Systematic Review
org
Hypertensive disorders of pregnancy and
long-term risk of maternal stroke—a systematic
review and meta-analysis
Matthew P. Brohan, MPH; Fionn P. Daly, MPH; Louise Kelly, MRCPI;
Fergus P. McCarthy, PhD; Ali S. Khashan, PhD; Karolina Kublickiene, PhD;
Peter M. Barrett, PhD
Introduction
Pregnancy has been established to be a
state of physiological stress for patients.
Hypertensive disorders of pregnancy
(HDPs), such as preeclampsia (PE), place
considerable strain on the body.1 There is a
growing body of evidence that HDPs
induce endothelial changes that impact a
patient’s physiology.2,3 These endothelial
changes have been linked to vascular pa-
From the School of Public Health, University
College Cork, Cork, Ireland (Mr Brohan, Drs
Khashan and Barrett), the School of Medicine,
University College Cork, Cork, Ireland (Mr
Brohan and Dr Daly), the Department of General
Medicine, Beaumont Hospital, Dublin, Ireland
(Dr Kelly), the Irish Centre for Maternal & Child
Health, University College Cork, Cork, Ireland (Dr
McCarthy), and the Division of Renal Medicine,
Department of Clinical Intervention, Science and
Technology, Karolinska Institutet, Stockholm,
Sweden (Dr Kublickiene).
Received Nov. 6, 2022; revised March 20, 2023;
accepted March 21, 2023.
The authors report no conflict of interest.
This study did not receive any funding.
The protocol for this systematic review was
registered in the International Prospective
Register of Systematic Reviews under identifier
CRD42021254660 in August 2021.
This study was presented at the Molecules to
People Conference of the University College
Cork, Cork, Ireland, September 2022; at the
Atlantic Corridor Conference of the University
College Cork and University of Galway,
November 2022; and at the Student Research
Conference of the Royal Academy of Medicine
Ireland, December 2022.
Corresponding author: Matthew P. Brohan,
MPH. 117307681@umail.ucc.ie
0002-9378
ª 2023 The Authors. Published by Elsevier Inc. This is
an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).
https://doi.org/10.1016/j.ajog.2023.03.034
thologies throughout the body, including
in the cardiovascular system, renal system,
and neurologic system.4e8 A consensus has
ajog.
OBJECTIVE: Hypertensive disorders of pregnancy are associated with a long-term risk for
cardiovascular disease among parous patients later in life. However, relatively little is
known about whether hypertensive disorders of pregnancy are associated with an
increased risk for ischemic stroke or hemorrhagic stroke in later life. This systematic
review aimed to synthesize the available literature on the association between hyper-
tensive disorders of pregnancy and the long-term risk for maternal stroke.
DATA SOURCES: PubMed, Web of Science, and CINAHL were searched from inception to
December 19, 2022.
STUDY ELIGIBILITY CRITERIA: Studies were only included if the following criteria were
met: case-control or cohort studies that were conducted with human participants, were
available in English, and that measured the exposure of a history of hypertensive dis-
orders of pregnancy (preeclampsia, gestational hypertension, chronic hypertension, or
superimposed preeclampsia) and the outcome of maternal ischemic stroke or hemor-
rhagic stroke.
METHODS: Three reviewers extracted the data and appraised the study quality following
the Meta-analyses of Observational Studies in Epidemiology guidelines and using the
Newcastle-Ottawa scale for risk of bias assessment.
RESULTS: The primary outcome was any stroke (undifferentiated) and secondary out-
comes included ischemic stroke and hemorrhagic stroke. The protocol for this systematic
review was registered in the International Prospective Register of Systematic Reviews
under identifier CRD42021254660. Of 24 studies included (10,632,808 study partici-
pants), 8 studies examined more than 1 outcome of interest. Hypertensive disorders of
pregnancy were significantly associated with any stroke (adjusted risk ratio, 1.74; 95%
confidence interval, 1.45e2.10). Preeclampsia was significantly associated with any
stroke (adjusted risk ratio, 1.75; 95% confidence interval, 1.56e1.97), ischemic stroke
(adjusted risk ratio, 1.74; 95% confidence interval, 1.46e2.06), and hemorrhagic stroke
(adjusted risk ratio, 2.77; 95% confidence interval, 2.04e3.75). Gestational hyper-
tension was significantly associated with any stroke (adjusted risk ratio, 1.23; 95%
confidence interval, 1.20e1.26), ischemic stroke (adjusted risk ratio, 1.35; 95% con-
fidence interval, 1.19e1.53), and hemorrhagic stroke (adjusted risk ratio, 2.66; 95%
confidence interval, 1.02e6.98). Chronic hypertension was associated with ischemic
stroke (adjusted risk ratio, 1.49; 95% confidence interval, 1.01e2.19).
CONCLUSION: In this meta-analysis, exposure to hypertensive disorders of pregnancy,
including preeclampsia and gestational hypertension, seems to be associated with an
increased risk for any stroke and ischemic stroke among parous patients in later life.
Preventive interventions may be warranted for patients who experience hypertensive
disorders of pregnancy to reduce their long-term risk for stroke.
Key words: cerebrovascular accident, chronic hypertension, gestational hypertension,
hemorrhagic, hypertensive disorders of pregnancy, ischemic, preeclampsia, stroke
been reached about the risk that HDPs
confer to the cardiovascular and renal
systems.5,9 HDPs have been shown to be
MONTH 2023 American Journal of Obstetrics & Gynecology 1.e1
Systematic Review ajog.org

overall stroke cases.18 It arises from the

AJOG at a Glance blockage of an artery supplying blood to a
Why was this study conducted? region of brain tissue.19 Known risk factors
This study aimed to synthesize the available literature on the association between include age, previous CVA, history of hy-
hypertensive disorders of pregnancy (HDP) and long-term risk of maternal stoke pertension, smoking, poor diet, elevated
because there is a relative lack of consensus on these associations. waist-to-hip ratio, diabetes mellitus, car-
diac disease, and depression.20,21 However,
Key findings the role of HDPs in the long-term risk for
HDP, preeclampsia, and gestational hypertension were all significantly associated ischemic stroke is uncertain.
with an increased risk for undifferentiated stroke, ischemic stroke, and hemor- By contrast, hemorrhagic stroke ac-
rhagic stroke. Chronic hypertension was significantly associated with ischemic counts for 10% to 25% of all new stroke
stroke. cases,18 and this arises when a weakened
blood vessel ruptures and bleeds into the
What does this add to what is known? surrounding brain tissue.22 Known risk
The strength of association between HDP subtypes and stroke subtypes varies factors for hemorrhagic stroke include
and is generally stronger for hemorrhagic stroke. There is a relative lack of hypertension, smoking, excessive
research on the association of chronic hypertension and superimposed pre- alcohol intake, and elevated waist-to-hip
eclampsia with undifferentiated and hemorrhagic stroke. Preventive in- ratio21 but it is also uncertain whether
terventions may be warranted for pregnant patients who experience HDP to HDPs play a role in the long-term risk.
reduce their long-term risk for stroke. To date, few studies have considered
the association between HDPs and
separate CVA subtypes, such as ischemic
associated with an increased risk for car- The clinical consequences of and ap- or hemorrhagic stroke. It is important to
diovascular disease (CVD). These condi- proaches to treatment vary considerably. identify whether patients with a history
tions have also been associated with an Previous systematic reviews exam- of HDPs are at varying risks for devel-
increased risk for chronic kidney disease ining the association between HDPs and oping these conditions. Given the
(CKD) and end-stage kidney disease cerebrovascular disease have reported inconsistency in the current evidence
(ESKD). However, the risk HDPs present varying results for PE and gestational base for long-term risk of stroke
to the neurovascular system is less estab- hypertension (GH). McDonald et al15 following HDPs and the underlying un-
lished. If HDPs induce endothelial changes examined a composite exposure of PE certainty over whether any association
that have an impact on vascular efficacy, it and eclampsia and found a positive as- may differ across the spectrum of cere-
is plausible that they may potentiate the sociation with cerebrovascular disease brovascular disease, a systematic review
risk for a cerebrovascular accident (CVA) (risk ratio [RR], 2.03; 95% confidence was planned to synthesize the current
in mothers later in life.8 Patients with a interval [CI], 1.54e2.67). Wu et al16 re- knowledge and to identify and discuss
history of PE may have augmented cerebral ported that a history of HDPs was asso- the limitations of the current literature
blood velocity that could lead to impaired ciated with an increased risk for stroke on this research topic. This study may
autoregulation of the cerebral microvas- (RR, 1.78; 95% CI, 1.58e2.00). Lo et al17 help to inform risk prevention strategies
culature with changes in the blood brain investigated the association between GH by examining a range of associations
barrier (BBB) function. This is reflected by and future risk of stroke. Following 1 across HDPs and stroke subtypes.
an increase in the circulating markers for pregnancy affected by GH or multiple
BBB injury among patients exposed to pregnancies affected by GH, this study Objectives
PE.10 reported no significant association with First, this systematic review aimed to pro-
The existing literature examining these future risk of stroke (RR, 1.26; 95% CI, vide an up-to-date, detailed synthesis of
associations is somewhat limited by the 0.96e1.65; and RR, 1.50; 95% CI, the available literature on the association
use of composite outcomes. A lot of the 0.75e2.99, respectively).17 It may be the between HDPs and long-term risk of
existing evidence has focused on stroke as case that individual HDPs (PE, GH, and maternal stroke. Second, it aimed to
part of a global cardiovascular other HDPs) are differentially associated examine the association between individ-
outcome11,12 instead of it as a separate with stroke subtypes. ual HDPs (PE, GH, chronic hypertension
biologic entity. By examining stroke as Strokes can be categorized broadly [CH], and superimposed PE on CH) and
part of a composite cardiovascular into ischemic and hemorrhagic strokes. ischemic and hemorrhagic stroke
outcome, the true impact of the associa- These disease entities vary in their separately.
tion between HDPs and cerebrovascular epidemiology, pathophysiology, risk
disease may be less apparent. Although factors, and prognoses. Materials and Methods
there are etiologic similarities between The incidence of ischemic stroke is Data sources and search strategy
CVD and cerebrovascular disease, these considerably higher than hemorrhagic Using the PubMed, Web of Science, and
conditions are biologically distinct.13,14 stroke, accounting for 55% to 95% of CINAHL databases, a systematic search

