Stress trajectory and hypertension 2–7 years after delivery: A nuMoM2b-HHS study

Abstract

Background:

Adverse pregnancy outcomes (APOs) are associated with higher risk of developing chronic hypertension. The objectives of this study were to determine whether patterns of perceived stress during and after pregnancy were associated with blood pressure and incident hypertension 2–7 years after delivery, and whether having an APO modified this association.

Methods:

Analyses utilized data from the prospective Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be (nuMoM2b) Heart Health Study (HHS) cohort. Perceived stress was assessed using the Perceived Stress Scale in the first and third trimester and 2–7 years after delivery. Latent class trajectory analysis characterized subgroups with similar patterns of perceived stress over time. APOs were abstracted from medical charts and included hypertensive disorders of pregnancy, preterm birth, small-for-gestational-age, and stillbirth. Multivariable regression models evaluated the independent effects of perceived stress on systolic and diastolic blood pressure and incident hypertension and 2–7 years after delivery.

Results:

Three distinct stress trajectory groups emerged, delineated by persistently low, moderate, and high stress levels. No associations between stress trajectory group and blood pressure or incident hypertension were observed after adjustment for covariates. However, there was a significant interaction between stress trajectory group and APO on blood pressure (P for interaction=0.04). Stress trajectory group was associated with higher blood pressure only among those with APO (β = 1.991 ± 0.819 mmHg, P=0.02) but not without APO (β = 0.040 ± 0.471 mmHg, P=0.93).

Conclusions:

These findings suggest that elevated perceived stress may contribute to higher blood pressure, specifically among women who had an APO.

Introduction

Stressful life events and perceived stress are associated with cardiovascular dysfunction and disease burden in women over the lifespan.^1,2 Pregnancy is associated with amplified psychosocial stress,³ attributable in part to pregnancy-specific correlates such as physical symptoms and discomfort with advancing gestation, occupational ramifications,⁴ interpersonal stress, or worry related to fetal/neonatal health status.⁵ Psychosocial stress has important clinical implications during pregnancy, including greater risk for suboptimal cardiovascular health,⁶ and the development of adverse pregnancy outcomes (APOs).^4,7 Importantly, increased maternal stress extends beyond the perinatal period. Indeed, increases in perceived stress across trimesters predict greater stress levels 2–7 years after delivery.⁸ However, it is not known whether patterns of psychosocial stress across pregnancy and in the years following delivery are linked to the accumulation of cardiovascular risk factors such as high blood pressure.

APOs are a robust risk factor for development of hypertension and cardiovascular events after pregnancy.^9–13 Experiencing an APO is associated with higher perceived stress 2–7 years after delivery, independent of pre-existing risk factors and changes in stress during pregnancy.⁸ This suggests that APO may contribute to persistent stress driven by peripartum sequalae.^14,15 However, it is unclear the extent to which psychosocial stress during and after pregnancy contributes to development of cardiovascular risk among women with a history of APO.

Accordingly, the objectives of the present analyses were to 1) determine whether patterns of perceived psychosocial stress during pregnancy and 2–7 years after delivery are associated with blood pressure and incident hypertension in the 2–7 years postpartum, and 2) assess whether these associations differ between women with and without an APO. We hypothesized that higher perceived stress across pregnancy and the 2–7 year follow-up period would be associated with higher maternal blood pressure, with a stronger association observed in women who experienced an APO.

Methods

Design:

Anonymized data from the parent nuMoM2b-HHS study are available through the National Institute of Child Health and Human Development (NICHD) and National Heart, Lung, and Blood Institute (NHLBI) data repositories DASH (https://dash.nichd.nih.gov/) and BioData Catalyst (https://biodatacatalyst.nhlbi.nih.gov/; See Major Resources Table in the Supplemental Materials). This is a secondary analysis of the multi-site, prospective pregnancy cohort Nulliparous Pregnancy Outcomes Study – Monitoring Mothers-to-be (nuMoM2b) Heart Health Study (HHS), for which detailed methods are reported elsewhere (NCT01322529, NCT02231398).^16,17 Briefly, participants were enrolled in the nuMoM2b study during the first trimester (<14 weeks gestational age) and followed through delivery and for 2–7 years from parturition to investigate associations between APOs and cardiovascular disease risk. Participants were nulliparas with viable singleton pregnancies. Exclusion criteria included maternal age <13, a history of ≥3 prior pregnancies, donor oocyte, and known fetal malformation. Participants were eligible for the in-person HHS visit when 2–7 years had elapsed since the nuMoM2b index birth, if they were not currently pregnant, and >6 months postpartum from a subsequent pregnancy. For the present analysis, exclusion criteria included index pregnancy loss <20 weeks or termination, hypertension at enrollment (defined by self-reported history of hypertension, anti-hypertensive medication use, or baseline blood pressure ≥130/80 mmHg), missing first trimester or HHS blood pressure readings, or missing all perceived stress assessments. Perceived stress and brachial blood pressure measures were obtained in the first trimester (gestational age 6 weeks +0 days to 13 weeks +6 days), third trimester (22 weeks +0 days to 29 weeks +6 days), and 2–7 years after the index pregnancy (mean 3 ± 1 years). All participants provided informed consent before enrollment. Each study site’s Institutional Review Board approved the study protocol, and this secondary analysis was approved by the University of Delaware Institutional Review Board.

