JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 75, NO.

2, 2020
ª 2020 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

PUBLISHED BY ELSEVIER

Acute Stroke During Pregnancy

and Puerperium
a, b, c d
Islam Y. Elgendy, MD, * Mohamed M. Gad, MD, * Ahmed N. Mahmoud, MD, Ellen C. Keeley, MD, MS, Carl J. Pepine,
d
MD

ABSTRACT

BACKGROUND Acute stroke during pregnancy or within 6 weeks of childbirth is devastating for the mother and her
family, yet data regarding incidence and contemporary trends are very limited.

OBJECTIVES This study sought to investigate the incidence and outcomes of acute stroke and transient ischemic
attack during pregnancy or within 6 weeks of childbirth in a large database.

METHODS The National Inpatient Sample was queried to identify women age $18 years in the United States with
pregnancy-related hospitalizations from January 1, 2007, to September 30, 2015. Temporal trends in acute stroke
(ischemic and hemorrhagic)/transient ischemic attack incidence and in-hospital mortality were extracted.

RESULTS Among 37,360,772 pregnancy-related hospitalizations, 16,694 (0.045%) women had an acute stroke. The
rates of acute stroke did not change (42.8 per 100,000 hospitalizations in 2007 vs. 42.2 per 100,000 hospitalizations in
2015; ptrends ¼ 0.10). Among those with acute stroke, there were increases in prevalence of obesity, smoking, hyper-
lipidemia, migraine, and gestational hypertension. Importantly, in-hospital mortality rates were almost 385-fold higher
among those who had a stroke (42.1 per 1,000 pregnancy-related hospitalizations vs. 0.11 per 1,000 pregnancy-related
hospitalizations; p < 0.0001). The rates of in-hospital mortality among pregnant women with acute stroke decreased
(5.5% in 2007 vs. 2.7% in 2015; ptrends < 0.001).

CONCLUSIONS In this contemporary analysis of pregnancy-related hospitalizations, acute stroke occurred in 1 of every
2,222 hospitalizations, and these rates did not decrease over approximately 9 years. The prevalence of most stroke risk
factors has increased. Acute stroke during pregnancy and puerperium was associated with high maternal mortality,
although it appears to be trending downward. Future studies to better identify mechanisms and approaches to pre-
vention and management of acute stroke during pregnancy and puerperium are warranted.
(J Am Coll Cardiol 2020;75:180–90) © 2020 by the American College of Cardiology Foundation.

A cute stroke remains a major cause of disability and mortality

worldwide (1). In recent years, the incidence of stroke has
cardiovascular risk factors such as hypertension, dia-betes, and
obesity among younger adults (4), as well as the advancing
maternal age at the time of birth (5), may contribute to increase the
been rising among the younger population, including pregnant risk of acute stroke during pregnancy. A previous analysis
women (2). Acute stroke during pregnancy is an infrequent but suggested that the incidence of acute stroke during pregnancy and
potentially devastating event for both the mother and her family. puerperium has been slowly rising from 1994 to 2007 (6). However,
During preg-nancy, hemodynamic changes, the hypercoagulable there are few studies evaluating these trends and the prevalence of
state, and other factors that have yet to be identified likely contribute risk factors in more recent years. To address this important knowl-
to the increased risk of cardiovascu-lar events (3). The increasing edge gap, we evaluated trends in the incidence and
prevalence of traditional
Listen to this manuscript’s
audio summary by Editor-in-
Chief
Dr. Valentin Fuster on

JACC.org. From the aDivision of Cardiology Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts; bHeart and Vascular Institute,
Department of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, Ohio; cDivision of Cardiol-ogy, University of Washington, Seattle,
Washington; and the dDivision of Cardiovascular Medicine, Department of Medicine, University of Florida, Gainesville, Florida. *Drs Elgendy and
Gad contributed equally to this work and are co-first authors. The authors have reported that they have no relationships relevant to the contents of this
paper to disclose.

Manuscript received September 19, 2019; revised manuscript received October 17, 2019, accepted October 28, 2019.

ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2019.10.056

JACC VOL. 75, NO. 2, 2020 Elgendy et al. 181
JANUARY 21, 2020:180–90 Stroke During Pregnancy

outcomes of acute stroke during pregnancy and puer-perium, as well failure, valvular heart disease, atrial septal defects, ABBREVIATIONS

AND ACRONYMS
as the trends in the prevalence of risk factors for acute stroke, using obesity, smoking history, chronic kidney disease, prior
a large contemporary nationwide database. stroke/TIA, malignancy, benign tumors, obstructive sleep
ICD-9-CM = International
apnea, alcohol abuse, illicit drug abuse, depression, Classification of Diseases-9th
Revision-Clinical Modification
rheumatoid arthritis, anemia, migraine, pre-
SEE PAGE 191
eclampsia/eclampsia, gestational hyperten-sion, NIS = National Inpatient
Sample
gestational diabetes, and systemic lupus erythematosus
METHODS SLE = systemic lupus
[SLE]) were identified with the corresponding ICD-9-CM
erythematosus
codes. The
DATA SOURCE. The Nationwide Inpatient Sample (NIS) database TIA = transient ischemic attack
was queried for data detailing hospital admissions between January
2007 and September 2015. The NIS is made publicly available by hospital-related characteristics included bed size (small, medium,