1.e2 American Journal of Obstetrics & Gynecology MONTH 2023

ajog.org
was conducted from inception of these
databases until December 19, 2022. We
used the following criteria to seek
appropriate studies: a population of
pregnant women, exposure to an HDP of
interest, a comparison group of women
with no such exposure, primary
outcome of any CVA, and/or secondary
outcome of ischemic or hemorrhagic
stroke. A detailed search strategy is
shown in Appendix 1. The Meta-analysis
of Observational Studies in Epidemi-
ology (MOOSE) guidelines were fol-
lowed throughout this review.23
The HDPs of interest included PE,
GH, CH, and superimposed PE on CH.
Exposure was defined by established
clinical criteria (eg, International Clas-
sification of Diseases codes), hospital
records or self-reporting of diagnosis.
CVA and ischemic and hemorrhagic
strokes could be defined by established
clinical criteria, hospital records, or self-
reporting of diagnosis. The terms CVA
and any stroke were considered to be
FIGURE
PRISMA flow diagram of study selection23
dentification
PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023.
equivalent. The protocol for this sys-
tematic review and meta-analysis was
prepared using the Preferred Reporting
Items for Systematic Review and Meta-
Analysis Protocols (PRISMA-P) guide-
lines24 and was registered in the Inter-
national Prospective Register of
Systematic Reviews (PROSPERO) data-
base in August 2021 under identifier
CRD42021254660.
Study selection
The study selection process is shown in
the Figure.24 Two independent reviewers
(M.P.B. and L.K.) reviewed the titles and
abstracts. Studies were only included if
the following criteria were met: case-
control or cohort studies that were
conducted on human participants,
available in English, reported on at least
20 cases, measured the exposure of a
history of HDPs and the outcome of
stroke (undifferentiated, hemorrhagic,
or ischemic stroke), reported a measure
of the association between the exposure
MONTH 2023 American Journal of Obstetrics & Gynecology
Systematic Review
and outcome, and demonstrated that the
diagnosis of the outcome followed the
diagnosis of the exposure. The diagnosis
of the outcome must have occurred by at
least 3 months postpartum. This is to
exclude studies that focused on the per-
ipartum period. Studies were excluded if
they involved the following: studies that
focused on CVA during pregnancy alone
or in the offspring, were not available in
English, case reports, case series, com-
mentaries, notes, editorials, or studies
published after December 19, 2022. A
table of eligibility criteria is shown in
Appendix 2.
Data extraction and quality appraisal
Two investigators (M.P.B. and F.P.D.)
worked independently to extract data
from the selected studies. A detailed table
of data extraction can be seen in Table 1.
Authors of studies were directly con-
tacted in the event of information not
being available. These emails were sub-
sequently followed up with a reminder
1.e3
 Systematic Review
1.e4 American Journal of Obstetrics & Gynecology MONTH 2023

TABLE 1
Summary table of included studies
Country;
follow-up Study design; Sample
Source period data source size (n)
Bhattacharya United Retrospective cohort
et al,25 2012 Kingdom; study; linkage of
24e26 y Aberdeen Maternity
and Neonatal
Databank, hospital
discharges, cancer
registrations and
death registrations,
record linkage
Blomstrand Sweden; Prospective cohort
et al,26 2022 44 y study; Population
Study of Women in
Gothenburg, self-
reported
Brown et al,27 United Case-control study;
2006 States; participants in The
1e3 y Stroke Prevention in
Young Women Study
(SPYW), self-
reported
Canoy et al,28 United Prospective cohort 1,105,568 GH
2016 Kingdom; study; participants in
mean, The Million Women
11.6 y Study, self-reported
Chuang Taiwan; Prospective cohort 1,338,334 PE
et al,29 2022 17 y study; Taiwan
National Health
Insurance Database,
record linkage
Outcome; RR (95% CI)
measure of Study
Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
34,854 GH Cerebrovascular Pregnancy with Year of birth, social Moderate GH: GH:
PE (in first disease; OR twins or other class, smoking risk of 1.21 (0.98e1.51) 1.28 (1.03e1.59)
pregnancy) multiples, women bias PE: PE:
with chronic 1.32 (0.90e1.94) 1.27 (0.87e1.87)
hypertension,
duplicate cases
1459 PE Any stroke, Those with a stroke Blood pressure, BMI, Low risk Ischemic stroke: Ischemic stroke:
ischemic stroke, before age of 60 y waist-to-hip ratio, of bias 1.46 (1.01e1.88) 1.31 (0.95e1.80)
hemorrhagic smoking, physical Hemorrhagic stroke: Hemorrhagic stroke:
stroke; HR activity, triglyceride 1.47 (0.68e3.20) 1.41 (0.61e3.27)
levels, education, teeth Any stroke: Any stroke:
number 1.35 (0.99e1.84) 1.30 (0.95e1.79)
682 PE Ischemic stroke; Women aged
14 Age, race, education, Moderate 1.59 (1.00e2.52) 1.38 (0.81e2.33)
OR or 45 y, pregnant parity, smoking, BMI, risk of
at time of stroke, diabetes mellitus, bias
those whose stroke elevated cholesterol,
occurred within 42 angina-MI, hypertension
d postpartum,
nulligravida women,
those with
incomplete data
First time History of heart Age, region, smoking, Low risk Cerebrovascular Cerebrovascular
cerebrovascular disease, stroke, or socioeconomic status, of bias disease: disease:
disease, cancer (except for exercise, alcohol intake, 1.18 (1.14e1.21) 1.23 (1.20e1.27)
ischemic stroke, nonmelanoma skin BMI, hormone Ischemic stroke: Ischemic stroke:
hemorrhagic cancer), nulliparity, replacement therapy 1.26 (1.20e1.32) 1.29 (1.23e1.35)
stroke; RR missing data on use, DM, Hemorrhagic stroke: Hemorrhagic stroke:
parity hypercholesterolemia, 1.04 (0.98e1.10) 1.14 (1.07e1.12)
baseline hypertension
Any stroke, Incomplete data, Age, delivery type, Low risk Any stroke: Any stroke:
ischemic stroke, history of stroke, gestation number, of bias 1.75 (1.42e2.14) 2.05 (1.67e2.52)
hemorrhagic age of delivery not hospital level, delivery Ischemic stroke: Ischemic stroke:
stroke; HR between 18 and season, living area, 1.68 (1.35e2.09) 1.98 (1.59e2.46)
45 y family income level, Hemorrhagic stroke: Hemorrhagic stroke:
chronic HTN, gestational 3.00 (1.91e4.71) 3.45 (2.18e5.47)

ajog.org
diabetes mellitus,
anemia, antepartum
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
 ajog.org
TABLE 1
Summary table of included studies (continued)
Country; Outcome; RR (95% CI)
follow-up Study design; Sample measure of Study
Source period data source size (n) Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
hemorrhage, and
postpartum hemorrhage
Garovic United Retrospective cohort 4782 but HDP (GH Stroke; HR Women with Network, race, family Moderate Not reported 1.86 (1.16e2.98)
et al,30 2010 States; study; Family Blood only 4726 and PE) prepregnancy history of CVD, smoking, risk of
4y Pressure Program analyzed chronic BMI, education, DM, bias
study, Self-reported because hypertension hypertension
of missing
data
Hannaford United Prospective cohort 23,000 PE Cerebrovascular History of oral Age, social class, Moderate Not reported 1.39 (0.89e2.16)
et al,31 1997 Kingdom; study; participants in disease; RR contraceptive use smoking risk of
up to 26 y the Royal College of bias
General Practitioners
Oral Contraception
Study, self-reported
Huang et al,32 Taiwan; Case-control study; 167,480 HDP (PE Stroke; HR HDP before index Age, area of residence, Low risk 2.70 (2.33e3.14) 2.13 (1.82e2.51)
2020 13 y Taiwan National and GH) date, stroke before urbanization level, of bias
Health Insurance tracking, age <12 y, annual income,
database, record male or gender hyperlipidemia, DM,
linkage unknown heart disease, HTN,
CKD, obesity, season
MONTH 2023 American Journal of Obstetrics & Gynecology

Hung et al,33 Taiwan; Prospective cohort 1,235,850 HDP Any stroke, History of stroke, Age, delivery type, Low risk Not reported HDP: any stroke
2022 17 y study; Taiwan GH ischemic stroke, delivery age not gestation number, of bias 1.71 (1.46e2.00)
National Health CH hemorrhagic between 18 and hospital level, delivery Ischemic stroke
Insurance database, PE stroke; HR 45 y season, living area, 1.60 (1.35e1.89)
record linkage Superimposed income level, gestational Hemorrhagic stroke
PE on CH diabetes mellitus, 2.98 (2.13e4.18)
anemia, antepartum GH:
hemorrhage, Any stroke
postpartum hemorrhage 1.68 (1.13e2.52)

Systematic Review
Ischemic stroke
1.40 (0.88e2.21)
Hemorrhagic stroke
4.81 (2.55e9.10)
CH:
Any stroke
1.27 (0.97e1.68)
Ischemic stroke
1.21 (0.90e1.63)
Hemorrhagic stroke
1.e5

Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
 Systematic Review
1.e6 American Journal of Obstetrics & Gynecology MONTH 2023

TABLE 1
Summary table of included studies (continued)
Country; Outcome; RR (95% CI)
follow-up Study design; Sample measure of Study
Source period data source size (n) Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
1.62 (0.85e3.09)
PE:
Any stroke
2.00 (1.63e2.45)
Ischemic stroke
1.91 (1.53e2.37)
Hemorrhagic stroke
3.50 (2.31e5.29)
Superimposed PE:
Any stroke
3.86 (1.91e7.82)
Ischemic stroke
3.72 (1.75e9.23)
Hemorrhagic stroke
2.97 (0.42e21.13)
Kuo et al,34 Taiwan; Retrospective 6475 PE, eclampsia Any stroke, History of MI, CHF, Age, date of delivery Low risk Any stroke: Any stroke:
2018 median, cohort study; Taiwan hemorrhagic, peripheral vascular of bias 3.42 (1.44e8.12) 3.47 (1.46e8.23)
9.8 y National Health cerebral disease, CVA, DM, Ischemic: Ischemic:
Insurance Research infarction; HR dyslipidemia, 4.54 (1.13e18.18) 4.76 (1.19e19.03)
database, record hypertension Hemorrhagic: Hemorrhagic:
linkage 0.91 (0.11e7.76) 0.93 (0.11e7.97)
Lin et al,35 Taiwan; Retrospective 141,730 GH ICH, IRR History of gestational Age, date of delivery Low risk Not reported GH:
2016 13 y cohort study; Taiwan PE DM or ICH, age of of bias 3.72 (3.63e3.81)
National Health <20 or >50 y PE:
Research database, 8.21 (8.12e8.31)
record linkage
Langlois Canada; Retrospective 165,558 PE Cerebrovascular Ages <16 y or >50 Model 1: neighborhood Low risk 2.26 (1.92e2.65) Model 1:
et al,36 2020 median, cohort study, disease; HR y, delivery <20 wk, income quintile, of bias 1.55 (1.28e2.88)
16 y administrative health women diagnosed residence, type 1 or type Model 2:
data sets linked with any cardiac, 2 nongestational DM, 1.16 (0.97e1.38)
using unique cerebrovascular, or chronic hypertension,
encoded identifiers, peripheral arterial renal disease, illicit drug
record linkage disease
5 y before or tobacco use,
time zero, non- dyslipidemia.
Ontario residents, Model 2: further
and those without adjusted for time-
Ontario Health varying DM, chronic
Insurance plan hypertension, renal

