Vascular health years after a hypertensive disorder of pregnancy: The EPOCH Study

Abstract

Background:

Preeclampsia is associated with a two-fold increase in a woman’s lifetime risk of developing atherosclerotic cardiovascular disease (ASCVD), but the reasons for this association are uncertain. The objective of this study was to examine the associations between vascular health and a hypertensive disorder of pregnancy among women ≥ 2 years postpartum.

Methods:

Pre-menopausal women with a history of either a hypertensive disorder of pregnancy (cases: preeclampsia or gestational hypertension) or a normotensive pregnancy (controls) were enrolled. Participants were assessed for standard ASCVD risk factors and underwent vascular testing, including measurements of blood pressure, endothelial function, and carotid artery ultrasound. The primary outcomes were blood pressure, ASCVD risk, reactive hyperemia index measured by EndoPAT and carotid intima-medial thickness. The secondary outcomes were augmentation index normalized to 75 beats per minute and pulse wave amplitude measured by EndoPAT, and carotid elastic modulus and carotid beta-stiffness measured by carotid ultrasound.

Results:

Participants had a mean age of 40.7 years and were 5.7 years since their last pregnancy. In bivariate analyses cases (N=68) were more likely than controls (N=71) to have hypertension (18% vs. 4%, p=0.034), higher calculated ASCVD risk (0.6 vs 0.4, p=0.02), higher blood pressures (systolic: 118.5 vs. 111.6 mm Hg, p=0.0004; diastolic: 75.2 vs 69.8 mm Hg, p=0.0004), and higher augmentation index values (7.7 vs. 2.3 p=0.03). They did not, however, differ significantly in carotid intima-media thickness (0.5 vs. 0.5, p=0.29) or reactive hyperemia index (2.1 vs 2.1, p=0.93), nor in pulse wave amplitude (416 vs 326, p=0.11), carotid elastic modulus (445 vs 426, p=0.36), or carotid beta stiffness (2.8 vs 2.8, p=0.86).

Conclusion:

Women with a prior hypertensive disorder of pregnancy had higher ASCVD risk and blood pressures several years postpartum, but did not have more endothelial dysfunction or subclinical atherosclerosis.

INTRODUCTION

Preeclampsia is associated with a two-fold increase in the lifetime risk of developing atherosclerotic cardiovascular disease (ASCVD), the leading cause of death in women.¹ Preeclampsia is also a leading cause of maternal morbidity and mortality, and affects 5% to 8% of all pregnancies, roughly eight million cases each year worldwide.² The hallmarks of preeclampsia are hypertension and proteinuria that develop after 20 weeks of gestation. The causes of preeclampsia are not fully understood, and the only effective therapy is delivery of the placenta.² Although the major clinical manifestations of preeclampsia typically resolve within a few days to weeks postpartum, a history of preeclampsia is associated with long-term adverse effects on the mother’s health, especially an increased risk of developing cardiovascular disease later in life.³ In prospective cohort studies, preeclampsia has been associated with significantly increased risk for subsequent ASCVD, independent of conventional cardiac risk factors such as hypertension, diabetes, and smoking, as well as factors such as socioeconomic status, education, race, and ethnicity.^3–17 The mechanistic links between preeclampsia and subsequent ASCVD are poorly understood, and difficult to investigate because the process develops over a long life course.

Preeclampsia is characterized during pregnancy by endothelial dysfunction, inflammation, immune alternations, and metabolic disturbances, but the extent to which these changes persist postpartum is unknown. Endothelial function, measured by non-invasive peripheral arterial tonometry, could be used to assess the impact of preeclampsia on subsequent vascular health, as it is central to the pathogenesis of both preeclampsia and ASCVD.^18–19 Additionally, measurement of carotid intima-media thickness (CIMT), an anatomic marker of early vascular changes and ASCVD risk, may also provide a measure of the association of preeclampsia with sub-clinical atherosclerosis.^20–22

The EPOCH (Effect of Preeclampsia on Cardiovascular Health) study is a prospective comparison of women who had preeclampsia or gestational hypertension with women who had a normotensive pregnancy, with a goal of improving understanding of the vascular and metabolic changes that increase the risk of late ASCVD after preeclampsia. Here we describe the clinical characteristics and several measures of vascular health among the women enrolled two to fifteen years postpartum.