the Agency for Healthcare Research and Quality for the Healthcare and large), location (urban vs. rural), hospital region (Northeast,
Cost and Utilization Project (7). The NIS represents the largest Midwest, South, and West), and teaching status.
publicly available all-payer database and contains discharge-level
administra-tive data on inpatient diagnoses and procedures from OUTCOMES MEASURED. The primary hypothesis that we tested
was that the incidence of acute stroke/TIA during pregnancy and
puerperium is increasing. The primary pre-specified outcome for
a stratified sample of approximately 20% of U.S. hospitals through this analysis was the incidence trend over time (years) for acute
2012. Starting from 2012, the NIS represents a sample of 20% of stroke during pregnancy and the puerperium. The secondary pre-
discharges from all hospitals. The NIS provides a weight variable specified outcomes included: 1) the trends of risk factors for acute
for establishing an estimate of national statistics. stroke during pregnancy; and 2) the rates and trends of in-hospital
mortality among women with acute stroke.
STUDY POPULATION. Women (age $18 years) who were
hospitalized during pregnancy, labor, and the post-partum period
due to pregnancy-related condi-tions (i.e., primary or secondary
STATISTICAL ANALYSIS. The patient baseline char-acteristics and
diagnoses) were identified by using the corresponding International
demographics as well as hospital-related outcomes were compared
Classification of Diseases-9th Revision-Clinical Modification (ICD-
between those who had a stroke or TIA and those who did not. Cat-
9-CM) diagnostic and procedure codes (Online Table 1). Stroke was
egorical variables were compared with the Mantel-Haenszel chi-
identified using ICD-9-CM diagnosis codes for hemorrhagic stroke
square test, and continuous variables were compared with analysis
(subarachnoid hemorrhage [430.xx] and intracerebral hemorrhage
of variance testing. To evaluate incidence and mortality trends (10),
[431.xx]), ischemic stroke (433.xx and 434.xx), and transient
the linear chi-square test was used, and the rates were expressed as
ischemic attack (TIA) (435.xx) (8,9). Studies have shown that
a percentage, per 1,000 hospitalizations, or per 100,000 pregnancy-
ischemic and hemor-rhagic stroke identification from administrative
related hospitalizations, as appropriate. The independent predictors
use of ICD-9-CM codes have a high specificity (99%) and positive
of acute stroke in pregnant women, as well as predictors of
predictive value (approximately 87%) (8). Because our analysis was
mortality in pregnant women with stroke, were examined with a
related to pregnant and post-partum women, we also used ICD-9-
hierarchical multivariable regression model to account for a
CM codes 674.0 and 997.02, which are specific codes for cere-
between-hospitals clustering effect. Variables included in the
brovascular disorders in puerperium and cesarean birth,
multivariable model were statistically significant on univariate
respectively. All available discharge diagnoses (i.e., primary or
analyses and were shown to affect outcomes based on previous
secondary) for stroke or TIA were included.
research. All statistical analyses were performed by using the
weighted values of observations as pro-vided by the NIS to measure
national estimates. Sta-tistical analyses were conducted using
RStudio software (RStudio, Boston, Massachusetts) or SPSS
software, version 25 (IBM SPSS Statistics, IBM, Armonk, New
York). A 2-sided value of p < 0.05 was set for statistical
PATIENT AND HOSPITAL CHARACTERISTICS. Base-line significance. Odds ratios and the 95% confidence intervals were
characteristics included demographics (age, race, length of hospital used to report the re-sults of the regression analysis.
stay, elective admission to the hospital, patient disposition, primary
payer informa-tion, and percentile of home income by residential
zip code) and medical comorbidities (e.g., hypertension,
hyperlipidemia, diabetes mellitus, coronary artery disease,
cardiomyopathy, atrial fibrillation, heart
182 Elgendy et al. JACC VOL. 75, NO. 2, 2020

Stroke During Pregnancy JANUARY 21, 2020:180–90

FIGURE 1 Study Flow Diagram

NIS hospitalizations from January

2007 through September 2015
N = 339,672,453

Male hospitalizations
N = 166,986,627

Female hospitalizations
N = 172,685,824

Nonpregnancy
hospitalizations
N = 135,325,052

Pregnancy hospitalizations
N = 37,360,772

No acute stroke/TIA Acute stroke/TIA

N = 37,344,078 N = 16,694

Flow diagram to identify how the study cohort was identified. NIS ¼ National Inpatient Sample; TIA ¼ transient ischemic attack.

RESULTS artery disease, atrial septal defect, migraine, and rheumatologic

diseases (i.e., rheumatoid arthritis and SLE). Pregnant women with
INCLUDED POPULATION. Among 37,360,772 hospi-talizations acute stroke/TIA also had
for pregnancy and puerperium from January 2007 through a higher proportion of pre-eclampsia/eclampsia but a lower
September 2015, 16,694 (0.045%) involved acute stroke/TIA: 7,872 proportion of gestational hypertension and gestational diabetes.
(47.2%) involved ischemic stroke/TIA, 5,169 (31.0%) involved Table 1 summarizes pertinent baseline patient and hospital-related
hemorrhagic stroke, and 3,652 (21.8%) involved un-specified stroke characteristics.
(Figure 1). In a secondary analysis in which diagnostic codes for On multivariable analysis, the following predictors were
TIA and pregnancy-specific codes were excluded, the incidence of independently associated with acute stroke during pregnancy:
acute stroke was 0.035% (13,041/37,360,772). The incidence of advancing maternal age, black race, prior history of stroke/TIA, pre-
acute stroke/TIA per 100,000 pregnancy-related hospitalizations eclampsia/ eclampsia, migraine, atrial septal defects, hyperlip-
increased with advancing maternal age (Figure 2). idemia, hypertension, diabetes mellitus, smoking, atrial fibrillation,
valvular heart disease, coronary artery disease, heart failure,
cardiomyopathy, malig-nancy, rheumatoid arthritis, SLE, anemia,
and depression.
Compared with those who did not have acute stroke/TIA, those
with acute stroke/TIA were older; were more likely to be black,
obese, and smokers; and had a higher proportion of hypertension, PRIMARY OUTCOME. During the study period, the incidence
diabetes mellitus, hyperlipidemia, atrial fibrillation, coronary
of acute stroke/TIA per 100,000 pregnancy-
related hospitalizations remained largely
unchanged
JACC VOL. 75, NO. 2, 2020 Elgendy et al. 183
JANUARY 21, 2020:180–90 Stroke During Pregnancy