ajog.org
numbers disease, illicit drug or
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
 ajog.org
TABLE 1
Summary table of included studies (continued)
Country; Outcome; RR (95% CI)
follow-up Study design; Sample measure of Study
Source period data source size (n) Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
tobacco use, and
dyslipidemia
Leon et al,37 United Retrospective cohort 1,303,365 Any HDP Ischemic stroke, Not reported Maternal ethnicity, Low risk HDP: PE:
2019 Kingdom; study; linked routine PE intracerebral maternal age, of bias Ischemic Ischemic stroke
median, electronic health hemorrhage, prepregnancy diabetes 2.25 (1.85e2.73) 2 (1.48e2.7)
9.25 y records from subarachnoid mellitus, prepregnancy Hemorrhagic ICH
CALIBER hemorrhage, hypertension, index of 1.85 (1.34e2.55) 1.52 (0.86e2.68)
(cardiovascular stroke not multiple deprivation, All stroke: Stroke not otherwise
research using otherwise cluster term to account 2.25 (1.97e2.57) specified
linked bespoke specified, for correlation within Term PE: 2.89 (1.92e4.34)
studies and all stroke; HR patients Ischemic All stroke
electronic health 2.71 (2.03e3.62) 1.9 (1.53e2.35)
records) resource, Hemorrhagic HDP:
record linkage 1.70 (0.98e2.96) Ischemic stroke
All stroke 1.72 (1.39e2.12)
2.4 (1.94e2.95) ICH
1.71 (1.24e2.36)
Stroke not otherwise
specified,
1.99 (1.46e2.71)
All stroke
1.83 (1.59e2.1)
MONTH 2023 American Journal of Obstetrics & Gynecology

Lin et al,38 Taiwan; up Retrospective 1,132,064 PE and Stroke; HR Extreme values of Age, years of Low risk 21 (2.5e174.0) 14.5 (1.3e165.1)
2011 to 5 y (1999 cohort study; linkage eclampsia maternal age, education, marital of bias
e2004) between birth infant birthweight, status, multiple
registries (1999 gestational weeks, gestations, infant sex,
e2003), NHI hospital or parity history of birthweight, parity, long-
discharge data, major adverse term HTN, pregnancy-
(1996e2004) and cardiovascular related HTN, anemia,
death files (1996 event diabetes mellitus,
e2004), record

Systematic Review
antepartum
linkage hemorrhage,
postpartum
hemorrhage, and SLE
Lykke et al,39 Denmark; Retrospective 782,287 All HDP First diagnosis Age <15 or >50 y Preterm delivery, SGA, Low risk Cohort 1: Cohort 1:
2009 median, 14.6 cohort study; Mild PE of stroke; HR placental abruption, of bias GH GH
y (cohort 1), National Patient Severe PE stillbirth, type 2 DM 1.68 (1.42e1.97) 1.51 (1.26e1.81)
12.9 y Registry and Central (HELLP) PE Cohort 2:
(cohort 2) Persons Registry and 1.60 (1.47e1.73) PE in 1st pregnancy
Cause of Death Cohort 2: 1.24 (1.10e1.40)
1.e7

Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
 Systematic Review
1.e8 American Journal of Obstetrics & Gynecology MONTH 2023

TABLE 1
Summary table of included studies (continued)
Country; Outcome; RR (95% CI)
follow-up Study design; Sample measure of Study
Source period data source size (n) Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
Registry, record PE in 1st pregnancy PE in second
linkage 1.23 (1.17e1.50) pregnancy
PE in second 1.65 (1.40e1.95)
pregnancy PE in both
1.86 (1.58e2.19) 1.48 (1.16e1.87)
PE in both
1.80 (1.42e2.27)
Männistö Finland; Retrospective cohort 10,314 GH Ischemic Events within 1st Prepregnancy BMI, Low risk GH: GH:
et al,40 2013 average, study; National PE cerebrovascular year, missing bp smoking before of bias 1.93 (1.50e2.48) 1.59 (1.24e2.04)
39.4 y Finland Birth Cohort CH disease; HR data, previous CVA pregnancy, parity, DM PE: PE:
eeroutine Superimposed before or during 1.28 (0.74e2.22) 1.19 (0.68e2.09)
collection by PE on CH pregnancy, CH: CH:
midwives at socioeconomic status 3.08 (2.39e3.97) 1.80 (1.39e2.34)
communal maternity Superimposed PE Superimposed PE on
welfare clinics, on CH: CH:
clinical 2.40 (1.31e4.42) 1.51 (0.83e2.76)
measurements
Miller et al,41 United Prospective cohort 83,749 HDP (GH, PE, All stroke before Older than 60 y at Age, race and ethnicity, Moderate Overall: Overall:
2019 States; study; women in and eclampsia) age 60 y; HR time of enrolment, smoking, migraine, risk of 1.3 (1.1e1.6) 1.3 (1.2e1.4)
mean, 18.1 y California Teachers history of stroke at obesity, DM, bias Before age 60 y: Before age 60 y:
(history of Study, self-reported time of enrolment hypertension 1.5 (1.1e2.1) 1.2 (0.9e1.7)
HDP), 18.2 y
(no history
of HDP)
Nelander Sweden; Retrospective cohort 3232 HDP All stroke; HR Male sex, BMI, education, Moderate 1.35 (0.99e1.79) 1.36 (1.00e1.81)
et al,42 2016 average, study; Swedish Twin nulliparous,
65 y smoking risk of
10.4 y Register, National old, not participating bias
Patient Register and in TELE test, minor
Cause of Death impairment in
Register, record cognitive function,
linkage dementia diagnosis
within 6 mo of
interview,
unavailable self-
reported HDP
Park et al,43 Republic of Retrospective 1,075,061 PE CVA in first 12 CVA before Age, primiparity, Moderate 1.71 (1.42e2.04) 1.64 (1.37e1.98)
2018 Korea; cohort study; mo; OR pregnancy, women multiple pregnancy, risk of
1 year Korea National with missing data, cesarean delivery, bias

ajog.org
Health Insurance women with 1 or induction, vacuum
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
 ajog.org
TABLE 1
Summary table of included studies (continued)
Country; Outcome; RR (95% CI)
follow-up Study design; Sample measure of Study
Source period data source size (n) Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
claims database, more of the delivery, DM, placenta
record linkage components of the previa, placental
Charlson abruption, peripartum
comorbidity index hysterectomy, uterine
arterial embolization,
postpartum
hemorrhage,
thromboembolism
Savitz et al,44 United Retrospective 849,639 PE ICH stroke or CVD, hypertension, Year of birth, maternal Low risk Not reported GH:
2014 States; cohort study; birth GH TIA; OR or diabetes mellitus age, maternal race and of bias ICH n/a
1y certificates and before delivery, ethnicity, health Stroke or TIA
hospital discharge nonsingleton births, insurance, gestational 1.2 (3 cases)
data, record linkage stillbirths, births DM, parity, maternal PE:
outside 1995e2004, education, prenatal ICH
those with missing smoking, prenatal care, 2.8 (1.3e6.2)
covariate data prepregnancy weight Stroke or TIA
2.8 (1.6e5.0)
Schokker The Case-control study; 113 HDP Stroke or Cases: Smoking, White race, Moderate 4.8 (2.0e11) 4.2 (1.6e11)
et al,45 2015 Netherlands; Stroke Database of TIA; OR Stroke owing to rare hypercholesterolemia, risk of
2008e2011 the department of cause, age >55 y DM, HTN bias
Neurology of the Controls:
MONTH 2023 American Journal of Obstetrics & Gynecology

Academic Medical Stroke, MI, kidney

Centre, record failure, deceased
linkage
Tang et al,46 Taiwan; Retrospective 1,132,019 PE Stroke, Extreme values of Age, years of education, Low risk 6-months 6-months
2009 1.25 y cohort study; birth ischemic stroke, maternal age, infant marital status, multiple of bias postpartum: postpartum:
certificates and hemorrhagic birthweight, gestation, birthweight, Ischemic stroke Ischemic stroke
National Health stroke; RR gestational age, or parity, anemia, DM, 14.18 (4.30e46.74) 11.60 (3.30e40.82)
Insurance hospital parity, history of cesarean delivery, CH, Hemorrhagic stroke Hemorrhagic stroke
discharge data, stroke pregnancy-related HTN, 13.34 (4.79e37.64) 11.76 (4.05e34.11)

Systematic Review
record linkage antepartum 12-months 12-months
hemorrhage, postpartum: postpartum:
postpartum hemorrhage Ischemic stroke Ischemic stroke
3.99 (0.55e29.18) 4.35 (0.58e32.92)
Hemorrhagic stroke Hemorrhagic stroke
24.70 (10.30 19.90 (7.75e51.11)
e59.21)
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
1.e9
 Systematic Review
1.e10 American Journal of Obstetrics & Gynecology MONTH 2023

TABLE 1
Summary table of included studies (continued)
Country; Outcome; RR (95% CI)
follow-up Study design; Sample measure of Study
Source period data source size (n) Exposures effect Exclusions Confounders adjusted quality Crude Adjusted
Tooher Australia; Retrospective 31,656 PE Stroke; OR Incomplete records Age, gestation, parity Low risk Not reported PE
et al,47 2017 median, cohort study; GH of bias 2.03 (0.75e5.49)
20 y medical records at a Any HDP (PE, GH
tertiary hospital and GH, CH, 0.57 (0.14e2.31)
admitted patient superimposed Any HDP
data collection in PE on CH) 1.94 (1.39e2.69)
New South Wales,
record linkage
Wilson et al,48 United Retrospective 3,593 GH Cerebrovascular Women with CH Age at delivery, social Moderate Hospital admission: Hospital admission:
2003 Kingdom; cohort study; linkage PE or disease; IRR class risk of GH GH
average, of Aberdeen eclampsia bias 1.42a 1.53 (0.72e3.27)
35.8 y maternity and PE PE
neonatal databank 2.00 a 2.10 (1.02e4.32)
with regional and Mortality: Mortality:
national registries, GH GH
record linkage 2.18 a 2.87 (0.81e10.2)
PE PE
2.44 a 3.59 (1.04e12.4)
BMI, body mass index; bp, blood pressure; CH, chronic hypertension; CHF, congestive heart failure; CKD, chronic kidney disease; CVA, cerebrovascular accident; CVD, cardiovascular disease; DM, diabetes mellitus; GH, gestational hypertension; HDP, hypertensive
disorders of pregnancy; HELLP, hemolysis, elevated liver enzymes and low platelet count; HR, hazard ratio; HTN, hypertension; ICH, intracranial hemorrhage; IRR, incidence rate ratio; MI, myocardial infarction; n/a, not appliable; NHI, national health insurance; OR, odds
ratio; PE, preeclampsia; RR, risk ratio; SGA, small for gestational age; SLE, systemic lupus erythematosus; TIA, transient ischemic attack.
a
95% confidence intervals not reported.
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023.