METHODS

Women were eligible for this study if they were between 21 and 55 years of age, pre-menopausal, and had delivered a baby at Lucille Packard Children’s Hospital (Stanford Health Care) at least two years prior to enrollment. Women were excluded if they had any of the following conditions prior to their index pregnancy: chronic hypertension, diabetes mellitus, heart disease, chronic kidney disease, autoimmune disease, or cancer. Women who had a subsequent pregnancy and live birth within the two years prior to potential study enrollment were also excluded. Cases met the 2013 criteria of the American College of Obstetricians and Gynecologists², which have been used in prior studies of preeclampsia and ASCVD. Healthy controls were identified as otherwise eligible women who had no history of preeclampsia, gestational diabetes, gestational hypertension, or proteinuria. This study was approved by the Stanford Institutional Review Board, and all participants provided written informed consent.

The study was funded by grant HL139844 from the National Heart, Lung, and Blood Institute, Bethesda, MD, with additional support from grants UL1 TR001085 and UL1 TR003142 from the National Institutes of Health, Bethesda, MD, and also from the Stanford Dunlevie MFM Center for Discovery, Innovation and Clinical Impact and the H&H Evergreen Fund. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the paper and its final contents.

Potential study participants were identified using the Stanford Research Repository (STARR) database of electronic health data, which includes diagnosis codes as well as information such as the mother’s body mass index and the baby’s gestational age at delivery. We excluded women who lived more than an hour’s driving distance from our research clinic. Potential study participants were approached about the study through an e-mail sent by the Stanford Research Participation Program; women without a valid e-mail address were contacted through the US Postal Service. We made vigorous attempts to recruit participants of all races, ethnicities, and levels of socioeconomic status, and translated all study materials into Spanish. Potential participants were directed to a website that described the rationale and design of the study and provided an eligibility screening survey.

We identified potential cases using discharge diagnosis codes for preeclampsia (International Classification of Disease, version 9 (ICD-9 codes) 624.4 – 624.6, or ICD-10 codes O14 – O15). We reviewed the medical records of interested and apparently eligible participants to determine if they met the inclusion criteria. We then contacted eligible cases to obtain informed consent and schedule a single research clinic visit. We then identified up to 50 potential healthy control subjects for each case, frequency matched on age (within one year), years since delivery (within one year), race and ethnicity, and approached them for participation using the same outreach methods as described above (i.e. initial contact by email or US mail, with an invitation to visit the study enrollment website).

Study visits

Participants completed self-administered surveys (in English or Spanish) in REDCap prior to the study visit to collect data on demographics, family history, medical history, reproductive history, and personal habits (e.g., diet, exercise, and smoking). Demographic data included age, self-reported race (participants could select multiple race categories), self-reported Hispanic ethnicity, and self-reported weight at 18 years of age.

We collected a fasting blood sample and a urine sample during a morning study visit, and measured the participant’s height and weight. We measured systolic and diastolic blood pressure (BP) four times and averaged the final three measurements. We used EndoPAT (Itamar Medical, Vinings, GA) to assess endothelial function based on its user dependability and previous use as a research tool in pregnant patients with preeclampsia. We performed EndoPAT testing in a warm, dimly lit, quiet room after a fast of at least eight hours to measure vascular function, specifically: reactive hyperemia index (RHI), a measure of endothelial-dependent vasodilation; augmentation index normalized to a heart rate of 75 beats per minute (AI@75), a marker of arterial stiffness; and pulse wave amplitude, a marker of arterial tone.²³ A sonographer then imaged the right and left carotid arteries using a linear array probe (L11–3 MHz, Philips iE33 Ultrasound System, Andover, MA) to obtain three 10-second loops at anterior, posterior, and lateral angles, which were stored as Digital Imaging and Communications in Medicine (DICOM) loops, with a resolution of 1024 × 768 pixels/cm. We measured carotid intima medial thickness (CIMT) online using a frame from each of the loops over a one-centimeter segment located in the far wall of the distal aspects of the right and left common carotids just proximal to the bifurcation, using semi-automated edge-detection software (QLab Philips, iE33, Andover, MA). Similar to EndoPAT, we chose CIMT because of its established use as a tool and risk marker for ASCVD, and its ability to be done readily in participants. We averaged measurements from the right and left carotid arteries of CIMT, carotid elastic modulus, and carotid beta stiffness from each participant.