FIGURE 2 Frequency of Acute Stroke/Transient Ischemic Attack per 100,000 Hospitalizations During Pregnancy and for Labor and Birth Stratified by Age Group

600 566.24
p = 0.001
500

400

300

200

100 49.64 62.5 92.1

30.52 36.96
0 ≥45
18-24 25-29 30-34 35-39 40-44
Age, Years
Acute Stroke per 100,000 Pregnancy-Related Hospitalizations

The incidence of acute stroke/TIA per 100,000 pregnancy-related hospitalizations increased with advancing maternal age. TIA ¼
transient ischemic attack.

(from 42.8 in 2007 to 49.5 in 2010, followed by a decrease to 42.2 prevalence of other traditional risk factors, such as hypertension and
in 2015; ptrends ¼ 0.10). There was an increase in the rates of acute diabetes mellitus, did not change (Table 2).
ischemic stroke/TIA (18.5 per 100,000 pregnancy-related
hospitalizations in 2007 vs. 22.3 per 100,000 pregnancy-related hos- A total of 703 patients (4.2%) with acute stroke/TIA died during
pitalizations in 2015; ptrends < 0.0001) and hemor-rhagic stroke the hospitalization. In-hospital mortality was almost 385-fold higher
(12.5 per 100,000 pregnancy-related hospitalizations in 2007 vs. among pregnant women with acute stroke/TIA versus those without
(42.1 per 1,000 pregnancy-related hospitalizations vs. 0.11 per
14.3 per 100,000 pregnancy-related hospitalizations in 2015; p trends
1,000 pregnancy-related hospitalizations, respec-tively; p < 0.0001).
< 0.0001). There was a decrease in the incidence of unspecified
stroke (11.38 per 100,000 pregnancy-related hospitalizations in The rates of in-hospital mortality among patients with acute
2007 vs. 5.6 per 100,000 pregnancy-related hospitalizations in 2015; stroke/TIA decreased during the study period (5.5% in 2007 vs.

ptrends < 0.0001) (Central Illustration, panel A). In the second-ary 2.7% in 2015; ptrends < 0.0001) (Central Illustration, panel B). On
analysis, in which TIA and the pregnancy-specific codes were multivariable analysis, the following predictors were independently
excluded, the incidence of acute stroke was increasing (29.8 per associated with in-hospital mortality among patients with acute
100,000 pregnancy-related hospitalizations in 2007 vs. 33.0 per stroke/TIA: age $40 years, black and Asian race, hemorrhagic
100,000 pregnancy-related hospitalizations in 2015; p trends < stroke (compared with ischemic stroke), anemia, heart failure,
0.0001) (Figure 3). cardiomyopathy, atrial fibrillation, hy-pertension, pre-
eclampsia/eclampsia, gestational diabetes, and cesarean delivery.

SECONDARY OUTCOMES. Among those with acute stroke/TIA,

there was an increase in the prevalence of the following risk factors: DISCUSSION
obesity, smoking, hyperlipidemia, migraine, atrial septal defects,
prior stroke, and gestational hypertension; however, the In this nationwide observational analysis of pregnancy-related
hospitalizations from 2007
184 Elgendy et al. JACC VOL. 75, NO. 2, 2020

Stroke During Pregnancy JANUARY 21, 2020:180–90

TIA has remained unchanged or might have increased during the

TABLE 1 Baseline Characteristics, Demographics, Comorbidities, and Pregnancy-
study period, as observed in the secondary analysis that used the
Related Complications
All Hospitalizations Stroke/TIA No Stroke/TIA most specific codes for acute stroke (i.e., by excluding TIA and the
(N ¼ 37,360,772) (n ¼ 16,694) (n ¼ 37,344,078) pregnancy-specific codes). There was also an increase in the
Age, yrs 28 (23–32) 30 (25–35) 28 (23–32) prevalence of the following risk factors among those with acute
Race stroke/TIA: obesity, smoking, hyperlipid-emia, migraine, atrial
White 52.4 45.1 52.4
septal defects, prior stroke, and gestational hypertension. Acute
Black 15.0 26.4 15.0
stroke/TIA was associated with high maternal mortality rates, but
Hispanic 21.7 18.3 21.7
Asian or Pacific Islander 5.3 4.1 5.3
the rates of in-hospital mortality with acute stroke/TIA have
Native American 0.8 0.7 0.8 decreased.
Other 4.8 5.2 4.8

Length of stay, days 2 (2–3) 4 (2–8) 2 (2–3)