ajog.org
ajog.org
email if the authors failed to respond.
The Newcastle-Ottawa Quality Assess-
ment Scale was used to appraise each
study’s quality and to assess the risk of
bias.49 The scale used can be seen in
Appendix 3. If the 2 reviewers (M.P.B.
and F.P.D.) were unable to reach a
consensus on the data extraction or
quality appraisal, a third reviewer (L.K.)
was consulted for arbitration until a
consensus was reached. A score of 7 on
the Newcastle-Ottawa scale indicated a
study with a low risk of bias, a score of 4
to 6 indicated a moderate risk of bias,
and a score of
3 indicated a high risk of
bias. A summary of the quality appraisal
is shown in Appendix 4.
Statistical analysis
Statistical analysis was undertaken using
Review Manager software (version 5.4)
(Cochrane Collaboration) (Cochrane
Collaboration, London, UK). To calculate
the overall pooled estimates for associa-
tions between the exposures and outcomes
of interest, we used random effects meta-
analyses. Both crude and adjusted esti-
mates were inputted into Review Manager
using the generic inverse variance method.
We presented adjusted effect estimates
based on the definitions outlined in each
individual study.
We used forest plots to demonstrate
the pooled estimates with 95% CIs. By
using I2 values and s2 statistics, the de-
gree of heterogeneity was investigated for
each meta-analysis. Odds ratios and RRs
were used as approximations of each
other under the rare disease assump-
tion,50 and hazard ratios were treated as
an extension of this principle because the
outcomes were uncommon.51 This is in
line with previous literature.5,11 Pre-
liminary analyses were conducted to
observe the pooled associations between
each of the exposures (HDP, PE, GH,
CH, and superimposed PE) and each of
the outcomes (any stroke [ie, undiffer-
entiated], ischemic stroke, and hemor-
rhagic stroke). For the purposes of data
extraction and meta-analysis, cerebro-
vascular disease was taken to be equiva-
lent to the outcome of stroke. The term
any stroke denotes cases for which the
studies reported on an outcome that is
either undifferentiated or a composite
outcome of all reported stroke subtypes.
If the outcome was known to be caused
by a subarachnoid hemorrhage, it was
excluded from analysis.
Subgroup analyses were planned by
study type, study quality, ethnic group,
study location, number of pregnancies
affected by the specific exposure (ie, spe-
cific hypertensive disorder), and method
of measurement of the exposure and
outcome data (ie, self-reported vs medical
records vs laboratory measurements). The
subgroup analyses that were completed
were by study type, study location, method
of data ascertainment, and study quality.
The Cochrane Handbook suggests that
there should be a minimum of 10 studies
included when conducting a subgroup
analysis.52 Interactions during subgroup
analyses were investigated and reported
using P values, and a P value <.05 indi-
cated heterogeneity across subgroups.
Following the subgroup analysis, publica-
tion bias was assessed using a funnel plot
when there were 10 or more studies avail-
able. Egger test was completed using Stata
version 17.0 (StataCorp, College Station,
TX).53
A priori, we planned sensitivity analyses
to produce pooled estimates of studies that
adjusted for a minimal suite of con-
founders, specifically, age, body mass index
or obesity, smoking, and diabetes mellitus.
However, this was not possible because
there were too few studies available to
generate pooled estimates. Even when
revised to a minimal suite of 2 of these 4
potential confounders, there were too few
studies to conduct a meaningful sensitivity
analysis.
A post hoc meta-regression was con-
ducted to further investigate heteroge-
neity. We followed the Cochrane
Handbook guidelines, which suggest
that meta-regressions should only be
conducted for comparison groups with
10 or more studies.54 We conducted
univariate meta-regressions according to
study type, study quality, method of data
ascertainment, and study location.
Ethical considerations
Because a systematic review involves the
secondary analysis of published studies,
there was little ethical concern. The only
concerns pertained to permissions to
MONTH 2023 American Journal of Obstetrics & Gynecology
Systematic Review
access studies, transparency of research
methods, and publication of a research
protocol. In cases for which data were
not routinely available from published
manuscripts, this was addressed through
direct correspondence with the original
authors. Detailed search methods were
considered a priori and published in the
PROSPERO database under identifier
CRD42021254660.
Results
The initial search yielded 5174 results. A
total of 24 studies were included in the
final review, including 10,632,808 study
participants in total.25e48 Nine of these
studies examined more than 1
HDP.25,33,35,37,39,40,44,47,48 Nine studies
examined the composite exposure of any
HDP.30,32,33,37,39,41,42,45,47 Six of the
eligible studies were identified through
searching the reference lists of included
studies.
There was a wide range of sample sizes
among studies, varying from 113 in a
Dutch case-control study45 to 1,338,334
in a Taiwanese prospective cohort
study.29 The study location also varied.
Ten studies used European
data,25,26,28,31,37,39,40,42,45,48 5 used North
American data,27,30,36,41,44 8 used Asian
data,32e35,38,43,46 and 1 study used
Australian data.47 The most frequently
observed location was Taiwan with 7
studies that were conducted there.
Fifteen studies were found to be of high
quality with a low risk of
bias26,28,29,32e40,44,46,47 and 9 were found
to have a moderate risk of
bias.25,27,30,31,41e43,45,48 No studies had a
high risk of bias.
The characteristics of the studies are
detailed in Table 1. Table 2 shows a
summary table of the meta-analyses
conducted. The subgroup analyses are
shown in Table 3.
Hypertensive disorders of pregnancy
(undifferentiated)
Of the 24 studies, 9 studies reported an
effect estimate for the association be-
tween the composite exposure of any
HDP and the outcome of any
stroke.30,32,33,37,38,41,42,45,47 One study
was removed from analysis because it did
not report an adjusted effect estimate. It
1.e11
Systematic Review ajog.org

TABLE 2
Summary of main meta-analyses
Number of Number of 95% prediction
Outcome studies Reference numbers participants Pooled RR (95% CI) I2 (%) s2 interval
Exposure: hypertensive disorders of pregnancya
Any strokeb
Crude 5c Huang et al,32 2020, Leon et al,37 2019, 1,557,939 2.00 (1.42e2.82) 93 0.12 0.72e5.55
Miller et al,41 2019, Nelander et al,42 2016,
Schokker et al,45 2015
Adjusted 8 Garovic et al,30 2010, Huang et al,32 2020, 2,830,171 1.74 (1.45e2.10) 85 0.05 0.99e3.05
Hung et al,33 2022, Leon et al,37 2019,
Miller et al,41 2019, Nelander et al,42 2016,
Schokker et al,45 2015, Tooher et al,47 2017
Ischemic stroke
Crude 1 Hung et al,33 2022 1,303,365 2.25 (1.85e2.73) n/a n/a
Adjusted 2 Hung et al,33 2022, Leon et al,37 2019 2,539,215 1.65 (1.44e1.88) 0 0.00
Hemorrhagic stroke
Crude 1 Hung et al,33 2022 1,303,365 1.85 (1.34e2.55) n/a n/a
33 37
Adjusted 2 Hung et al, 2022, Leon et al, 2019 2,539,215 2.26 (1.32e3.87) 81 0.12
Exposure: preeclampsia
Any stroke
Crude 8c Bhattacharya et al,25 2012, Blomstrand 4,707,393 1.80 (1.55e2.10) 76 0.03 1.16e2.79
et al,26 2022, Chuang et al,29 2022, Kuo
et al,34 2018, Langlois et al,36 2020, Leon
et al,37 2019, Lykke et al,39 2009, Park
et al,43 2018
Adjusted 13 Bhattacharya et al,25 2012, Blomstrand 6,073,570 1.75 (1.56e1.97) 42 0.06 1.01e3.02
et al,26 2022, Chuang et al,29 2022, Garovic
et al,30 2010, Hannaford et al,31 1997, Hung
et al,33 2022, Kuo et al,34 2018, Langlois
et al,36 2020, Leon et al,37 2019, Park
et al,43 2018, Savitz et al,44 2014, Tooher
et al,47 2017, Wilson et al,48 2003
Ischemic stroke
Crude 6c Blomstrand et al,26 2022, Brown et al,27 2,660,629 1.80 (1.42e2.25) 56 0.04 1.03e3.16
2006, Chuang et al,29 2022, Kuo et al,34
2018, Leon et al,37 2019, Männistö et al,40
2013
Adjusted 7 Blomstrand et al,26 2022, Brown et al,27 3,896,479 1.74 (1.46e2.06) 41 0.02 1.18e2.56
2006, Chuang et al,29 2022, Hung et al,33
2022, Kuo et al,34 2018, Leon et al,37 2019,
Männistö et al,40 2013
Hemorrhagic stroke
Crude 4c Blomstrand et al,26 2022, Chuang et al,29 2,649,633 2.04 (1.34e3.09) 31 0.06 0.84e4.94
2022, Kuo et al,34 2018, Leon et al,37 2019
Adjusted 7 Blomstrand et al,26 2022, Chuang et al,29 4,876,852 2.77 (2.04e3.75) 64 0.08 1.30e5.90
2022, Hung et al,33 2022, Kuo et al,34 2018,
Lin et al,35 2016, Leon et al,37 2019, Savitz
et al,44 2014
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)