Definitions

We defined preeclampsia using the 2013 criteria of the American College of Obstetricians and Gynecologists,² which required an elevated BP (>= 140 systolic or >=90 diastolic, measured twice at least four hours apart) after 20 weeks gestation in a woman who was previously normotensive, plus one additional feature (proteinuria, thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms). Cases with new onset elevated BP without evidence of an additional feature after chart review were classified as having gestational hypertension. Preeclampsia with severe features was defined by any of the following features: BP >= 160 systolic or >=110 diastolic; platelet count < 100,000/microliter; liver enzymes more than twice the upper limit of normal; progressive renal insufficiency (serum creatinine >=1.1 mg/dl or double prior levels); pulmonary edema; or new-onset cerebral or visual disturbances.²

We defined clinical hypertension at the study visit as either a measured BP >= 140/90 mm Hg, or reported use of an antihypertensive medication. We considered participants who were not taking an anti-hypertensive medication to have borderline hypertension if they had either a systolic BP between 130 and 139 mm Hg or a diastolic BP between 80 and 89 mm Hg. We measured fasting levels of total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and glucose. We defined hypercholesterolemia as either an LDL-C > 100 mg/dL, a total cholesterol > 200 mg/dL, or reported use of a lipid-lowering medication (i.e., a statin). We defined diabetes as a prior diagnosis of diabetes, treatment with a glucose-lowering medication, or a fasting blood sugar > 125 mg/dl. We defined microalbuminuria as an albumin/creatinine ratio of >= 30 mg/g in a urine sample. We calculated each participant’s ten-year ASCVD risk based on Pooled Cohort Equations, as recommended by the ACC/AHA guidelines.^{24, 25}

Statistical methods

The primary study outcomes were RHI from the EndoPAT, CIMT from the carotid ultrasound, blood pressure, and ASCVD ten-year risk. Secondary outcomes were measures of vascular stiffness (AI@75) and vascular tone (pulse wave amplitude) from the EndoPAT, and carotid elastic modulus and beta-stiffness from the carotid ultrasound.

We present categorical data as counts and percentages, and compared study groups using the chi-squared test (or Fisher’s exact test when a cell count was ≤5). We present continuous data as box-and-whisker plots, with means and standard deviations, and compared study groups using a Student’s t-test. We used a generalized linear regression model to determine the relationship between case versus control and each outcome, adjusted for age, body mass index, White, Black, and Asian race, Hispanic ethnicity, systolic blood pressure, and use of an antihypertensive medication.

RESULTS:

We identified and attempted to contact 3236 women who were potentially eligible as cases based on diagnosis codes for preeclampsia in their electronic medical records (Figure 1). Of the 436 women who responded, 165 completed online surveys and indicated interest in participating, 100 provided informed consent, and 68 completed a study visit with performance of either an EndoPAT or carotid ultrasound test (cases). We identified and attempted to contact 8990 control subjects without a hypertensive disorder of pregnancy who had been frequency matched to cases on age, race, ethnicity, and time since last known pregnancy (Figure 1); 184 responded, 86 provided informed consent for participation, and 71 completed a study visit with performance of either an EndoPAT or carotid ultrasound test (controls).

Figure 1.