Perceived stress:

The exposure of interest was maternal perceived stress, as assessed by the ten-item Perceived Stress Scale (PSS) in the first and third trimester and at the 2–7 years follow-up visit. The PSS assesses the degree to which life events are perceived as uncontrollable, unpredictable, or overwhelming in the previous month on a five-point frequency Likert scale, such that higher score indicates higher perceived stress (range, 0–40).¹⁸ The PSS has good internal and test-retest (<4 week testing intervals) reliability and high construct validity with criterion such as severity of stressful life events, affect, and general health in pregnant women.^19,20

Latent class linear mixed models (LCLMMs) were employed to identify subgroups of participants with similar patterns of PSS scores across timepoints (e.g., trajectory).²¹ Under this model, the outcome is assumed to be normally distributed, and each subgroup (also called a “latent class” in this context) is identified by a specific trajectory that is generally common among participants that belong to the subgroup. Participants with at least one PSS score were assigned to a trajectory group, with a missing at random assumption for participants missing two timepoints. Given the heightened sensitivity of normality tests to mild deviations from normality at large sample sizes, we opted to rely on visual inspection of the empirical distribution of the observed data to assess feasible normality of the PSS scores.^22,23 All three PSS scores demonstrated a clear, single mode coupled with approximate symmetry in support of feasible normality. Skewness and kurtosis were ≤ 0.67 and ≤ 3.4, respectively, for all three PSS scores. To complement this approach, 10,000 samples of size N = 25 were drawn with replacement and Shapiro-Wilk test applied to each, as this hypothesis test is most reliable when N < 50. Formal hypothesis testing indicated normal distribution in 84%, 80%, and 92% of samples for first trimester, third trimester, and postpartum scores respectively.

Blood pressure and hypertension:

The primary outcomes were resting systolic blood pressure and diastolic blood pressure measured in clinic 2–7 years following the index pregnancy, modeled as continuous outcomes. Blood pressure was measured using a standardized protocol: Blood pressure readings were obtained in triplicate following five minutes of seated rest via automated oscillometric device (Omron HEM-907XL, Omron Healthcare Incorporated, Lake Forest, Illinois).¹⁷ The average of the last 2 systolic and diastolic pressures were used as the primary outcome. The secondary outcome was incident hypertension, defined by blood pressure ≥130/80mmHg or initiation of anti-hypertensive medication.²⁴

Adverse pregnancy outcomes:

Pregnancy outcomes were determined from review of medical records by a certified clinical chart abstractor and adjudicated by principal investigators as appropriate. APOs included hypertensive disorders of pregnancy, preterm birth, small-for-gestational-age birth, and stillbirth defined per standard diagnostic criteria. Hypertensive disorders of pregnancy included all types of preeclampsia regardless of severity or timing of onset, eclampsia, and gestational hypertension.²⁵ Preterm birth was defined as a livebirth at <37 weeks gestational age regardless of indication. Small-for-gestational-age birth was classified using Alexander growth curves and defined as birthweight for gestational age less than 5^th percentile.²⁶ Stillbirth is defined as fetal demise after 20 weeks gestational age.

Baseline demographics and covariates:

Covariates included baseline (first trimester) blood pressure, maternal demographics (e.g., age, race, ethnicity), body mass index (BMI), time between delivery and HHS visit, and subsequent pregnancy. Baseline blood pressure readings were collected in duplicate by trained research staff using an aneroid sphygmomanometer with appropriate cuff size following a 10 minute rest period. Detailed interviews were conducted to collect demographic characteristics and medical history at enrollment. BMI (kg/m²) was assessed by research personnel from measures of height and weight obtained via balance-beam scale and stadiometer, respectively. Interviews were conducted at 6-month intervals from the nuMoM2b index birth until the HHS visit to determine if participants had a subsequent pregnancy.