1.e12 American Journal of Obstetrics & Gynecology MONTH 2023

ajog.org Systematic Review

TABLE 2
Summary of main meta-analyses (continued)
Number of Number of 95% prediction
Outcome studies Reference numbers participants Pooled RR (95% CI) I2 (%) s2 interval
Exposure: gestational hypertension
Any stroke
Crude 2c Bhattacharya et al,25 2012, Canoy et al,28 1,140,422 1.18 (1.14e1.21) 0 0.00
2016
Adjusted 5 Bhattacharya et al,25 2012, Canoy et al,28 2,411,521 1.23 (1.20e1.26) 0 0.00 1.20e1.26
2016, Hung et al,33 2022, Tooher et al,47
2017, Wilson et al,48 2003
Ischemic stroke
Crude 2 Canoy et al,28 2016, Männistö et al,40 2013 1,115,882 1.53 (1.01e2.32) 91 0.08
28 33
Adjusted 3 Canoy et al, 2016, Hung et al, 2022, 2,351,732 1.35 (1.19e1.53) 26 0.00 1.19e1.53
Männistö et al,40 2013
Hemorrhagic stroke
Crude 1c Canoy et al,28 2016 1,105,568 1.04 (0.98e1.10) n/a n/a
28 33
Adjusted 3 Canoy et al, 2016, Hung et al, 2022, 2,483,148 2.66 (1.02e6.98) 100 0.69 0.07e107.74
Lin et al,35 2016
Exposure: chronic hypertension
Any stroke
Crudec Not reported
Adjusted 1 Hung et al,33 2022 1,235,850 1.27 (0.97e1.68) n/a n/a
Ischemic stroke
Crude 1 Männistö et al,40 2013 10,314 3.08 (2.39e3.97) n/a n/a
33 40
Adjusted 2 Hung et al, 2022, Männistö et al, 2013 1,246,164 1.49 (1.01e2.19) 75 0.06
Hemorrhagic stroke
Crudec Not reported
Adjusted 1 Hung et al,33 2022 1,235,850 1.62 (0.85e3.09) n/a n/a
Exposure: superimposed preeclampsia on chronic hypertension
Any stroke
Crudec Not reported
Adjusted 1 Hung et al,33 2022 1,235,850 3.86 (1.91e7.82) n/a n/a
Ischemic stroke
Crude 1 Männistö et al,40 2013 10,314 2.40 (1.31e4.42) n/a n/a
33 40
Adjusted 2 Hung et al, 2022, Männistö et al, 2013 1,246,164 2.30 (0.95e5.55) 70 0.29
Hemorrhagic stroke
Crudec Not reported
Adjusted 1 Hung et al,33 2022 1,235,850 2.97 (0.42e21.13) n/a n/a
CI, confidence interval; RR, risk ratio.
a
Hypertensive disorders of pregnancy refer to an exposure that is undifferentiated. Cases for which the specific HDP was categorized were placed in their own groups and were not included as part of
this category; b Any stroke refers to an outcome that is undifferentiated; c Certain studies failed to report crude estimates.
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023.

MONTH 2023 American Journal of Obstetrics & Gynecology 1.e13

 Systematic Review
1.e14 American Journal of Obstetrics & Gynecology MONTH 2023

TABLE 3
Table of subgroup analyses
P value for
Reference Number of Pooled RR subgroup 95%
Subgroups Number of studies numbers participants (95% CI) I2 (%) s2 interaction prediction interval
PE and any strokea: study quality
Low risk of bias
8 Blomstrand et al,26 2022, Chuang et al,29 4,932,336 1.85 (1.59e2.15) 42 0.02 1.28e2.67
2022, Hung et al,33 2022, Kuo et al,34 2018,
Langlois et al,36 2020, Leon et al,37 2019,
Savitz et al,44 2014, Tooher et al,47 2017
Moderate risk of bias
5 Bhattacharya et al,25 2012, Garovic et al,30 1,141,234 1.59 (1.37e1.83) 0 0.00 .15 1.37e1.83
2010, Hannaford et al,31 1997, Park et al,43
2018, Wilson et al,48 2003
PE and any stroke: study location
North America
3 Garovic et al,30 2010, Hannaford et al,31 1,019,923 1.83 (1.34e2.45) 52 0.04 0.74e4.56
1997, Langlois et al,36 2020
Europe
5 Bhattacharya et al,25 2012, Blomstrand 1,366,271 1.72 (1.28e2.31) 62 0.06 0.82e3.61
et al,26 2022, Leon et al,37 2019, Savitz
et al,44 2014, Wilson et al,48 2003
Asia
4 Chuang et al,29 2022, Hung et al,33 2022, 3,655,720 1.91 (1.63e2.24) 42 0.01 1.34e2.73
Kuo et al,34 2018, Park et al,43 2018
Australia
1 Tooher et al,47 2017 31,656 2.03 (0.75e5.49) n/a n/a .35
PE and any stroke: data ascertainment method
Record linkage
10 Bhattacharya et al,25 2012, Chuang et al,29 5,944,385 1.83 (1.61e2.07) 42 0.00 1.61e2.07
2022, Hung et al,33 2022, Kuo et al,34 2018,
Langlois et al,36 2020, Leon et al,37 2019,
Park et al,43 2018, Savitz et al,44 2014,
Tooher et al,47 2017, Wilson et al,48 2003
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)

ajog.org
 ajog.org
TABLE 3
Table of subgroup analyses (continued)
P value for
Reference Number of Pooled RR subgroup 95%
Subgroups Number of studies numbers participants (95% CI) I2 (%) s2 interaction prediction interval
Self-reported
3 Blomstrand et al,26 2022, Garovic et al,30 29,185 1.43 (1.15e1.80) 0 0.00 .07 0.45e4.56
2010, Hannaford et al,31 1997
PE and any stroke: study type
Retrospective cohort study
9 Bhattacharya et al,25 2012, Garovic et al,30 3,474,927 1.75 (1.51e2.01) 32 0.01 1.33e2.30
2010, Kuo et al,34 2018, Langlois et al,36
2020, Leon et al,37 2019, Park et al,43 2018,
Savitz et al,44 2014, Tooher et al,47 2017,
Wilson et al,48 2003
Prospective cohort study
4 Blomstrand et al,26 2022, Chuang et al,29 2,598,643 1.73 (1.39e2.15) 62 0.03 .94 0.96e3.13
2022, Hannaford et al,31 1997, Hung et al,33
2022
HDPb and any stroke: study quality
Low risk of bias
4 Huang et al,32 2020, Hung et al,33 2022, 2,738,351 1.88 (1.70e2.10) 25 0.00 1.70e2.10
MONTH 2023 American Journal of Obstetrics & Gynecology

Leon et al,37 2019, Tooher et al,47 2017

Moderate risk of bias
4 Garovic et al,30 2010, Miller et al,41 2019, 91,820 1.53 (1.17e2.00) 61 0.04 .15 0.77e3.05
Nelander et al,42 2016, Schokker et al,45
2015
HDP and any stroke: study location
North America

Systematic Review
2 Garovic et al,30 2010, Miller et al,41 2019 88,475 1.44 (1.05e1.97) 53 0.03
Europe
3 Leon et al,37 2019, Nelander et al,42 2016, 1,306,710 1.78 (1.26e2.51) 67 0.06 0.59e5.40
Schokker et al,45 2015
Asia
2 Huang et al,32 2020, Hung et al,33 2022 1,403,330 1.91 (1.54e2.37) 73 0.02
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023. (continued)
1.e15
 Systematic Review
1.e16 American Journal of Obstetrics & Gynecology MONTH 2023

TABLE 3
Table of subgroup analyses (continued)
P value for
Reference Number of Pooled RR subgroup 95%
Subgroups Number of studies numbers participants (95% CI) I2 (%) s2 interaction prediction interval
Australia
1 Tooher et al,47 2017 31,656 1.94 (1.39e2.71) n/a n/a .49
HDP and any stroke: study type
Retrospective cohort study
4 Garovic et al,30 2010, Leon et al,37 2019, 1,342,979 1.76 (1.54e2.01) 11 0.00 1.54e2.01
Nelander et al,42 2016, Tooher et al,47 2017
Prospective cohort study
2 Hung et al,33 2022, Miller et al,41 2019 1,319,599 1.48 (1.13e1.93) 89 0.03
Case-control study
2 Huang et al,32 2020, Schokker et al,45 2015 83,862 2.51 (1.42e4.45) 46 0.11 .22
HDP and any stroke: data ascertainment method
Record linkage
6 Huang et al,32 2020, Hung et al,33 2022, 2,657,965 1.84 (1.61e2.11) 54 0.01 1.38e2.46
Leon et al,37 2019, Nelander et al,42 2016,
Schokker et al,45 2015, Tooher et al,47 2017
Self-reported
2 Garovic et al,30 2010, Miller et al,41 2019 88,475 1.44 (1.05e1.97) 53 0.03 .16
CI, confidence interval; HDP, hypertensive disorder of pregnancy; PE, preeclampsia; RR, risk ratio.
a
Any stroke refers to an outcome that is undifferentiated; b Hypertensive disorders of pregnancy refer to an exposure that is undifferentiated. Cases for which the specific HDP was categorized were placed in their own groups and were not included as part of this
category.
Brohan. Hypertensive disorders of pregnancy and long-term risk for stroke. Am J Obstet Gynecol 2023.