Elective admission 47.5 20.4 47.5 Although the incidence of stroke has been decreasing in the
Control/ownership of hospital United States (11), some studies have shown that this decrease is
Government, nonfederal 13.4 15.6 13.4
mainly driven by a reduction in the stroke incidence among men but
Private, not for profit 72.2 74.2 72.2
not women (12). The findings from this study extended our
Private, investor owned 14.4 10.1 14.4
knowledge by showing that although acute stroke/TIA are rare
Bed size of hospital
Small 11.9 6.6 11.9 events during pregnancy and puerperium, the incidence has
Medium 27.2 21.9 27.2 remained unchanged or might be increasing. The incidence of acute
Large 60.9 71.5 60.9 stroke/ TIA during pregnancy and puerperium was slightly higher
Location/teaching status of hospital than that in a pooled analysis of 11 studies, which showed an
Rural 10.7 4.0 10.7
incidence of 0.03% (13). That meta-analysis was composed mainly
Urban nonteaching 37.7 24.7 37.7
of an analysis from the NIS database years 1994 through 2011 and
Urban teaching 51.6 71.3 51.6
was restricted to young women 25 to 34 years old, along with other
Hospital U.S. census region
Northeast 16.5 18.7 16.5 smaller single-center studies from several countries that enrolled
Midwest or North Central 21.2 20.4 21.2 patients from the 1990s until 2008 (13). Our more contemporary
South 38.3 40.4 38.3 analysis did not apply any age restrictions. Our analysis, as well as
West 24.0 20.4 24.0 other studies (14), showed that the cumulative inci-dence of acute
Patient disposition stroke increases with advancing maternal age, and this might
Routine 97.2 68.6 97.3
explain why the inci-dence of acute stroke/TIA was higher in our
Transfer to short-term hospital 0.4 7.0 0.4
study. Our estimates are also higher than that of a recent Canadian
Transfer to skilled nursing facility or 0.1 12.5 0.1
registry (13.4 per 100,000 births) (15); however, that study did not
intermediate care facility
Home health care 1.9 6.8 1.9 include TIA. Our esti-mates remained higher than that of this
Against medical advice 0.3 1.1 0.3 Canadian registry in our secondary analysis, which excluded TIA
Died 0.01 4.0 0.01 and pregnancy-specific codes for stroke. Furthermore, black
Primary expected payer women, who are known to be at a higher risk of stroke (16), are not
Medicare 0.9 4.9 0.9
as largely repre-sented in Canada as in the United States. Another
Medicaid 43.0 40.4 43.0
possible explanation for the relatively higher inci-dence of acute
Private insurance 49.9 47.0 49.9
Self-pay 3.2 4.0 3.2 stroke/TIA in our study compared with other studies (13,15) is our
No charge 0.2 0.3 0.2 inclusion of any stroke diagnosis (rather than only a primary diag-
Other 2.9 3.3 2.9 nosis of stroke). The increases in the trends in ischemic and
Continued on the next page hemorrhagic stroke noted in our study suggest that we must
understand the reasons behind this rise. We also noted that the
proportion of un-specified stroke has been decreasing, which likely
through 2015, we showed that acute stroke/TIA occurred in reflects improvement in coding of stroke etiology in
approximately 1 of every 2,222 hospitali-zations. Acute stroke/TIA
was independently associ-ated with advancing maternal age, black
race, prior history of stroke/TIA, pre-eclampsia/eclampsia,
migraine, atrial septal defects, hyperlipidemia, hy-pertension,
diabetes mellitus, atrial fibrillation, cor-onary artery disease,
valvular heart disease, heart failure, malignancy, rheumatoid
arthritis, SLE, ane-mia, and depression. The incidence of acute
stroke/
JACC VOL. 75, NO. 2, 2020 Elgendy et al. 185
JANUARY 21, 2020:180–90 Stroke During Pregnancy

more recent years rather a true decrease in the inci-dence of the

TABLE 1 Continued
events. Nevertheless, we showed that the incidence of acute stroke
All Hospitalizations Stroke/TIA No Stroke/TIA
is likely increasing in the secondary analysis by using the most
(N ¼ 37,360,772) (n ¼ 16,694) (n ¼ 37,344,078)
specific codes for acute stroke.
Median household income, percentile
<25 27.8 31.1 27.8
During pregnancy and puerperium, there is a state of 25–49 25.2 25.9 25.2
hypervolemia and increased venous stasis, associ-ated with an 50–74 24.6 22.3 24.6

increase in prothrombotic factors, that contribute to the increased $75 22.4 20.7 22.4

risk of ischemic stroke and cerebral vein thrombosis (3). Comorbidities

Hypertension 0.9 11.9 0.9
Hypertension re-mains the most common preventable risk factor for
Hyperlipidemia 0.2 6.1 0.2
stroke (ischemic or hemorrhagic) in the general pop-ulation (16), Diabetes mellitus 1.3 4.5 1.3
and efforts have been directed toward reducing the burden of Ischemic heart disease 0.1 2.7 0.1
hypertension along with other traditional risk factors (17). Our Cardiomyopathy 0.1 1.7 0.1
findings showed that hypertension is a prevalent risk factor and is Atrial fibrillation 0.04 1.4 0.04

inde-pendently associated with acute stroke/TIA. Furthermore, the Heart failure 0.1 4.6 0.1
Atrial septal defects 0.0 3.0 0.0
prevalence of hypertension did not decrease during the study period
Obesity 4.8 7.7 4.8
among pregnant women with stroke. We also found that the preva-
Smoking 2.1 7.0 2.1
lence of other traditional cardiovascular risk factors (i.e., obesity,
Chronic kidney disease 0.1 1.6 0.1
smoking, hyperlipidemia) has been increasing whereas the Malignancy 0.1 1.9 0.1
prevalence of diabetes melli-tus did not change. These findings Benign tumors 1.5 3.5 1.5
suggest that additional efforts should be directed toward reducing Obstructive sleep apnea 0.1 0.7 0.1
the burden of these risk factors among women in the childbearing Alcohol abuse 0.1 0.3 0.1

period. Notably, studies have shown that traditional risk factors are Illicit drug abuse 0.5 1.0 0.5
Depression 2.1 6.0 2.1
less prevalent among women with pregnancy-related stroke
Rheumatoid arthritis 0.1 0.4 0.1
compared with non–pregnancy-related stroke of the same group,
Anemia 11.3 22.4 11.2
which suggests that pregnancy-related stroke might have some Migraine 0.7 9.6 0.7
unique pathophysiologic mechanisms (18). Systemic lupus erythematosus 0.1 1.0 0.1
Valvular lesions 0.4 4.3 0.4
Prior stroke 0.1 4.0 0.1