Statistical analyses:

Trajectory group characteristics were compared using analysis of variance with post hoc pairwise comparisons with Tukey’s adjustment or chi-square tests with post hoc Bonferroni adjustment as appropriate. Multivariable linear models were used to examine associations between stress trajectory group and blood pressure, and logistic regression models were used to examine associations between stress trajectory group and incident hypertension. Similarly, multivariable linear models were used to examine the associations between continuous PSS score at each time point (first trimester, third trimester, 2–7 years after index pregnancy) and blood pressure, and logistic regression models were used to examine the association between PSS scores at each time point and incident hypertension. PSS scores were z-transformed such that regression coefficients represent change in blood pressure or odds of incident hypertension per standard deviation increase in PSS score.

To assess whether APOs moderated the associations between psychosocial stress and blood pressure, all linear and logistic regression models were repeated with the addition of an interaction (stress trajectory group x APO; perceived stress score x APO). All analyses were then repeated with stratification by APO history. APOs were modeled as a binary exposure defined by ≥1 APO (a composite of hypertensive pregnancy disorders, small-for-gestational-age or preterm birth, stillbirth) versus none. Exploratory analyses were also conducted to investigate associations by individual APO compared with no APO.^8,12 Exploratory analyses were only conducted for hypertensive disorders of pregnancy compared with no APO, due to insufficient sample sizes for small-for-gestational-age birth, preterm birth, and stillbirth outcomes.

Unadjusted and adjusted models are reported. A priori covariates in adjusted models included maternal age, BMI, first trimester blood pressure, race, ethnicity, insurance payor, tobacco use, time elapsed since the index pregnancy, and subsequent pregnancy. Analyses were conducted using R (Ver. 4.3.2). Statistical significance was set to α=0.05 for all analyses.

Results

Stress trajectory group characteristics:

After exclusion criteria were applied, a total sample of n=3,322 (74% of nuMoM2b-HHS cohort) were available for the present analysis (Figure 1). Most participants had PSS score data at all three timepoints (n=3,058 [92.1%]) used to generate the latent classes. Three stress trajectory groups were identified, characterized by persistently low (n=785), moderate (n=1,593), and high (n=944) PSS score levels (Figure 2).

Figure 1:

Inclusion and exclusion flowchart. NuMoM2b-HHS includes participants enrolled in Nulliparous Pregnancy Outcomes Study – Monitoring Mothers-to-be (nuMoM2b) during pregnancy and Heart Health Study (HHS) 2–7 years after the index pregnancy. BP indicates blood pressure; APO, adverse pregnancy outcome; HDP, hypertensive disorders of pregnancy; PTB, preterm birth; SGA, small-for-gestational age; BMI, body mass index.

Figure 2:

Perceived stress scale score trajectories across pregnancy trimesters and 2–7 years postpartum (PP). Stress trajectory groups were created using latent trajectories.

Participant characteristics by stress trajectory group are presented in Table 1. Compared with the low stress trajectory group, participants with moderate or high stress trajectories were younger, had higher body mass index and lower educational attainment, were more likely to be of Black or African American race, Hispanic, and to have history of tobacco use, but were less likely to be partnered. The proportion of those in the high and moderate stress groups with two prior pregnancies at baseline (7%) was greater than that of the low stress groups (4%, P < 0.05). The prevalence of index pregnancy APO was similar between stress trajectory groups, except for greater prevalence of preterm birth in the high stress group (10%) compared with the moderate and low stress groups (7%, P<0.05). Sixty-four percent of participants had at least one subsequent pregnancy between the index birth and the HHS visit (median: 1; IQR: 0–2 pregnancies). Participants in the high stress trajectory group were less likely to have a subsequent pregnancy compared with those in the low and moderate stress groups (P < 0.001, Table 1).