ajog.org
ajog.org
was deemed inappropriate to compare
this with the adjusted figures extracted
from other studies.38 The pooled effect
estimates for HDP and any stroke gave
an adjusted RR (aRR) of 1.74 (95% CI,
1.45e2.10).
Subgroup analyses.
Several subgroup analyses were con-
ducted. Subgroups were classified based
on study quality, study location, data
ascertainment method, and study type.
Subgroup analyses by ethnic group were
not possible because studies either failed
to report on ethnicity or they comprised
populations of mixed ethnicity.
The suite of confounders indicated in
the study protocol (age, body mass in-
dex, smoking status, and diabetes mel-
litus) were too restrictive to generate any
subgroup that satisfied this requirement.
Two studies examined the association
between HDP and ischemic stroke.33,37
The aRR pooled effect measure across
these studies was 1.65 (95% CI,
1.44e1.88). There was no observed
heterogeneity between these studies
(I2¼0).
Similarly, there were 2 studies that
reported on the association between
HDP and hemorrhagic stroke.33,37 The
aRR pooled effect estimate was 2.26
(95% CI, 1.32e3.87). There was mod-
erate heterogeneity observed (I2¼81).
Preeclampsia
A total of 14 studies reported associa-
tions between PE and any
stroke.25,26,29e31,33e37,43,44,47,48 One of
these studies was removed from the
analysis because it was a clear outlier
with a disproportionate effect on the
pooled estimates.35 It is plausible that
this outlier could have been a conse-
quence of the time frame of the study
because it only followed participants for
a short duration after pregnancy, and the
association between PE and any stroke
may be stronger in the immediate post-
partum period than in later life. Before
removal of this study, the aRR pooled
effect measure generated was 2.05 (95%
CI, 1.63e2.57). The remaining 13
studies generated an aRR pooled effect
measure of 1.75 (95% CI, 1.56e1.97).
These 13 studies had relatively low het-
erogeneity (I2¼42). Seven studies
examined the association between PE
and ischemic stroke.26,27,29,33,34,37,40 An
aRR pooled effect estimate of 1.74 (95%
CI, 1.46e2.06) was observed. Seven
studies discussed the association
between PE and hemorrhagic
stroke.26,29,33,34,35,37,44 The aRR pooled
effect estimate for this association was
2.77 (95% CI, 2.04e3.75). When
compared with the meta-analysis con-
ducted for PE and ischemic stroke, the
meta-analysis for PE and hemorrhagic
stroke had a higher degree of heteroge-
neity with I2 values of 41 and 64,
respectively.
Gestational hypertension
Five studies examined the association
between GH and any stroke.25,28,29,47,48
The aRR pooled effect measure for
these studies was 1.23 (95% CI,
1.20e1.26).
Three studies examined the associa-
tion between GH and ischemic stroke
specifically.28,33,40 The aRR pooled effect
estimate was 1.35 (95% CI, 1.19e1.53).
Three studies reported on the associa-
tion between GH and hemorrhagic
stroke.28,33,35 The aRR pooled effect es-
timate was 2.66 (95% CI, 1.02e6.98).
However, the comparability of these 3
studies is limited. Maximal heterogene-
ity was observed between these studies
(I2¼100%) and the pooled effect esti-
mate had a wide 95% confidence
interval.
Chronic hypertension
There were 2 studies that examined CH
as an exposure variable.33,40 Both of
these studies examined the association
between CH and ischemic stroke,
generating an aRR pooled effect estimate
of 1.49 (95% CI, 1.01e2.19). One study
reported on the association between CH
and any stroke with an aRR of 1.27 (95%
CI, 0.97e1.68) and on CH and hemor-
rhagic stroke with an aRR of 1.62 (95%
CI, 0.85e3.09).33
Superimposed preeclampsia on
chronic hypertension
Two studies reported on superimposed
PE as an exposure variable.33,40 Both
studies examined the association be-
tween this exposure and the outcome of
MONTH 2023 American Journal of Obstetrics & Gynecology
Systematic Review
ischemic stroke. No significant pooled
effect estimate was generated (aRR, 2.30;
95% CI, 0.95e5.55). One study reported
on the association between super-
imposed PE and any stroke (aRR, 3.86;
95% CI, 1.91e7.82) and between
superimposed PE and hemorrhagic
stroke (aRR, 2.97; 95% CI,
0.42e21.13).33
Publication bias
Because evaluation of the association
between PE and any stroke included >10
studies, publication bias was assessed
using a Begg’s funnel plot (Appendix 7).
The funnel plot showed slight asymme-
try. However, the distribution of studies
on either side of the axis was relatively
even and diffuse. The result of Egger’s
test was not significant (P¼.20), indi-
cating that there was no small-study ef-
fect and suggesting that there was no
notable publication bias.
Meta-regression
A post hoc meta-regression was con-
ducted to further investigate heteroge-
neity. According to the Cochrane
Handbook, a meta-regression should
only be conducted on comparison
groups with 10 or more studies.54 We
followed this threshold and conducted
univariate meta-regressions according to
study type, study quality, method of data
ascertainment and study location. The
only comparison group that met the
threshold of 10 studies was that be-
tween PE and undifferentiated stroke.
The results of these analyses can be seen
in Appendix 8. The meta-regressions
demonstrated no statistically significant
difference between the groups. However,
given the small number of studies in each
subgroup, the difference between sub-
groups may be clinically relevant and
should not be overlooked.
Discussion
Principal findings
This systematic review and meta-
analysis aimed to synthesize the pub-
lished literature on the association be-
tween HDPs and stroke among parous
patient in later life. We included case-
control and cohort studies, used a vali-
dated quality assessment tool, and
1.e17
Systematic Review
included multiple reviewers to avoid
selection bias. Although the available
literature on this topic was sparse given
the eligibility criteria, the review pro-
duced several key findings.
First, HDPs seemed to be associated
with any stroke, ischemic stroke, and
hemorrhagic stroke in parous patients in
later life. Patients who have a history of
HDPs seem to be more likely to develop
stroke later in life than those without a
history of HDPs.
The observed associations may, how-
ever, differ by HDP subtype. PE and GH
may be positively associated with any
(undifferentiated) stroke, ischemic
stroke, and hemorrhagic stroke. PE is the
most prevalent HDP and the evidence
for its association with a future risk for
stroke was consistent across the litera-
ture. The association between GH and
hemorrhagic stroke is based on 3 studies
with maximal heterogeneity and, as
such, should be interpreted with
caution. CH seems to be associated with
ischemic stroke. The association be-
tween superimposed PE and ischemic
stroke is based on 2 studies and should
also be interpreted with caution.
The strength of the association also
seems to vary across HDP and stroke
subtype. Ischemic and hemorrhagic
strokes have a distinct epidemiology,
pathophysiology, risk factors, and prog-
noses. For example, hemorrhagic stroke is
associated with higher mortality and
morbidity than ischemic stroke.55
Improved understanding of the variations
in the strength of the associations with
stroke subtypes may be relevant for
informing future preventive interventions.
There was a relative lack of data on the
individual exposures of CH and super-
imposed PE on CH and their associa-
tions with certain outcomes, such as
undifferentiated stroke and hemorrhagic
stroke. A cohort study by Hung et al33
was the only study to investigate these
exposures in the context of their associ-
ation with undifferentiated stroke and
hemorrhagic stroke.
Comparison with existing literature
Most of the results of this study demon-
strate several findings that are in keeping
with existing literature. However, this
1.e18 American Journal of Obstetrics & Gynecology MONTH 2023
study also highlights some results that are
in contrast with existing literature.
First, as an example to demonstrate
the similarity between this study and the
existing literature, the significant posi-
tive association observed between PE
and stroke in this study is consistent with
existing literature. When compared with
another large systematic review, such as
that of McDonald et al15 with a pooled
aRR association of 2.03 (95% CI,
1.54e2.67), the pooled estimate
observed in our study is similar (aRR,
1.75; 95% CI, 1.56e1.97). The associa-
tion reported by McDonald et al15 has a
wider confidence interval, which may be
because it was based on 5 studies instead
of 13 studies as was the case in this study.
Such similarities can also be seen in the
literature when considering the associa-
tion between HDP and stroke. A sys-
tematic review by Wu et al16 found that
HDPs had a significant positive associa-
tion with stroke (aRR, 1.72; 95% CI,
1.50e1.97).16 This result is consistent
with the findings of our study (aRR, 1.74;
95% CI, 1.45e2.10).
Second, the results of this study are in
contrast with existing literature in some
regards. An example of this is seen in the
association between GH and stroke. In
this study, there is a significant positive
association between GH and stroke, 1.23
(1.20-1.26), however, this is not seen
consistently in existing literature. A large
systematic review and meta-analysis by
Lo et al17 reported a nonsignificant as-
sociation between GH and stroke (1.50;
95% CI, 0.75e2.99). Lo et al17 analyzed
3 studies in their meta-analysis, whereas
in this study, 5 studies were analyzed.
The studies identified and included for
analysis in each study also differed,
which may explain the difference in the
observed associations. This may also
point to a lack of consensus in the liter-
ature that was included in both cases.
There is extensive literature on HDPs
and stroke, however, the underlying
mechanisms behind these exposures
remain uncertain. As described in the
literature, it is possible that individuals
who experience HDPs have a predispo-
sition to develop stroke. Most patients
with HDPs experience a transient period
of hypertension confined to pregnancy,
ajog.org
except for those with CH. However, pa-
tients who experience HDPs are at
higher risk of developing hypertension
in later life,56 and hypertension is an
independent risk factor for stroke.57
Development of an HDP may indicate
an underlying predisposition to hyper-
tension and vascular disease. Control for
postpartum hypertension among studies
was poor. This is a key factor that may
confound or mediate the association
between HDP and stroke in later life.
More research that controls for this fac-
tor is warranted.
HDPs are thought to induce endo-
thelial changes that lead to vascular
dysfunction.58 PE has been hypothesized
to cause down-regulation of nitric oxide
in the maternal endothelium.59 Re-
ductions in the concentration of this key
vasodilator are understood to influence
vascular dysfunction among these pa-
tients in the long term. The maternal
vascular endothelium also seems to
remain sensitized to the vasoconstrictor