Study flow diagram

Study participants had a mean age of 40.7 years, and were a mean of 5.7 years since their most recent live birth, with no significant differences between cases and controls (Table 1). Participants were representative of the demographics of the local population: 58% were White, 28% were Asian, 4% were Black, and 23% reported Hispanic ethnicity. Among the 68 cases, 34 (50%) had preeclampsia with severe features, 25 (37%) had preeclampsia without severe features, and 9 (13%) had gestational hypertension.

Table 1.

Demographics and laboratory measures of cases and controls at the study visit

	Cases (n=68)	Controls (n=71)	P-value*
Age (years)	40.6 (6.5)	40.8 (5.3)	0.86
Race (Self-reported categories) Ϯ
White	37 (54%)	43 (61%)	0.46
Asian or Pacific Islander	16 (24%)	23 (32%)	0.24
Black/African American	2 (3%)	3 (4%)	1
American Indian or Alaskan Native	1 (1%)	2 (3%)	1
No race reported Ϯ	15 (22%)	8 (11%)	0.09
Hispanic ethnicity Ϯ	19 (28%)	13 (18%)	0.18
Parity **			0.43
Primiparous	24 (36%)	21 (30%)
Multiparous	43 (64%)	50 (70%)
Education			0.11
High school diploma or less, vocational training	10 (15%)	3 (4%)
Some college, bachelor’s or associate degree	18 (26%)	20 (28%)
Some graduate school, graduate or professional degree)	40 (59%)	48 (68%)
Health insurance			0.12
Private insurance	55 (81%)	66 (93%)
MediCal/Medicaid	9 (13%)	4 (6%)
No health insurance	4 (6%)	1 (1%)
Years since last pregnancy **	5.9 (3.6)	5.4 (3.8)	0.39
Gestational age (weeks) at index pregnancy	37.5 (3.3)	38.8 (2.7)	0.01
Body mass index (kg/m2)	26.6 (4.9)	24.2 (5.0)	0.005
Body mass index categories			0.003
<25 kg/m2	31 (46%)	52 (73%)
25–29.9 kg/m2	18 (26%)	11 (15%)
≥30 kg/m2	19 (28%)	8 (11%)
Ever smoker	9 (13%)	9 (13%)	0.93
Diabetes	1 (1%)	0 (0%)	0.49
Use of antihypertensive medication	7 (10%)	2 (3%)	0.09
Hypertension			0.03
No	44 (65%)	56 (79%)
Borderline hypertension	12 (18%)	12 (17%)
Clinical hypertension	12 (18%)	3 (4%)
Hypercholesterolemia	39 (57%)	32 (45%)	0.15
Family history of heart disease or stroke	21 (31%)	20 (28%)	0.73
Total cholesterol (mg/dL)	188.9 (36.9)	183.5 (38.4)	0.40
LDL cholesterol (mg/dL)	111.3 (33.8)	99.9 (36.7)	0.06
HDL cholesterol (mg/dL)	59.9 (15.6)	68.6 (15.1)	0.001
Fasting glucose (mg/dL)	89.4 (8.2)	87.5 (7.7)	0.16
Microalbuminuria (≥30 mg/g creatinine)	2 (3%)	1 (1%)	0.61
ASCVD risk	0.6 (0.5)	0.4 (0.5)	0.02

Data are reported as mean (standard deviation) or number (percentage)

^*Categorical variables: Chi-square test or Fisher exact test (for cells with < 5); Continuous variables: t-test

^**Missing data from one case on parity and years since last pregnancy

^ϮAdds to more than 100% because 3 cases and 7 controls selected multiple race categories, and 14 of the 15 cases, and all 8 controls, who did not select a race category reported being Hispanic.

Compared with controls, cases had a higher mean body mass index at the study visit (26.6 vs. 24.2, p=0.005), and had an earlier earlier gestational age at delivery in the index pregnancy (37.5 vs. 38.8 weeks, p=0.01). Cases were more likely to have developed clinical hypertension (18% vs. 4%, p=0.03) and had a higher estimated ten-year ASCVD risk (0.6 vs 0.4, p=0.02). At the study visit, cases also had higher mean systolic (118.5 vs. 111.6 mm Hg, p=0.0004) and diastolic blood pressures (75.2 vs. 69.8 mm Hg, p=0.0004) even though 10% of cases versus 3% of controls were on an antihypertensive medication (Table 2, Figure 2). Cases had lower levels of HDL-cholesterol (59.9 vs. 68.6 mg/dl, p=0.001) but similar levels of total cholesterol and LDL-cholesterol (Table 1). Microalbuminuria was infrequent in both groups (Table 1).

