregnancies at visits 1 and 2. At both visits women with normotensive pregnancies were more likely to screen with moderate/severe anxiety relative to women with a HDP (p = 0.11 at visit 1; p = 0.07 at visit 2).

Likewise, perceived stress decreased in both group of women between visits 1 (p = 0.22) and 2 (p = 0.09), with normotensive women being more likely to report higher levels of stress relative to women with a HDP (Table 3).

Among HDP women, 53.8% (n = 14) women had PreE vs. a diagnosis of cHTN/gHTN (n = 12). There were no statistically significant differences between women in any of the demographic factors with the exception of BMI. Women who developed PreE/siPreE were significantly heavier at the time of enrollment (48.14 ± 15.1 vs. 36.12 ± 8.5 kg/m²; Supplementary Table 2) compared to women with cHTN/gHTN. When labor complications were examined, women with PreE/siPreE were more likely to have a preterm delivery relative to women with cHTN/gHTN (78.6 vs. 25%, p = 0.02; Supplementary Table 2). However, there were not any significant differences between any other variables. Women with PreE/siPreE were also not more likely to have a PMAD at either visit 1 (p = 0.99) or visit 2 (p = 0.99).

In the current analysis there 27 women were categorized as PMAD and compared to 38 non-PMAD women. There was not a significant relationship between HDP and PMAD at either visit 1 (p = 0.21) or visit 2 (p = 0.07). The overall percent of women with a PMAD decreased 16% (p = 0.06) between visit 1 and visit 2 regardless of their hypertensive status, with no significant changes in outcomes due to a change in PMAD status.

No significant differences were found in baseline demographic factors between women who had a PMAD and those who did not (Table 4). However, when race was collapsed into black vs. non-black, black women were 40% more likely to report a PMAD, O.R. 1.39 95% CI (0.54–3.63) relative to non-black women. Current employment status (p = 0.39) or education level (p = 0.62) did not have an impact on PMAD, with the majority of women experiencing a PMAD (59%) having completed a high school diploma/GED or some college (Table 4). Relationship status also had no significant impact on PMAD (p = 0.11). DCI was not associated with PMAD (p = 0.12) as the majority of women lived in at risk or distressed communities regardless of their PMAD status. Experiencing a PMAD was not associated with an increased gestational age at delivery (p = 0.51), pre-term birth (p = 0.43) or decreased infant birth weight (p = 0.53, Table 4). Similarly women with PMAD did not experience more labor complications (p = 0.15) or infant complications (p = 0.71) relative to women without PMAD (Table 4).

A significant relationship was found between PSS and PMAD, as women with PMADs had significantly higher levels of PSS at both visit 1 (p < 0.0001) and visit 2 (p < 0.0001) relative to women without PMADs. Additionally, the mean level of PSS did not significantly change between visit 1 and visit 2 among women with (p = 0.99) or without PMADs (p = 0.94; Figure 2A). To further explore the impact of perceived stress, we examined the levels of perceived stress and demographic factors.

At both visit 1 (p = 0.006) and visit 2 (p = 0.03) women who were married were more likely to experience moderate levels of stress, while non-married women in relationships or who were single were more likely to experience severe levels of stress (Figure 2B). Interestingly, a significant relationship was found between PSS and DCI at visit 2 (p = 0.02), however, upon further examination it was found that 68.9% of the women (22/32) living in distressed zip codes reported moderate-severe stress. However, PSS was not significantly associated with any other socioeconomic or demographic factors nor was it associated with labor or birth outcomes (Supplementary Data).

Biological variables were assessed to determine if they differed in the 2nd or 3rd trimester among women with and without PMADs. At visit 1 there was a significant difference in IL-17 levels between women with and without PMADs (p = 0.03, Figure 3A). Circulating levels of TNFα were also found to be significantly increased (p = 0.009; Figure 3B), with significant differences at visit 1 (p = 0.04). We also examined, anti-angiogenic growth factors that are commonly associated with a HDP to see if they changed and we did not see any significant differences between groups in circulating soluble endoglin (p = 0.12). There was not a group effect on circulating sFlt-1 levels, but there was a time effect as sFlt-1 significantly increased between visit 1 and visit 2 regardless of maternal mood (p = 0.01, Figure 3C). The same effect occurred with circulating PlGF levels, which increased as pregnancy increased (p < 0.0001, Figure 3D).