angiotensin II for a long period of
time.60 Less vasodilation and increased
vasoconstriction may encourage
increased turbulent flow, which in turn
may lead to an increased risk for vessel
damage or thrombus formation.
HDPs have been associated with other
long-term central nervous system disor-
ders. These include cognitive dysfunc-
tion,61 seizure disorders,62 white matter
changes,63 and dementia.64 It is thought
that these conditions are indicative of
changes in microvascular functionality
as a consequence of HDPs.65 It is plau-
sible that such changes may also have a
contributory role in the increased long-
term risk for maternal stroke.
Furthermore, although not a focus of
the current review, increasing experi-
mental evidence also suggests that HDPs
may lead to cerebrovascular sequelae
among the offspring of affected patients.
Rodent models of PE have shown
increased levels of proinflammatory cy-
tokines, chemokines, a higher incidence
of microbleeds, reduced microglial den-
sity,10 and impaired cerebral angiogen-
esis among the offspring66 as a
consequence of susceptibility to hypoxia-
induced injury.67 Moreover, reduced
blood flow to both the occipital and
ajog.org
parietal lobes in the brains of children
born to patients with PE68 suggests that
cerebral vascular alterations may be
present in those children, although the
clinical significance of such changes re-
mains uncertain.
Strengths and limitations
To the best of our knowledge, this is the
most comprehensive systematic review
and meta-analysis on HDP and the long-
term risk of stroke or stroke subtypes in
patients in later life. We used 3 databases
and supplemented the study selection
with manual searching of reference lists
of the selected studies. Most of the
studies included had large sample sizes,
were longitudinal by design, and had a
low or moderate risk of bias.
This study had some key strengths.
First, it was based on a published
research protocol. We followed a system
of established guidelines, specifically the
MOOSE guidelines and PRISMA-P
study selection guidelines. This study
also used the Newcastle-Ottawa Scale for
assessment of bias as an established and
validated quality appraisal tool. Data
extraction and quality appraisal took
place in a structured manner with in-
dependent data extraction and bias
assessment. There was a high level of
concordance between reviewers at both
the quality appraisal and data extraction
stages. A detailed series of subgroup
analyses were undertaken to provide an
understanding of the consistency of the
respective associations. Moreover, an
assessment of publication bias was con-
ducted, both graphically and statistically.
There were inherent limitations to
some of the included studies. One study
failed to report the number of study
participants, but instead reported the
number of births.44 Efforts to contact
authors for clarification were unsuc-
cessful. We used the number of births as
a proxy for study participants, but this
may not be entirely accurate, and it may
have impacted the reported effect esti-
mates. Furthermore, some studies did
not record the follow-up times for par-
ticipants. An approximation of missing
follow-up times was estimated from
survival curves or were given as a date
range.
The studies included in this systematic
review were determined to have a low or
moderate risk of bias. However, it is
likely that control for residual con-
founding was limited among these
studies. Consequently, it is likely that
residual confounding remains. This may
have influenced the measures of the as-
sociation that were observed.
Because of the low number of studies
included in some subgroups, not all
planned subgroup analyses could take
place. Subgroup analyses by ethnicity,
number of pregnancies, and according to
whether the study controlled for a min-
imum suite of confounders could not
take place.
Because the inclusion criteria required
that studies must be available in the
English language, this may have limited
the number of studies that were
included. This could have led to a biased
search strategy and may have limited the
studies that were accepted for review.
Two studies were not available in En-
glish. Upon translating these studies,
they did not meet the eligibility
criteria.69,70
This systematic review examined
studies from a wide range of ethnic
backgrounds. However, there was no
representation by studies conducted in
Africa or South America. This may have
limited the ethnic diversity of the studies
and, ultimately, the generalizability of
the results.
High levels of heterogeneity were
observed in a number of the meta-
analyses as assessed by the tau2 and I2
statistics. Despite using random effect
models, these measures should be
interpreted with caution and should not
be taken as a substitution for the
assessment of clinical and methodolog-
ical diversity. When possible outlying
effect estimates were excluded, however,
some clinical and methodological di-
versity may have persisted.
Another important limitation of
meta-analyses is the risk of introducing
bias to any pooled effects through the
original studies. A meta-analysis cannot
eliminate bias, improve precision, or
reduce the methodological errors of the
included studies.54 We endeavored to
overcome this limitation by using
MONTH 2023 American Journal of Obstetrics & Gynecology
Systematic Review
detailed eligibility criteria and by using a
standardized bias assessment tool for all
included studies. A subgroup analysis
was also conducted based on risk of bias
and demonstrated that there was no
statistically significant differences be-
tween studies with a low risk of bias and
those with a moderate or high risk of bias
as determined by use of the Newcastle-
Ottawa scale.
Finally, several of the associations
observed in this review were obtained by
pooling low numbers of studies.
Consequently, the pooled associations
should be interpreted with caution,
particularly when high levels of hetero-
geneity were observed in meta-analyses.
Conclusions and implications
HDPs are prevalent in the community.
Even if the individual relative risk of
stroke is low, because of the prevalence
of HDP in the community, the absolute
long-term risk of stroke following
exposure to HDP may be large. This
study suggests that patients with an HDP
may warrant systematic monitoring into
later life to limit the long-term impact of
HDP. Many clinical guidelines highlight
the importance of a patient’s obstetrical
history in gaining an understanding of
their risk stratification for cardiovascular
disease.71,72 This systematic review in-
dicates a potential need for the inclusion
of obstetrical factors in risk stratification
for stroke.
This systematic review investigated a
spectrum of associations across HDP
and stroke subtypes. It seems that the
strength of these separate associations
varies considerably. However, common
subtypes of HDP, such as PE and GH,
seem to be markers for future risk of
both ischemic and hemorrhagic stroke
even if their relative impact varies. In
absolute terms, any primary or second-
ary preventive measures that reduce the
impact of HDP may be effective in
reducing the high incidence of disease
outcomes in the population. Thus, for
cerebrovascular disease prevention, any
measures that reduce HDPs or their
long-term effects are likely to have a
greater impact on reducing the absolute
number of ischemic strokes than on
reducing the number of hemorrhagic
1.e19
Systematic Review
strokes. Nonetheless, there may be po- 3. Lyall F, Greer IA. The vascular endothelium in
tential benefits for prevention of both normal pregnancy and pre-eclampsia. Rev
Reprod 1996;1:107–16.
stroke subtypes. 4. Arnott C, Nelson M, Alfaro Ramirez MA, et al.
Furthermore, given the growing evi- Maternal cardiovascular risk after hypertensive
dence that HDPs are associated with disorder of pregnancy. Heart 2020;106:
cardiovascular and renal disease,5,37 in 1927–33.
addition to cerebrovascular disease, 5. Barrett PM, McCarthy FP, Kublickiene K,
et al. Adverse pregnancy outcomes and long-
HDPs should be considered a risk term maternal kidney disease: a systematic re-
marker for a wide spectrum of cardio- view and meta-analysis. JAMA Netw Open
metabolic disease outcomes in later life. 2020;3:e1920964.
Any efforts to reduce the burden of 6. Lederer M, Wong A, Diego D, Nguyen D,
HDPs will likely yield preventive benefits Verma U, Chaturvedi S. Tracking the develop-
ment of cerebrovascular risk factors following
for a range of other chronic disease pregnancy with preeclampsia. J Stroke Cere-
outcomes.73 brovasc Dis 2020;29:104720.
There is relatively little data available 7. Cunningham MW Jr, LaMarca B. Risk of
on the association of CH and super- cardiovascular disease, end-stage renal dis-
imposed PE with undifferentiated and ease, and stroke in postpartum women and their
fetuses after a hypertensive pregnancy. Am J
hemorrhagic stroke. High quality Physiol Regul Integr Comp Physiol 2018;315:
observational studies with a large pop- R521–8.
ulation and mixed demographics and 8. Younes ST, Ryan MJ. Pathophysiology of ce-
comorbidities are necessary to gain an rebral vascular dysfunction in pregnancy-induced
understanding of the true association hypertension. Curr Hypertens Rep 2019;21:52.
9. Harvey RE, Coffman KE, Miller VM. Women-
between these factors. specific factors to consider in risk, diagnosis and
Patients with a history of HDP may be treatment of cardiovascular disease. Womens
at a greater risk for CVA, ischemic stroke, Health (Lond) 2015;11:239–57.
and hemorrhagic stroke in later life than 10. Giambrone AB, Logue OC, Shao Q,
patients without a history of HDP. PE Bidwell GL, Warrington JP. Perinatal micro-
bleeds and neuroinflammation in E19 rat fe-
seems to be associated with maternal tuses exposed to utero-placental ischemia. Int J
stroke most consistently. Patients with a Mol Sci 2019;20:4051.
history of GH may also be at a greater risk 11. Dall’Asta A, D’Antonio F, Saccone G,
for any stroke, ischemic stroke, or hem- et al. Cardiovascular events following preg-
orrhagic stroke than patients without nancy complicated by pre-eclampsia with
emphasis on comparison between early- and
GH. There is a relative lack of data on the late-onset forms: systematic review and meta-
associations of CH and superimposed PE analysis. Ultrasound Obstet Gynecol 2021;57:
on CH with undifferentiated and hem- 698–709.
orrhagic stroke, and this warrants further 12. Grandi SM, Filion KB, Yoon S, et al. Car-
investigation. The mechanisms of this diovascular disease-related morbidity and mor-
tality in women with a history of pregnancy
increased risk for cerebrovascular disease complications. Circulation 2019;139:1069–79.
need further exploration in studies using 13. Brass LM. The impact of cerebrovascular
models that may reflect cerebrovascular disease. Diabetes Obes Metab 2000;2(Suppl2):
outcomes in patients and their children. S6–10.
However, most importantly, patients 14. Lakatta EG. Age-associated cardiovascular
changes in health: impact on cardiovascular dis-
who experience HDPs may warrant ease in older persons. Heart Fail Rev 2002;7:
closer monitoring for cerebrovascular 29–49.
disease in later life - 15. McDonald SD, Malinowski A, Zhou Q,
Yusuf S, Devereaux PJ. Cardiovascular
sequelae of preeclampsia/eclampsia: a sys-
tematic review and meta-analyses. Am Heart J
2008;156:918–30.
REFERENCES 16. Wu R, Wang TT, Gu RH, et al. Hypertensive
1. Lao TT. The mother–the long-term implica- disorders of pregnancy and risk of cardiovas-
tions on metabolic and cardiovascular compli- cular disease-related morbidity and mortality: a
cations. Best Pract Res Clin Obstet Gynaecol systematic review and meta-analysis. Cardiol-
2015;29:244–55. ogy 2020;145:633–47.
2. Ijomone OK, Osahon IR, Okoh COA, 17. Lo CCW, Lo ACQ, Leow SH, et al. Future
Akingbade GT, Ijomone OM. Neurovascular cardiovascular disease risk for women with gesta-
dysfunctions in hypertensive disorders of preg- tional hypertension: a systematic review and meta-
nancy. Metab Brain Dis 2021;36:1109–17. analysis. J Am Heart Assoc 2020;9:e013991.
1.e20 American Journal of Obstetrics & Gynecology MONTH 2023
ajog.org
18. Béjot Y, Bailly H, Durier J, Giroud M.
Epidemiology of stroke in Europe and trends for
the 21st century. Presse Med 2016;45:e391–8.
19. Kanyal N. The science of ischemic stroke:
pathophysiology & pharmacological treatment.
Int J Pharm Res Rev 2015;4:65–84.
20. Hajat C, Dundas R, Stewart JA, et al. Ce-
rebrovascular risk factors and stroke subtypes:
differences between ethnic groups. Stroke
2001;32:37–42.
21. O’donnell MJ, Xavier D, Liu L, et al. Risk
factors for ischaemic and intracerebral hae-
morrhagic stroke in 22 countries (the INTER-
STROKE study): a case-control study. Lancet
2010;376:112–23.
22. Montaño A, Hanley DF, Hemphill JC 3rd.
Hemorrhagic stroke. Handb Clin Neurol.
2021;176:229‒248
23. Brooke BS, Schwartz TA, Pawlik TM.
MOOSE reporting guidelines for meta-analyses
of observational studies. JAMA Surg
2021;156:787–8.
24. Page MJ, McKenzie JE, Bossuyt PM, et al.
The PRISMA 2020 statement: an updated
guideline for reporting systematic reviews. BMJ
2021;372:n71.
25. Bhattacharya S, Prescott GJ, Iversen L,
Campbell DM, Smith WC, Hannaford PC. Hy-
pertensive disorders of pregnancy and future
health and mortality: A record linkage study.
Pregnancy Hypertens 2012;2:1–7.
26. Blomstrand A, Blomstrand C, Hakeberg M,
Sundh V, Lissner L, Björkelund C. Forty-four-
year longitudinal study of stroke incidence and
risk factors - the Prospective Population Study
of Women in Gothenburg. Scand J Prim Health
Care 2022;40:139–47.
27. Brown DW, Dueker N, Jamieson DJ, et al.
Preeclampsia and the risk of ischemic stroke
among young women: results from the Stroke
Prevention in Young Women Study. Stroke
2006;37:1055–9.
28. Canoy D, Cairns BJ, Balkwill A, et al. Hy-
pertension in pregnancy and risk of coronary
heart disease and stroke: a prospective study in
a large UK cohort. Int J Cardiol 2016;222:
1012–8.
29. Chuang C-J, Chiou W-Y, Yang H-J, et al.
Impact of preeclampsia/eclampsia on hemor-
rhagic and ischemic stroke risk: a 17 years
follow-up nationwide cohort study. PLoS One
2022;17:e0276206.
30. Garovic VD, Bailey KR, Boerwinkle E, et al.
Hypertension in pregnancy as a risk factor for
cardiovascular disease later in life. J Hypertens
2010;28:826–33.
31. Hannaford P, Ferry S, Hirsch S. Cardiovas-
cular sequelae of toxaemia of pregnancy. Heart
1997;77:154–8.
32. Huang CC, Huang CC, Lin SY, et al. Asso-
ciation between hypertensive pregnancy disor-
ders and future risk of stroke in Taiwan: a
nationwide population-based retrospective
case-control study. BMC Pregnancy Childbirth
2020;20:217.
33. Hung SK, Lee MS, Lin HY, et al. Impact of
hypertensive disorders of pregnancy on the risk
ajog.org
of stroke stratified by subtypes and follow-up
time. Stroke 2022;53:338–44.
34. Kuo YL, Chan TF, Wu CY, Ker CR, Tu HP.
Preeclampsia-eclampsia and future cardiovas-
cular risk among women in Taiwan. Taiwan J
Obstet Gynecol 2018;57:364–9.
36. Langlois AWR, Park AL, Lentz EJM, Ray JG.
Preeclampsia brings the risk of premature car-
diovascular disease in women closer to that of
men. Can J Cardiol 2020;36:60–8.
37. Leon LJ, McCarthy FP, Direk K, et al. Pre-
eclampsia and cardiovascular disease in a large
UK pregnancy cohort of linked electronic health
records: a CALIBER study. Circulation
2019;140:1050–60.
35. Lin LT, Tsui KH, Cheng JT, et al. Increased
risk of intracranial hemorrhage in patients with
pregnancy-induced hypertension: a nationwide
population-based retrospective cohort study.
Medicine (Baltimore) 2016;95:e3732.
38. Lin YS, Tang CH, Yang CYC, et al. Effect of
pre-eclampsia-eclampsia on major cardiovas-
cular events among peripartum women in
Taiwan. Am J Cardiol 2011;107:325–30.
39. Lykke JA, Langhoff-Roos J, Sibai BM,
Funai EF, Triche EW, Paidas MJ. Hypertensive
pregnancy disorders and subsequent cardio-
vascular morbidity and type 2 diabetes mellitus
in the mother. Hypertension 2009;53:944–51.
40. Männistö T, Mendola P, Vääräsmäki M, et al.
Elevated blood pressure in pregnancy and
subsequent chronic disease risk. Circulation
2013;127:681–90.
41. Miller EC, Boehme AK, Chung NT, et al.
Aspirin reduces long-term stroke risk in women
with prior hypertensive disorders of pregnancy.
Neurology 2019;92:e305–16.
42. Nelander M, Cnattingius S, Åkerud H,
Wikström J, Pedersen NL, Wikström AK. Preg-
nancy hypertensive disease and risk of dementia
and cardiovascular disease in women aged 65
years or older: a cohort study. BMJ Open
2016;6:e009880.
43. Park Y, Cho GJ, Kim LY, Lee TS, Oh MJ,
Kim YH. Preeclampsia increases the incidence of
postpartum cerebrovascular disease in Korean
population. J Korean Med Sci 2018;33:e35.
44. Savitz DA, Danilack VA, Elston B,
Lipkind HS. Pregnancy-induced hypertension
and diabetes and the risk of cardiovascular
disease, stroke, and diabetes hospitalization in
the year following delivery. Am J Epidemiol
2014;180:41–4.
45. Schokker SA, Van Oostwaard MF,
Melman EM, et al. Cerebrovascular, cardiovas-
cular and renal hypertensive disease after hy-
pertensive disorders of pregnancy. Pregnancy
Hypertens 2015;5:287–93.
46. Tang CH, Wu CS, Lee TH, et al. Pre-
eclampsia-eclampsia and the risk of stroke
among peripartum in Taiwan. Stroke 2009;40:
1162–8.
47. Tooher J, Thornton C, Makris A, Ogle R,
Korda A, Hennessy A. All hypertensive disorders of
pregnancy increase the risk of future cardiovascular
disease. Hypertension 2017;70:798–803.
48. Wilson BJ, Watson MS, Prescott GJ, et al.
Hypertensive diseases of pregnancy and risk of
hypertension and stroke in later life: results from
cohort study. BMJ 2003;326:845.
49. Wells G, Shea B, O’connell D, et al. The
Newcastle-Ottawa Scale (NOS) for assessing the
quality of nonrandomized studies in meta-
analyses. Canada: Department of Epidemiology
and Community Medicine, University of Ottawa;
2020.
50. Greenland S, Thomas DC. On the need for
the rare disease assumption in case-control
studies. Am J Epidemiol 1982;116:547–53.
51. Symons MJ, Moore DT. Hazard rate ratio
and prospective epidemiological studies. J Clin
Epidemiol 2002;55:893–9.
52. Higgins JP, Thomas J, Chandler J, et al.
Cochrane handbook for systematic reviews of
interventions. Chichester, United Kingdom:
John Wiley & Sons; 2019.
53. StataCorp. Stata statistical software.
Release 16. College Station, TX: StataCorp
LLC; 2019.
54. Deeks JJ, Higgins JPT, Altman DG (editors).
Chapter 10: Analysing data and undertaking
meta-analyses. In: Higgins JPT, Thomas J,
Chandler J, Cumpston M, Li T, Page MJ, Welch
VA (editors). Cochrane Handbook for System-
atic Reviews of Interventions, version 6.3
(updated February 2022). Cochrane, 2022.
Available at www.training.cochrane.org/
handbook. Accessed January 2, 2023.
55. Sennfält S, Norrving B, Petersson J, Ullberg T.
Long-term survival and function after stroke: a
longitudinal observational study from the Swedish
Stroke Register. Stroke 2019;50:53–61.
56. Heida KY, Franx A, Van Rijn BB, et al. Earlier
age of onset of chronic hypertension and type 2
diabetes mellitus after a hypertensive disorder of
pregnancy or gestational diabetes mellitus. Hy-
pertension 2015;66:1116–22.
57. Boehme AK, Esenwa C, Elkind MS. Stroke
risk factors, genetics, and prevention. Circ Res
2017;120:472–95.
58. Grand’Maison S, Pilote L, Okano M,
Landry T, Dayan N. Markers of vascular
dysfunction after hypertensive disorders of
pregnancy: a systematic review and meta-
analysis. Hypertension 2016;68:1447–58.
59. Amaral LM, Cunningham MW Jr,
Cornelius DC, LaMarca B. Preeclampsia: long-
term consequences for vascular health. Vasc
Health Risk Manag 2015;11:403–15.
60. Granger JP, Alexander BT, Llinas MT,
Bennett WA, Khalil RA. Pathophysiology of
preeclampsia: linking placental ischemia/hyp-
oxia with microvascular dysfunction. Microcir-
culation 2002;9:147–60.
MONTH 2023 American Journal of Obstetrics & Gynecology
Systematic Review
61. Fields JA, Garovic VD, Mielke MM, et al.
Preeclampsia and cognitive impairment later
in life. Am J Obstet Gynecol 2017;217:74.
e1–11.
62. Nerenberg KA, Park AL, Vigod SN, et al.
Long-term risk of a seizure disorder after
eclampsia. Obstet Gynecol 2017;130:
1327–33.
63. Aukes AM, De Groot JC, Wiegman MJ,
Aarnoudse JG, Sanwikarja GS, Zeeman GG.
Long-term cerebral imaging after pre-
eclampsia. BJOG 2012;119:1117–22.
64. Andolf EG, Sydsjö GC, Bladh MK, Berg G,
Sharma S. Hypertensive disorders in pregnancy
and later dementia: a Swedish National Register
Study. Acta Obstet Gynecol Scand 2017;96:
464–71.
65. Coutinho T, Lamai O, Nerenberg K. Hyper-
tensive disorders of pregnancy and cardiovas-
cular diseases: current knowledge and future
directions. Curr Treat Options Cardiovasc Med
2018;20:56.
66. Lara E, Acurio J, Leon J, Penny J,
Torres-Vergara P, Escudero C. Are the
cognitive alterations present in children
born from preeclamptic pregnancies the
result of impaired angiogenesis? Focus on
the potential role of the VEGF family. Front
Physiol 2018;9.
67. Miller SL, Huppi PS, Mallard C. The conse-
quences of fetal growth restriction on brain
structure and neurodevelopmental outcome.
J Physiol 2016;594:807–23.
68. Rätsep MT, Paolozza A, Hickman AF, et al.
Brain structural and vascular anatomy is altered
in offspring of pre-eclamptic pregnancies: a pilot
study. AJNR Am J Neuroradiol 2016;37:
939–45.
69. 于凌, 于国平, 高波. 可逆性后部脑病综合
征: 临床和影像学分析. 现代生物医学进展.
2010(11):2095-6.
70. 유진구, 대형이. Hypertension in preg-
nancy. Journal of the Korean Medical Associa-
tion. 2016 Jan 1;59(1):24-30.
71. Mosca L, Appel LJ, Benjamin EJ, et al. Evi-
dence-based guidelines for cardiovascular dis-
ease prevention in women. Circulation
2004;109:672–93.
72. Whelton PK, Carey RM, Aronow WS, et al.
2017 ACC/AHA/AAPA/ABC/ACPM/AGS/
APhA/ASH/ASPC/NMA/PCNA guideline for the
prevention, detection, evaluation, and manage-
ment of high blood pressure in adults: a report of
the American College of Cardiology/American
Heart Association Task Force on Clinical Prac-
tice Guidelines. J Am Coll Cardiol 2018;71:
e127–248.
73. Barrett PM, Khashan AS, McCarthy FP,
Kublickiene K. Adverse pregnancy outcomes
and maternal health: action needed for long-
term benefit. Acta Obstet Gynecol Scand
2020;99:1107–9.
1.e21