Table 1:

Participant characteristics

Characteristics	Low Stress N=785	Moderate Stress N=1593	High Stress N=944	p-value
Maternal age, years	28 ± 5	27 ± 5a	25 ± 6a,b	<0.001
Body mass index, kg/m2	25.0 ± 5.0	25.4 ± 5.3	26 .8± 6.4a,b	<0.001
Race, n (%)				<0.001
• American Indian or Alaska Native	2 (<1%)	3 (<1%)	2 (<1%)
• Asian	25 (3%)	55 (4%)	30 (4%)
• Black or African American	67 (9%)	164 (11%)	153 (18%)a,b
• Pacific Islander	2 (<1%)	2 (<1%)	4 (<1%)
• White	612 (83%)	1151 (79%)	592 (70%)a,b
• Multi-racial	34 (5%)	90 (6%)	64 (6%)a
Hispanic, n (%)	92 (12%)	232 (15%)	179 (19%)a,b	<0.001
Highest degree attained, n (%)				<0.001
• Less than high school	19 (2%)	73 (5%)a	119 (13%)a,b
• Highschool	64 (8%)	147 (9%)	151 (16%)a,b
• Some college	112 (14%)	280 (18%)	235 (25%)a,b
• Associates	73 (9%)	178 (11%)	124 (13%)a
• Bachelors	255 (33%)	524 (33%)	196 (21%)a,b
• Graduate	262 (33%)	391 (25%)a	119 (13%)a,b
Health insurance payor				<0.001
• Private/Commercial	641 (82%)	1202 (76%)a	542 (56%)a,b
• Public/Military	117 (15%)	334 (21%)a	367 (40%)a,b
• Self-pay/other	23 (3%)	53 (3%)	30 (3%)
Partnered, n (%)	764 (97%)	1532 (96%)	854 (91%)a,b	<0.001
Gravidity at enrollment, n (%)				0.03
• 1 (current)	611 (78%)	1198 (75%)	698 (74%)
• 2	142 (18%)	279 (18%)	185 (20%)
• 3	32 (4%)	116 (7%)a	61 (7%)a
Tobacco use within 3 mos of index pregnancy, n (%)	65 (8%)	209 (13%)a	236 (25%)a,b	<0.001
Systolic blood pressure, mmHg	107 ± 9	107 ± 9	108 ± 10	0.67
Diastolic blood pressure, mmHg	66 ± 6	66 ± 7	65 ± 7	0.06
Pre-pregnancy Diabetes, n (%)	3 (<1%)	11 (<1%)	9 (1%)	0.69
Follow-up duration, years	3.1 ± 0.9	3.2 ± 0.9	3.2 ± 0.9	0.61
Subsequent pregnancy, n (%)	524 (67%)	1049 (66%)	553 (57%)a,b	<0.001
Index pregnancy outcomes
Any APO, n (%)	228 (29%)	452 (29%)	291 (31%)	0.46
Small for gestational age, n (%)	31 (4%)	64 (4%)	40 (4%)	0.95
Stillbirth, n (%)	4 (<1%)	7 (<1%)	0 (0%)	0.11
HDP, n (%)	167 (21%)	345 (22%)	195 (21%)	0.85
Preterm birth, n (%)	51 (7%)	117 (7%)	93 (10%)a	0.02
Perceived stress scale score (0–40)
• First trimester	6.2 ± 3.3	11.4 ± 3.9a	19.5 ± 5.6a,b	<0.001
• Third trimester	4.9 ± 3.0	10.4 ± 4.0a	18.0 ± 5.8a,b	<0.001
• Postpartum	7.5 ± 3.6	14.9 ± 4.4a	20.9 ± 6.5a,b	<0.001

Comparisons are by low, moderate, and high stress trajectory groups as identified by latent class analyses. APO indicates adverse pregnancy outcome; HDP, hypertensive disorders of pregnancy.

^aP<0.05 vs Low;

^bP<0.05 vs Moderate in post hoc pairwise comparisons.

Associations between stress trajectory group and blood pressure 2–7 years after pregnancy:

There was no significant association between stress trajectory group and systolic blood pressure in the entire sample (Figure 3A; Supplemental Table S1). High stress trajectory was associated with greater diastolic blood pressure relative to the low stress trajectory group in unadjusted models (β = 1.249, P < 0.01), but this relation was attenuated with adjustment (Figure 3B; Table S1). Stress trajectory was not significantly associated with odds of incident hypertension (Table 2).

Figure 3:

Association of stress trajectory with (A) systolic and (B) diastolic blood pressure 2–7 years postpartum among the entire cohort (closed circles) and with (C) systolic and (D) diastolic blood pressure stratified by adverse pregnancy outcomes (APO; open squares) and without APO (open circles). Blood pressures values are adjusted means and standard error. All crude and adjusted parameter estimates are reported in Table S1. *P<0.05, indicates significant association between stress trajectory and blood pressure among those with APO.

Table 2:

Crude and adjusted associations between stress trajectory and incident hypertension.