Pregnancy complications
Gestational hypertension 3.4 2.8 3.4
Pre-eclampsia/eclampsia 4.4 19.3 4.4
Gestational diabetes 7.4 6.8 7.4
Besides the traditional cardiovascular risk factors, we identified
Cesarean delivery 30.0 17.9 30.0
other independent risk factors that were associated with acute
stroke/TIA. Consistent with the growing body of evidence linking
Values are median (interquartile range) or %.
migraine to the risk of ischemic and hemorrhagic stroke (19–21), TIA ¼ transient ischemic attack.
we found that migraine was associated with acute stroke/TIA during
pregnancy and puerperium. Migraine is not well distinguished from
pre-eclampsia in pregnancy in administrative data; thus, Furthermore, studies have shown that the ability of administrative
administrative data-bases might introduce misclassification for the databases to accurately diagnose atrial septal defects is limited (24).
diag-nosis of migraine in pregnant women (22). Atrial septal defects Thus, the strong asso-ciation between atrial septal defects and acute
(which includes patent foramen ovale) were associated with acute stroke observed in our study is likely an overestimation of the true
stroke/TIA. With the emergence of randomized controlled trial data effect. Atrial fibrillation and valvular heart disease (including
in recent years supporting the benefit of patent foramen ovale prosthetic valves) are known risk factors for cardioembolic
closure in patients with cryptogenic stroke (23), physicians are more ischemic events and hem-orrhagic stroke (as a complication of
likely to look for a patent fora-men ovale in younger patients with anticoagulation therapy) (25,26). Similar to previous studies, which
stroke. Hence, the association between acute stroke/TIA observed in showed that pre-eclampsia/eclampsia is associated with both
our study is likely due to the fact that physicians are unlikely to ischemic and hemorrhagic stroke (27,28), pre-eclampsia/eclampsia
search for an atrial septal defect/patent foramen ovale in patients was associated with acute stroke/TIA in our analysis. Although data
without stroke. suggest that the rates of severe pre-eclampsia have been increasing

in the United States (29), our data were

 186 Elgendy et al. JACC VOL. 75 , NO . 2, 2020
Stroke During Pregnancy JANUARY 21, 2 0 2 0:180–90

CENTRAL ILLUSTRATION Stroke During Pregnancy and Puerperium

Temporal Trends of Acute Stroke During Pregnancy and Puerperium

A
in the United States
Incidence of Stroke in Females Hospitalized During Pregnancy
Incidence per 100,000 Overall Incidence ptrend = 0.106

Hospitalization/Years 50

40

30

20

10

0
2007 2008 2009 2010 2011 2012 2013 2014 2015
Overall Stroke and Transient Ischemic Attack Hemorrhagic Stroke
Ischemic Stroke and Transient Ischemic Attack Unspecified Stroke Subtype

B Temporal Trends of Mortality Related to Acute Stroke During Pregnancy

in the United States
Mortality of Pregnant Females Diagnosed With Stroke During Hospitalization
ptrend < 0.001
6.00%

In-Hospital
Mortality
4.00%

2.00%

0.00%
2007 2008 2009 2010 2011 2012 2013 2014 2015
Year
Elgendy, I.Y. et al. J Am Coll Cardiol. 2020;75(2):180–90.

(A) Temporal trends in the incidence of acute stroke/transient ischemic attack complicating pregnancy and puerperium in the United States from 2007
through 2015. The incidence of acute stroke/transient ischemic attack per 100,000 pregnancy-related hospitalizations has remained largely
unchanged: there was an increase from 42.8 in 2007 to 49.5 in 2010, followed by a decrease to 42.2 in 2015 (ptrends ¼ 0.10). (B) Temporal trend of
pregnancy-related stroke mortality in the United States from 2007 through 2015. The rates of in-hospital mortality among patients with acute

stroke/transient ischemic attack decreased (5.5% in 2007 vs. 2.7% in 2015; ptrends < 0.0001).
JACC VOL. 75, NO. 2, 2020 Elgendy et al. 187
JANUARY 21, 2020:180–90 Stroke During Pregnancy

FIGURE 3 Temporal Trends in the Incidence of Acute Stroke Complicating Pregnancy and Puerperium in the United States From 2007 Through 2015 After
Excluding the ICD-9-CM Codes of Transient Ischemic Stroke and Pregnancy-Related Codes

Incidence of Stroke in Females Hospitalized During Pregnancy - Sensitivity Analysis

Incidence
Overall Incidence ptrend < 0.001
per 40
100,000
Hospitaliz
ation/Year
s 30

20

10

0
2007 2008 2009 2010 2011 2012 2013 2014 2015

The incidence of acute stroke has been increasing: 29.8 per 100,000 pregnancy-related hospitalizations in 2007 versus 33.0 per 100,000
pregnancy-related hospitalizations in 2015 (ptrends < 0.0001). ICD-9-CM, International Classification of Diseases-9th Revision-
Clinical Modification.