Table 2.

Mean vascular measurements in cases of hypertensive disorders of pregnancy and healthy controls

Variable	Cases (n=68)	Controls (n=71)	p-value
Systolic blood pressure (mm Hg)	118.5 (10.7)	111.6 (11.5)	0.0004
Diastolic blood pressure (mm Hg)	75.2 (8.5)	69.8 (9.2)	0.0004
Reactive hyperemia index	2.1 (0.7)	2.1 (0.7)	0.93
Augmentation index @ 75 bpm	7.7 (13.4)	2.3 (15.6)	0.03
Pulse wave amplitude (au)	416 (336)	326 (302)	0.11
Carotid intima-medial thickness (mm)	0.50 (0.1)	0.50 (0.1)	0.29
Carotid elastic modulus (mmHg)	445 (105)	426 (131)	0.36
Carotid beta stiffness	2.8 (0.2)	2.8 (0.2)	0.86

Data presented as mean (standard deviation)

Missing (n=) vascular outcome measurements by case/control, respectively: RHI 1/3; AI@75:1/2; PWV 1/3; CIMT: 1/3; CEM 2/2; CBS 2/2

Figure 2.

Box-and-whisker plots of outcome measures (vertical axis) in cases (left) and controls (right). The top of the box is the 75^th percentile, the line in the middle of the box is the 50^th percentile, and the bottom of the box is the 25^th percentile. The low and high whiskers indicate the most extreme point that is less than or equal to 1.5 times the 25^th or 75^th percentile, respectively. Red points represent data from cases with preeclampsia with severe features, blue points represent data from cases who did not have preeclampsia with severe features, and the black dots represent control subjects. Abbreviations: PE w/SF = preeclampsia with severe features; PE = preeclampsia without severe features; GHTN = gestational hypertension; CTL: control.

There were no significant differences between cases and controls in vascular reactivity (measured by RHI) or subclinical atherosclerosis (measured by CIMT) (Table 2). The mean RHI was 2.1 in both cases and controls (p=0.93); the mean CIMT was 0.50 in both cases and controls (p=0.29) (Table 2, Figure 2). In adjusted models, RHI was not significantly associated with any baseline characteristic except Asian race (p=0.049) (Table 3), whereas CIMT was significantly associated with older age (p<.0001), higher systolic BP (p=0.01), and higher body mass index (p =0.0006).

Table 3.

Multivariable linear regression model of selected vascular outcomes, adjusted for baseline covariates

Factor	Reactive Hyperemia Index		Carotid intima medial thickness		Augmentation index normalized to a heart rate of 75 bpm
	Adjusted parameter (std error)	P-value	Adjusted parameter (std error)	P-value	Adjusted parameter (std error)	P-value
Case *	0.01 (0.13)	0.91	0.001 (0.01)	0.92	3.17 (2.36)	0.18
Age (per 1 year)	0.01 (0.01)	0.61	0.01 (0.001)	<.0001	0.82 (0.2)	0.0001
White Race	0.26 (0.19)	0.17	0.02 (0.01)	0.24	0.19 (3.51)	0.96
Black Race	−0.08 (0.37)	0.84	0.03 (0.02)	0.16	2.57 (6.78)	0.71
Asian Race	0.42 (0.21)	0.049	0.03 (0.01)	0.06	2.66 (3.86)	0.49
Hispanic Ethnicity	0.31 (0.19)	0.11	−0.002 (0.01)	0.86	6.67 (3.54)	0.06
Systolic blood pressure (per 1 mm Hg)	0.0002 (0.01)	0.97	0.001 (0.0004)	0.01	0.43 (0.11)	0.0001
Body Mass Index (per 1 kg/m2)	0.005 (0.01)	0.72	0.004 (0.001)	0.0006	−0.39 (0.25)	0.13
Antihypertensive medication use	0.02 (0.26)	0.95	−0.01 (0.02)	0.41	−0.95 (4.8)	0.84