		Entire cohort
Incident hypertension	Low Stress N=785	Moderate Stress N=1593		High Stress N=944
	OR (95% CI)	OR (95% CI)	p-value	OR (95% CI)	p-value
Model 1	Reference	1.13 (0.90, 1,42)	0.30	1.20 (0.93, 1.54)	0.16
Model 2	Reference	1.13 (0.89, 1.45)	0.32	0.98 (0.77, 1.35)	0.90
		Without APO
	N=557	N=1141		N=653
	OR (95% CI)	OR (95% CI)	p-value	OR (95% CI)	p-value
Model 1	Reference	1.10 (0.82, 1.48)	0.52	1.12 (0.80, 1.55)	0.52
Model 2	Reference	1.06 (0.77, 1.45)	0.73	0.85 (0.58, 1.23)	0.38
		With APO
	N=228	N=452		N=291
	OR (95% CI)	OR (95% CI)	p-value	OR (95% CI)	p-value
Model 1	Reference	1.16 (0.8, 1.69)	0.43	1.23 (0.82, 1.84)	0.31
Model 2	Reference	1.27 (1.17, 1.91)	0.24	1.29 (0.82, 2.02)	0.27

Data are odds ratio (OR) and 95% confidence interval (CI). Model 1: Unadjusted. Model 2: adjusted for age at enrollment, race, ethnicity, insurance payor, smoking, body mass index at enrollment, follow-up duration, and subsequent pregnancy.

Associations between perceived stress and blood pressure 2–7 years after pregnancy:

There was no significant association between perceived stress score at any timepoint and systolic blood pressure 2–7 years after the index pregnancy (Table 3). Unadjusted associations were present for first trimester (β = 0.416, P = 0.01), third trimester (β = 0.417, P = 0.01), and postpartum (β = 0.573, P < 0.001) perceived stress score with greater diastolic blood pressure, but these associations similarly did not persist with adjustment for covariates (Table 3). There was no association between perceived stress score at any timepoint and odds of incident hypertension (Table S2).

Table 3:

Crude and adjusted associations between perceived stress scores at each time point and postpartum blood pressure

			Entire Cohort
Blood pressure 2–7 years postpartum	First trimester PSS N=3311		Third trimester PSS N=3215		Postpartum PSS N=3172
Model 1	β (SE)	p-value	β (SE)	p-value	β (SE)	p-value
Systolic blood pressure, mmHg	0.061 (0.179)	0.73	0.092 (0.181)	0.61	0.050 (0.183)	0.78
Diastolic blood pressure, mmHg	0.416 (0.163)*	0.01	0.417 (0.165)*	0.01	0.573 (0.166)*	<0.001
Model 2	β (SE)	p-value	β (SE)	p-value	β (SE)	p-value
Systolic blood pressure, mmHg	−0.107 (0.195)	0.59	−0.112 (0.193)	0.56	−0.180 (0.191)	0.35
Diastolic blood pressure, mmHg	0.250 (0.176)	0.16	0.222 (0.174)	0.20	0.302 (0.172)	0.08
			Without APO
	N=2311		N=2278		N=2213
Model 1:	β (SE)	p-value	β (SE)	p-value	β (SE)	p-value
Systolic blood pressure, mmHg	−0.023 (0.201)	0.91	0.150 (0.205)	0.46	0.035 (0.206)	0.86
Diastolic blood pressure, mmHg	0.344 (0.184)	0.06	0.526 (0.187)*	0.005	0.574 (0.189)*	0.002
Model 2:
Systolic blood pressure, mmHg	−0.207 (0.222)	0.35	−0.093 (0.221)	0.67	−0.211 (0.219)	0.34
Diastolic blood pressure, mmHg	0.134 (0.201)	0.50	0.225 (0.201)	0.26	0.301 (0.198)	0.13
			With APO
	N=969		N=923		N=928
Model 1:	β (SE)	p-value	β (SE)	p-value	β (SE)	p-value
Systolic blood pressure, mmHg	0.165 (0.361)	0.65	−0.315 (0.364)	0.37	−0.163 (0.364)	0.65
Diastolic blood pressure, mmHg	0.481 (0.329)	0.15	−0.057 (0.331)	0.86	0.312 (0.332)	0.35
Model 2:
Systolic blood pressure, mmHg	0.237 (0.395)	0.59	−0.274 (0.382)	0.47	−0.229 (0.376)	0.54
Diastolic blood pressure, mmHg	0.625 (0.354)	0.08	0.108 (0.344)	0.75	0.178 (0.335)	0.60

Data are unstandardized parameter estimates and standard error (SE). PSS indicates z-transformed perceived stress scale score (continuous). Model 1: Unadjusted. Model 2: adjusted for age at enrollment, race, ethnicity, insurance payor, smoking, body mass index at enrollment, systolic or diastolic blood pressure at enrollment, follow-up duration, and subsequent pregnancy.