restricted to those women who had a stroke/TIA, which represents a severe pre-eclampsia. Nevertheless, early identifica-tion of pregnant
very small proportion of pregnant and postpartum women. In women who are at high risk of pre-eclampsia and the offering of
addition, we did not attempt to make a distinction between mild and preventive measures such as low-dose aspirin might help reduce the
risk of

TABLE 2 Trends of the Prevalence of Risk Factors and Comorbidities Among Pregnant Women With Acute Stroke

Risk Factor 2007 2008 2009 2010 2011 2012 2013 2014 2015 p Value

Hypertension 13.7 11.5 10.6 10.6 12.7 12.7 13.1 11.5 10.5 0.431
Hyperlipidemia 2.3 5.6 5.6 4.3 7.0 7.0 8.1 10.4 5.4 <0.0001
Diabetes mellitus 3.9 5.0 3.0 4.8 6.4 5.7 2.4 5.3 4.3 0.433
Coronary artery disease 3.0 2.0 2.7 3.2 3.1 1.9 2.6 2.5 3.1 0.983
Cardiomyopathy 1.9 1.5 1.9 1.2 1.2 1.6 3.1 1.4 1.6 0.445
Atrial fibrillation 1.5 0.5 2.2 0.5 2.2 0.8 2.1 1.7 1.6 0.072
Heart failure 4.7 4.7 7.3 3.9 3.8 4.1 5.8 3.9 5.1 0.319
Atrial septal defects 2.6 1.5 3.0 3.2 3.5 3.2 3.4 3.9 4.3 <0.0001
Obesity 3.6 3.7 5.8 8.0 7.8 8.1 11.0 11.2 11.7 <0.0001
Smoking 3.3 7.6 6.1 4.5 5.6 7.8 10.5 10.4 8.6 <0.0001
Chronic kidney disease 1.4 1.6 3.3 0.2 0.8 1.4 2.1 2.0 1.9 0.385
Rheumatoid arthritis 0.5 1.0 0.3 0.2 1.0 0.3 0.0 0.0 0.8 0.020
History of stroke 0.2 3.4 3.3 5.8 4.8 4.3 4.5 5.3 4.7 <0.0001
Migraine 6.2 7.7 6.2 6.9 12.7 9.7 13.9 10.4 15.6 <0.0001
Valvular disease 4.4 6.3 3.8 3.7 3.8 6.2 3.4 3.7 3.5 0.014
Systemic lupus erythematosus 0.7 1.7 0.8 1.0 1.0 0.8 0.8 1.4 0.8 0.729
Anemia 16.2 18.3 19.6 24.5 22.7 27.6 24.7 24.2 26.1 <0.0001
Gestational hypertension 1.7 1.5 3.2 1.4 4.0 2.2 3.1 4.5 4.7 <0.0001
Pre-eclampsia/eclampsia 21.7 18.0 18.9 19.6 19.1 21.6 20.2 13.2 21.8 0.066
Gestational diabetes 6.5 4.9 5.8 8.9 7.6 7.0 6.0 7.9 6.6 0.072

Values are %.
188 Elgendy et al. JACC VOL. 75 , NO . 2, 2020
Stroke During Pregnancy JANUARY 21, 2 0 2 0:180–90

acute stroke (30). Studies have also shown that the prevalence of codes and is thus subject to coding errors (i.e., mis-coding and
gestational diabetes is increasing (31); however, our analysis undercoding). Third, the NIS database lacks important clinical
showed that the prevalence of gestational diabetes among pregnant information such as the stroke severity and subtype, imaging data,
women with stroke/TIA did not change. We found that cesarean and data regarding medications during the hospital encounter.
birth was less likely associated with acute/stroke TIA. Some studies Fourth, we could not comment on the outcomes beyond the index
have shown that outcomes of vaginal and cesarean delivery are hospitalization because the NIS database is restricted to in-hospital
probably similar after stroke (3,32); however, we found that data only. Fifth, the NIS database relies on discharge rather than
cesarean birth was associated with increased mortality after acute admission diagnoses, so we could not determine whether stroke or
stroke/TIA, suggesting that sicker women with acute stroke/TIA TIA was the primary reason for hospital admission or developed
were likely offered cesarean delivery in our study. later during that admission. Sixth, the diagnosis codes for TIA and
the pregnancy-specific codes for stroke (ICD-9-CM codes 674.0 and
997.02) have not been validated in studies validating the ICD-9-CM
codes of stroke from the administrative database (8,40). Thus, we
In this investigation, the maternal mortality rate with acute performed a secondary analysis by excluding TIA and the
stroke/TIA was high (approximately 4.2%). Interestingly, we also pregnancy-specific codes. Furthermore, these vali-dation studies
observed that the rates of in-hospital mortality were decreasing were conducted by using Medicare data (8); thus, the positive
during the study period. These findings are consistent with the predictive value for these codes is likely to be lower in a younger
decrease in stroke-related mortality observed among the general population with a relatively lower prevalence of stroke. Finally,
population worldwide (33). Improve-ments in timely computed although we had information regarding maternal mortality rates,
tomographic imaging, thrombolytic therapy, and the recent newborn mortality data are not available. Despite these limitations,
introduction of mechanical thrombectomy contributed to the this study pro-vides important data regarding the trends and out-
improved outcomes among patients with stroke in recent years (34– comes of acute stroke/TIA by using a large, contemporary
36); however, the role of these therapies remains unclear among nationally representative sample of women during pregnancy and
pregnant women with ischemic stroke because this population has puerperium.
been excluded from randomized trials of these ther-apies. Data from
the Get With The Guidelines Stroke Registry suggest that
reperfusion therapy (defined as intravenous tissue plasminogen
activator, catheter-based thrombolysis, thrombectomy, or any
combination of these) was associated with similar favorable
outcomes and reperfusion rates among pregnant or postpartum
women compared with nonpregnant women (37). A recent
consensus docu-ment from Canada suggests that these reperfusion CONCLUSIONS
therapies could be offered to pregnant and post-partum women who
otherwise meet criteria (38). Future studies in this area remain In this large, contemporary, nationally representa-tive sample of
warranted. Consistent with the general population (39), we also pregnancy-related hospitalizations, acute stroke/TIA occurred in 1
found that hemorrhagic stroke was associated with higher odds of of every 2,222 hospi-talizations. The incidence of acute stroke/TIA
in-hospital mortality compared with ischemic stroke/TIA during during pregnancy and puerperium has remained unchanged or might
pregnancy and puerpe-rium, because hemorrhagic strokes are be increasing. Among those with acute stroke/TIA, the prevalence
usually more severe and tend to result in more extensive injury (39). of traditional cardiovas-cular risk factors and pregnancy-related
conditions such as pre-eclampsia/eclampsia increased or did not
change during the study period. Acute stroke during pregnancy and
puerperium was associated with high maternal mortality. Future
studies focusing on identification of mechanisms and novel
prevention and management strategies for acute stroke during
pregnancy and puerperium are warranted.