^{* -}Unadjusted parameter estimates for case vs. control status (standard error, p-value): RHI = 0.01 (0.12, p=0.92); CIMT = 0.01 0.01, p=0.29); AI@75 = 5.44 (2.49, p=0.03)

The RHI and CIMT models included 67 cases and 68 controls; the AI@75 model included 67 cases and 69 controls.

Cases had a significantly higher mean augmentation index normalized to a heart rate of 75 beats per minute (AI@75, p=0.03) compared with the controls (Table 2), but did not differ significantly in pulse wave amplitude, carotid elastic modulus and carotid beta-stiffness (Table 2). In an adjusted model, AI@75 was significantly associated with age (p=0.0001) and systolic BP (p=0.0001), but not with case vs. control status (Table 3).

DISCUSSION

In this prospective study women who were at least two years after a hypertensive disorder of pregnancy (primarily preeclampsia) had significantly higher systolic and diastolic blood pressures compared with controls, as well as an elevated 10-year risk of ASCVD based on the Pooled Cohort Equations. Cases were, however, not significantly different from controls in subclinical atherosclerosis (measured by carotid intima-media thickness) or endothelial function (measured by the reactive hyperemia index). These vascular changes in relatively young women (average age 40.8 years) suggest they may have a greater lifetime burden of hypertension and a higher risk of ASCVD.

Women who have had preeclampsia or gestational hypertension had higher measured blood pressures in this study, a higher prevalence of chronic hypertension, and greater use of antihypertensive medications. These differences developed after pregnancy, since women with chronic hypertension prior to pregnancy had been excluded from this study. This finding is consistent with previous studies that have shown that subsequent development of hypertension is more common after a hypertensive disorder of pregnancy.^1,3 It is uncertain whether women who develop clinical hypertension after experiencing a hypertensive disorder of pregnancy have an underlying predisposition to develop hypertension, or whether their vascular function was altered by the development of preeclampsia. Genome-wide association studies have reported associations between preeclampsia and genetic variants related to hypertension²⁶, suggesting that a tendency towards higher blood pressure may predate pregnancy. Regardless of the mechanism, these findings suggest that women who develop a hypertensive disorder of pregnancy should be followed more closely to detect the subsequent development of chronic hypertension.

Carotid intima-medial thickness (CIMT) is a measure of subclinical atherosclerosis that is associated with elevated risk of subsequent ASCVD. We found that CIMT increased significantly with older age and higher blood pressure levels (Figure 3), but was not affected by a history of a hypertensive disorder of pregnancy. Some, but not all, previous studies have reported higher levels of CIMT in women with a history of preeclampsia compared with control subjects²⁰. Garovic and associates studied 40 women with a history of preeclampsia at an average age of 59.4 years, and reported a significantly higher CIMT in cases than in age and parity-matched comparison subjects²⁰. By contrast, McDonald and colleagues found no differences in CIMT between 109 women with a history of preeclampsia and 218 age-matched women with no such history, at a median age of 49 years in both groups²⁷. A meta-analysis of ten studies of women at least ten years postpartum found significantly higher CIMT among women with versus without prior preeclampsia²⁰. We studied women at younger ages (mean 40.8 years) and at shorter intervals postpartum (mean 5.7 years) than these previous studies, so differences in CIMT may develop over time, especially considering the significant association of CIMT with age (Figure 3).

Figure 3.