^*P<0.05.

Associations between stress and blood pressure 2–7 years after pregnancy among those with and with APO:

There was a significant interaction between stress trajectory group and APO on diastolic blood pressure (P for interaction = 0.04; Figure 3D) and a similar effect on systolic blood pressure, albeit not statistically significant (P for interaction = 0.06; Figure 3C). In stratified analyses, moderate stress trajectory group was only associated with higher diastolic blood pressure among women with APO (β = 1.991 ± 0.819 mmHg, P=0.02), but there was no association among those without APO (β = 0.040 ± 0.471 mmHg, P=0.93; Figure 3D, Table S1). There were no interactions of PSS score and APO on postpartum blood pressure at any time point (P for interaction all ≥ 0.17). There were no interactions of stress trajectory group or PSS score and APO on incident hypertension (P for interaction all ≥ 0.25). There were no interactions between stress trajectory group or PSS score and hypertensive disorders of pregnancy on blood pressure or incident hypertension (P for interaction all ≥ 0.10).

Discussion

In the nuMoM2b HHS cohort, we characterized patterns of perceived psychosocial stress across pregnancy trimesters through 2–7 years after delivery among nulliparous women and related these patterns to blood pressure and incident hypertension. We identified three distinct trajectories of perceived stress delineated by low, moderate, and high stress levels. Contrary to our hypothesis, stress trajectory groups were not independently associated with blood pressure or incident hypertension 2–7 years after the index birth in the entire cohort after covariate adjustment. APO moderated the effect of stress on diastolic blood pressure such that stress trajectory group was associated with greater blood pressure among women who had an APO but not among women who had an uncomplicated pregnancy, consistent with our hypothesis. This result persisted after accounting for demographic and lifestyle factors that are associated with both APO risk and blood pressure. Taken together, these findings indicate that women with a history of APO may be particularly susceptible to adverse effects of psychosocial stress on cardiovascular health.

This study prospectively examines stress trajectories across pregnancy and postpartum, implications for maternal blood pressure, and the moderating role of APO. Pregnancy, childbirth, and becoming a parent represent major life events associated with exposure to novel stressors like physical symptom burden, interpersonal stress, fear of childbirth, or fetal health concerns.³ In a previous analysis of the nuMoM2b HHS cohort, prenatal PSS score and change in PSS score between the first and third trimester were positively associated with PSS score 2–7 years after delivery.⁸ Notably, this association was not attenuated with elapsed time since delivery. Consistent with this result, we observed three linear, positive stress trajectories. These findings are concordant with evidence that the PSS in part reflects stable traits including stress appraisal, and that increases in PSS score correspond to significant changes in life circumstances such as the transition to parenthood.²⁷

Stress trajectory groups differed across demographic, lifestyle behavior, and socioeconomic indicator variables in the first trimester, which parallels the contribution of social determinants of health (SDOH) – the multifaceted circumstances or environment in which a person lives – to perceived stress.^28,29 Other pregnancy-specific factors associated with amplified perceived stress during pregnancy include pregnancy intendedness, history of infertility, and prior pregnancy loss.³⁰ The former factors were not assessed; however, we observed higher gravidity in the high and moderate stress trajectory groups relative to the low group, indicative of prior pregnancy loss or termination in this nulliparous cohort. Taken together, it is plausible that these SDOH or pregnancy history factors may be causally linked to perceived stress during and after pregnancy. Further research is needed to understand the key determinants of perceived stress trajectories during and after pregnancy, including the types of stressors and exacerbating or protective factors.