STUDY LIMITATIONS. The findings of this investiga-tion should be

interpreted in the context of potential limitations. First, this is an
observational, non-randomized design. Although we adjusted for
po-tential confounders, the risk of unmeasured confounding could ADDRESS FOR CORRESPONDENCE: Dr. Islam Y. Elgendy, 55
not be excluded. Second, the NIS is an administrative database Fruit St., GRB-800, Boston, Massachu-setts 02114. E-mail:
relying on ICD-9-CM
[email protected]. Twitter: @islamelgendy83.
JACC VOL. 75, NO. 2, 2020 Elgendy et al. 189
JANUARY 21, 2020:180–90 Stroke During Pregnancy

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: Acute TRANSLATIONAL OUTLOOK: Future studies should

stroke is rare during pregnancy and the puerperium but is focus on identifying the mechanisms responsible for
associated with high maternal mortality. The incidence acute stroke during pregnancy and the puerperium,
has changed little, if at all, over time, paralleling trends in characterizing women at risk, and developing effective
the prevalence of cardiovascular risk factors among methods for prevention.
pregnant women with acute stroke or transient ischemic
attack.

REFERENCES

1. Lozano R, Naghavi M, Foreman K, et al. Global Cincinnati/Northern Kentucky Stroke Study. identification and classification of adult congenital
and regional mortality from 235 causes of death Neurology 2017;89:990–6. heart disease. J Am Heart Assoc 2018;7(2):
for 20 age groups in 1990 and 2010: a systematic e007378.
13. Swartz RH, Cayley ML, Foley N, et al. The
analysis for the Global Burden of Disease Study
incidence of pregnancy-related stroke: a system- 25. Dagres N, Nieuwlaat R, Vardas PE, et al.
2010. Lancet 2012;380:2095–128.
atic review and meta-analysis. Int J Stroke 2017; Gender-related differences in presentation, treat-
2. Kissela BM, Khoury JC, Alwell K, et al. Age at 12:687–97. ment, and outcome of patients with atrial fibril-
stroke: temporal trends in stroke incidence in a lation in Europe: a report from the Euro Heart
14. Miller EC, Gatollari HJ, Too G, et al. Risk of
large, biracial population. Neurology 2012;79: Survey on Atrial Fibrillation. J Am Coll Cardiol
pregnancy-associated stroke across age groups in
1781–7. 2007;49:572–7.
New York State. JAMA Neurol 2016;73:1461–7.
3. Grear KE, Bushnell CD. Stroke and pregnancy: 26. Labaf A, Svensson PJ, Renlund H, Jeppsson A,
15. Liu S, Chan WS, Ray JG, Kramer MD, Joseph
clinical presentation, evaluation, treatment, and Själander A. Incidence and risk factors for throm-
KS. Stroke and cerebrovascular disease in
epidemiology. Clin Obstet Gynecol 2013;56: 350– boembolism and major bleeding in patients with
pregnancy: incidence, temporal trends, and risk
9. mechanical valve prosthesis: a nationwide population-
factors. Stroke 2019;50:13–20.
4. Yandrapalli S, Nabors C, Goyal A, Aronow based study. Am Heart J 2016;181:1–9.
16. Benjamin EJ, Muntner P, Alonso A, et al.
WS, Frishman WH. Modifiable risk factors in 27. Miller EC, Gatollari HJ, Too G, et al. Risk fac-
Heart disease and stroke statistics—2019 update:
young adults with first myocardial infarction. J Am tors for pregnancy-associated stroke in women
a report from the American Heart Association. Cir-
Coll Cardiol 2019;73:573–84. with preeclampsia. Stroke 2017;48:1752–9.
culation 2019;139:e56–528.
5. Mathews TJ, Hamilton BE. Mean age of 28. Bateman B, Schumacher H, Bushnell CD, et
17. Wright JS, Wall HK, Ritchey MD. Million Hearts
mothers is on the rise: United States, 2000-2014. al. Intracerebral hemorrhage in pregnancy. Fre-
2022: small steps are needed for cardiovascular
NCHS Data Brief 2016;(232):1–8. quency, risk factors, and outcome. Neurology
disease prevention. JAMA 2018;320:1857–8.
6. Kuklina EV, Tong X, Bansil P, George MG, 2006;67:424–9.
Callaghan WM. Trends in pregnancy hospitaliza- 18. Miller EC, Yaghi S, Boehme AK, Willey JZ,
29. Ananth CV, Keyes KM, Wapner RJ. Pre-
tions that included a stroke in the United States Elkind MS, Marshall RS. Mechanisms and out-
eclampsia rates in the United States, 1980-2010:
from 1994 to 2007: reasons for concern? Stroke comes of stroke during pregnancy and the post-
age-period-cohort analysis. BMJ 2013;347:f6564.