Plots of outcome measures (vertical axis) versus participant age (left panels) and mean systolic blood pressure (right panels). In the left panels, red circles indicate cases and black squares indicate controls; in the right panels, filled circles and squares represent participants not taking an antihypertensive medication, and empty circles and squares represent participants who had taken an antihypertensive medication within 48 hours of the study visit. Simple linear regression lines were fitted separately to the data from cases and controls, and drawn in red for cases and in black for controls. Abbreviations: HTN med = antihypertensive medication

Endothelial dysfunction is central to the pathogenesis of preeclampsia during pregnancy and may persist long after delivery and resolution of the acute manifestations of preeclampsia. EndoPAT testing is a measure of endothelial function that is based on the concept of endothelial-dependent flow-mediated vasodilation. In contrast to some prior studies, we found no difference in the reactive hyperemia index (RHI) between cases and controls at an average of 5.7 years postpartum, and we also found no association of RHI with the age of the participant, or with any other baseline factor (Table 3). Orabona and colleagues studied 30 women with a history of early-onset preeclampsia, 30 women with a history of late-onset preeclampsia, and 30 women with normotensive pregnancies at a mean 2.3 years postpartum, and found that RHI was significantly lower in the women with early onset preeclampsia than in controls or in women with late onset preeclampsia²⁸. By contrast, Kvehaugen and colleagues found no difference in RHI between 26 women with preeclampsia and 15 controls with a mean of 6.0 years postpartum, although the subgroup who delivered an infant small for gestational age had a significantly lower RHI²⁹. Other studies have measured endothelial function during and after preeclampsia using a different technique than we used, flow-mediated vasodilation³⁰. A meta-analysis of 37 studies found that flow-mediated vasodilation was significantly reduced in women with preeclampsia during pregnancy and for up to six months postpartum, but was not significantly different after ten or more years of follow-up³¹. The associations between preeclampsia and flow-mediated vasodilation were quite heterogeneous in nine cross-sectional studies conducted between six months and three years postpartum. The evidence from available studies suggests that endothelial dysfunction develops during preeclampsia, persists for some time after delivery, but appears to resolve over longer follow-up intervals. Our study was performed a mean of 5.7 years postpartum, which is later than prior studies that reported endothelial dysfunction after preeclampsia.

Limitations

Our study has several strengths. We recruited participants who had well documented medical and pregnancy histories, used validated assessments of vascular function performed by trained personnel, and assessed vascular health and ASCVD risk at a mean of 5.7 years after pregnancy. Our study also has several limitations. Only a small percentage of women who we contacted about potential enrollment were sufficiently interested to respond, visit the study website, and complete an eligibility survey. Consequently, study participants may differ from the overall population in a variety of ways. We made vigorous attempts to enroll a diverse group of subjects with respect to race, ethnicity, and socioeconomic status, and the study participants were generally representative of the local population demographics. Since cardiovascular disease and hypertensive disorders of pregnancy affect Black women disproportionately, we made special attempt to recruit Black subjects, but were limited by the relatively low percentage of Black residents in our local area (2.9% in Santa Clara County). We recruited for this study between 2020 and 2022, and our efforts were adversely affected by the COVID-19 pandemic; some eligible women were unable, or unwilling, to complete the research visit under pandemic restrictions. The final study sample size was modest, which may have affected the statistical power to detect differences between cases and controls. We did not have enough participants with either early preeclampsia or severe preeclampsia to analyze their potential effects on the vascular measures. Finally, participants were relatively young (mean age of 40.7 years), so sufficient time may not have elapsed for differences in sub-clinical ASCVD to have developed.

Conclusions

Women who had a hypertensive disorder of pregnancy have increased blood pressures and an elevated ASCVD risk several years postpartum. Closer follow-up for development of hypertension after pregnancy and more intensive management of persistent elevated blood pressures may reduce late cardiovascular disease after a hypertensive disorder of pregnancy.

Non-standard Abbreviations and Acronyms:

AI@75

Augmentation index normalized to a heart rate of 75 beats per minute

ASCVD

Atherosclerotic cardiovascular disease

CIMT

Carotid intima-media thickness

EPOCH

Effect of Preeclampsia on Cardiovascular Health

HDP

Hypertensive disorders of pregnancy

REDCap

Research Electronic Data Capture

RHI

Reactive hyperemia index