The presence of stressors and perceived stress have been associated with cardiovascular risk.^1,31–33 In the Women’s Health Initiative study, exposure to stressful life events (e.g., serious illness, job loss) and social strain were associated with greater risk for coronary heart disease and stroke among postmenopausal women, independent of demographic and socioeconomic factors.¹ Similarly, psychosocial stress has implications for cardiovascular health during pregnancy. In a prior analysis of the nuMoM2b cohort, psychosocial stress exposure groups were identified via latent class analyses based on scales such as perceived stress, anxiety and depression symptoms, economic stress, social support, and discrimination.⁶ Greater stressor exposure was associated with poorer cardiovascular health defined by the presence of Life’s Essential 8 risk factors. In the present study, stress trajectory group was associated with blood pressure in unadjusted models in the entire cohort. This finding was specific to diastolic blood pressure, congruent with prevalent isolated diastolic hypertension in women within ten years after pregnancy.^33,34 However, this association between stress trajectory group and blood pressure was abolished with adjustment for key demographic, lifestyle behavior, and socioeconomic factors. Thus, these interconnected traditional cardiovascular risk factors (e.g., BMI, tobacco use) and SDOH may explain the association between stress during and after pregnancy and blood pressure. Indeed, SDOH including neighborhood-level deprivation and individual-level socioeconomic disadvantage are associated with the development of persistent postpartum hypertension after hypertensive disorders of pregnancy³⁵ and greater cardiovascular disease risk 2–7 years after delivery regardless of APO history.^36,37 Accordingly, domains of SDOH may contribute to blood pressure indirectly via higher perceived stress. Additional research is needed to elucidate the complex pathways between SDOH, perceived stress, and cardiovascular health during and after pregnancy.

The robust association between APO and subsequent cardiovascular disease is well established.^9–13 In the nuMoM2b cohort, a history of APO is associated with a 2-fold increased risk of incident hypertension 2–7 years after the index pregnancy.¹² However, the pathways through which APO augments cardiovascular risk later in life are not fully understood. In a prior analysis, certain APO including hypertensive disorders of pregnancy and indicated preterm birth were also positively associated with PSS score 2–7 years after delivery.⁸ Of note, this finding reflected an independent contribution of APO to stress after pregnancy, rather than APO acting as a mediator between early pregnancy APO risk factors or stress and postpartum stress.⁸ We did not observe differences in APO prevalence by stress trajectory group, possibly due in part to a different analytic approach and sample. Although the relative contribution of APO to postpartum perceived stress warrants additional investigation, the current study further supports the idea that APO do not serve a mediating role between stress during and after pregnancy and cardiovascular risk. Indeed, we found that APO modified the link between stress trajectory group and blood pressure, such that an association between greater stress exposure and higher postpartum blood pressure was present among those with APO but not uncomplicated pregnancy. Among women with APO, those in the moderate and high stress exposure groups had systolic blood pressure 1 mmHg higher and diastolic blood pressure 2 mmHg higher than those in the low stress group after adjustment for covariates, with average blood pressures remaining in the normotensive range. In this regard, the direct clinical implications of these modest blood pressure differences remain to be determined. However, cardiovascular risk in women increases at blood pressure levels well below clinical hypertension thresholds. Each 10-mmHg increase in systolic blood pressure above 100 mmHg is associated with a 25% increase in incident cardiovascular disease risk in women, with this relationship being more pronounced in younger women (<50 years).³⁸ Altogether, our findings indicate that while patterns of perceived stress during and after pregnancy are similar in women with and without APO, women with APO may be particularly vulnerable to the adverse influence of stress on cardiovascular function.

The physiological mechanisms by which higher stress is associated with greater blood pressure specifically among women who had an APO are unclear and likely multifactorial. In the present study, APO were defined by the presence of at least one uteroplacental-mediated complication including de novo hypertensive disorders of pregnancy (i.e., preeclampsia, gestational hypertension, eclampsia), preterm birth, small-for-gestational-age birth, or stillbirth. Impaired vascular endothelial function is a common feature thought to be involved in the pathogenesis of these APO that persists after delivery, related in part to an exaggerated response to vasoconstrictor signaling.^39–45 Endothelial dysfunction is a hallmark of vascular aging and predicts the development of hypertension in women.⁴⁶ Importantly, psychosocial stress has also been linked with endothelial dysfunction in young adults^47–49 via similar vasoconstrictor pathways observed after APO.⁴⁹ In this regard, the vascular endothelium may be uniquely vulnerable to the adverse impact of stress among women who had an APO, whereas endothelium-mediated vasodilation is preserved in women with uncomplicated pregnancy. A complementary pathway that may explain this association is altered control of blood pressure via the autonomic nervous system after APO. Indices of autonomic dysfunction including greater sympathetic activity^50–52 and blunted baroreflex sensitivity⁵³ are observed during and years after delivery among women who had a hypertensive disorder of pregnancy. Perceived stress during an acute mental stressor is similarly implicated in heightened sympathetic outflow,^54,55 and stressor exposure is directly related to greater sympathetic-mediated vasoconstriction.⁵⁶ It follows that higher psychosocial stress may further exacerbate existing autonomic dysfunction in women who had an APO. These putative endothelium-dependent and sympathetic system pathways regulate vascular tone, which may explain in part our findings specific to diastolic blood pressure given that vascular tone is a principal hemodynamic correlate of peripheral diastolic pressure.⁵⁷ It is important to note that this mechanistic insight is primarily supported by studies in women with history of hypertensive disorders of pregnancy, and less is known about the independent impact of other APOs examined in the present study as a composite exposure. Taken together, we posit that psychosocial stress may amplify the sustained cardiovascular dysfunction attributable APO, such that women with APO are more susceptible to deleterious effect of stress on blood pressure.

This study is strengthened by prospective assessment of both blood pressure and stress, the application of latent trajectory analyses to identify latent patterns of stress across pregnancy and postpartum in a large, diverse, nulliparous cohort, and APO status abstracted from medical records. Nevertheless, several limitations should be noted. First, stress was quantified by subjective appraisal of stressors (PSS score), and we are not able to characterize other components of the stress experience such as stressor exposures or affective, behavioral, or physiological stress responses (e.g, mood states, physical symptoms) which are salient to stress-related health outcomes.⁵⁸ Stressor exposures and responses may have important implications for interventional approaches in the postpartum period. PSS score was not collected in the second trimester in the parent study. In a previous pregnancy cohort study, PSS scores were not significantly different between the first and the second trimester in nulliparous women, and the difference in PSS score between the first and third trimester was comparable to those observed in this cohort (i.e., less than 2 points).⁵⁹ Thus, it is unlikely that the absence of second trimester PSS measures affected the latent stress trajectory patterns characterized in the present study. Second, we modeled APO as a composite, binary exposure and were not able to evaluate the effects of individual or comorbid APO due to insufficient sample size. It is possible that specific individual or combinations of APO (e.g., preeclampsia with fetal growth restriction) have distinct effect on stress trajectory, blood pressure, or the association thereof. Third, this cohort only included nulliparous women. Future work is needed to establish potential effects of psychosocial stress on hypertension risk among women with cumulative pregnancy exposures. Finally, causal inference is limited due to the observational nature of this study.

In conclusion, we show that longitudinal patterns of perceived psychosocial stress across pregnancy through 2–7 years after delivery may contribute to greater blood pressure following an APO, but not after uncomplicated pregnancy. Our findings further indicate that psychosocial stress may contribute to elevated cardiovascular disease risk among women with a history of APO. Future work is needed to investigate exacerbating and mitigating factors related to maternal perceived stress patterns and to identify the underlying mechanisms that link stress to cardiovascular health after APO.

Perspectives:

This study offers valuable insights into temporal patterns of psychosocial stress across first pregnancy through 2–7 years after delivery, and the implications of psychosocial stress for maternal cardiovascular risk. Three distinct stress trajectories emerged in this pregnancy cohort, characterized by stable low, moderate, and high stress levels. Higher stress trajectories over time were linked to increased blood pressure 2–7 years after delivery only among women who had an adverse pregnancy outcome, but not among women who had an uncomplicated pregnancy. This result suggests that women who had an adverse pregnancy outcome are more susceptible to the adverse effects of stress on long-term cardiovascular health. These novel findings may have important implications for preventative strategies to mitigate cardiovascular disease risk after an adverse pregnancy outcome. Current guidelines emphasize blood pressure monitoring and management after adverse pregnancy outcomes, but psychosocial stress represents an emerging cardiovascular risk factor and potential therapeutic target. Future research should examine determinants of psychosocial stress in this population, elucidate the physiological mechanisms by which adverse pregnancy outcomes amplify vulnerability to stress-related vascular risk, and evaluate interventional approaches to address the influence of psychosocial stress on cardiovascular health in this high risk population.

Supplementary Material

Tables S1 and S2

Novelty and Relevance:

What is new?

• This study identified three distinct temporal patterns of psychosocial stress across pregnancy through 2–7 years postpartum in a large, diverse cohort of first-time mothers.

• Persistently higher stress levels were associated with increased blood pressure among women who had an adverse pregnancy outcome, but not women who had an uncomplicated pregnancy.

What is relevant?

• Adverse pregnancy outcomes are established risk factors for future hypertension and cardiovascular disease.

• Women who had an adverse pregnancy outcome may be uniquely vulnerable to the adverse impact of stress on long term cardiovascular health.

Clinical/Pathophysiological Implications?

• Psychosocial stress represents a potential modifiable therapeutic target for cardiovascular risk reduction after adverse pregnancy outcomes.

Abbreviations:

APO

Adverse Pregnancy Outcomes

nuMoM2b-HHS

Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be Heart Health Study

PSS

Perceived Stress Scale

SDOH

Social Determinants of Health