2011;42:2564–70. partum period: a cross-sectional study. Neurol
Clin Pract 2016;6:29–39. 30. Lavery JA, Friedman AM, Keyes KM, Wright
7. NIS, HCUP Nationwide Inpatient Sample.
JD, Ananth CV. Gestational diabetes in the
Healthcare Cost and Utilization Project (HCUP). 19. Mahmoud AN, Mentias A, Elgendy AY, et al.
United States: temporal changes in prevalence
Available at: https://www.hcup-us.ahrq.gov/ Migraine and the risk of cardiovascular and cere-
rates between 1979 and 2010. BJOG 2017;124:
nisoverview.jsp. Accessed April 20, 2019. brovascular events: a meta-analysis of 16 cohort
804 –13.
studies including 1 152 407 subjects. BMJ Open
8. Lakshminarayan K, Larson JC, Virnig B, et al.
2018;8:e020498. 31. LeFevre ML, US Preventive Services Task
Comparison of Medicare claims versus physician
Force. Low-dose aspirin use for the prevention of
adjudication for identifying stroke outcomes in the 20. Sacco S, Ornello R, Ripa P, Pistoia F, Carolei
morbidity and mortality from preeclampsia: US
Women’s Health Initiative. Stroke 2014;45:815–21. A. Migraine and hemorrhagic stroke: a meta-
Preventive Services Task Force recommen-
analysis. Stroke 2013;44:3032–8.
9. Elgendy IY, Omer MA, Kennedy KF, et al. 30- dation statement. Ann Intern Med 2014;161:
day readmissions after endovascular thrombec- 21. Elgendy IY, Nadeau SE, Bairey Merz CN, 819–26.
tomy for acute ischemic stroke. J Am Coll Cardiol Pepine CJ. Migraine headache: an under-appreci-
32. Tate J, Bushnell C. Pregnancy and stroke risk in
Intv 2018;11:2414–24. ated risk factor for cardiovascular disease in
women. Womens Health (Lond) 2011;7:363–74.
women. J Am Heart Assoc 2019;8:e014546.
10. Elgendy IY, Ha LD, Elbadawi A, et al.
33. Krishnamurthi RV, Moran AE, Feigin VL, et al.
Temporal trends in inpatient use of intravascular 22. Wabnitz A, Bushnell C. Migraine, cardiovascu-
Stroke prevalence, mortality and disability-adjusted
imaging among patients undergoing lar disease, and stroke during pregnancy: sys-
life years in adults aged 20-64 years in 1990-2013:
percutaneous coro-nary intervention in the United tematic review of the literature. Cephalalgia 2015;
data from the Global Burden of Disease 2013 Study.
States. J Am Coll Cardiol Intv 2018;11:913–5. 35:132–9.
Neuroepidemiology 2015;45:190–202.
11. Koton S, Schneider AL, Rosamond WD, et al. 23. Mojadidi MK, Zaman MO, Elgendy IY, et al.
34. Fonarow GC, Zhao X, Smith EE, et al. Door-
Stroke incidence and mortality trends in US com- Cryptogenic stroke and patent foramen ovale. J
to-needle times for tissue plasminogen activator
munities, 1987 to 2011. JAMA 2014;312:259–68. Am Coll Cardiol 2018;71:1035–43.
administration and clinical outcomes in acute
12. Madsen TE, Khoury J, Alwell K, et al. Sex- 24. Khan A, Ramsey K, Ballard C, et al. Limited ischemic stroke before and after a quality
specific stroke incidence over time in the Greater accuracy of administrative data for the improvement initiative. JAMA 2014;311:1632–40.
190 Elgendy et al. JACC VOL. 75, NO. 2, 2020

Stroke During Pregnancy JANUARY 21, 2020:180–90

35. Elgendy IY, Kumbhani DJ, Mahmoud A, Bhatt DL, in ischemic stroke in pregnancy or post-partum 40. Tirschwell DL, Longstreth WT Jr. Validating
Bavry AA. Mechanical thrombectomy for acute period. Am J Obstet Gynecol 2016; 214:723. administrative data in stroke research. Stroke
ischemic stroke: a meta-analysis of random- ized trials. 2002;33:2465–70.
J Am Coll Cardiol 2015;66:2498–505.
38. Ladhani NNN, Swartz RH, Foley N, et al. Ca-
36. Elgendy IY, Mahmoud AN, Mansoor H, nadian stroke best practice consensus statement:
Mojadidi MK, Bavry AA. Evolution of acute acute stroke management during pregnancy. Int J KEY WORDS hemorrhagic stroke, ischemic
ischemic stroke therapy from lysis to thrombec- Stroke 2018;13:743–58. stroke, pregnancy, puerperium
tomy: similar or different to acute myocardial
39. Andersen KK, Olsen TS, Dehlendorff C,
infarction? Int J Cardiol 2016;222:441–7.
Kammersgaard LP. Hemorrhagic and ischemic
37. Leffert LR, Clancy CR, Bateman BT, et al. strokes compared: stroke severity, mortality, and APPENDIX For a supplemental table, please see
Treatment patterns and short-term outcomes risk factors. Stroke 2009;40:2068–72. the online version of this paper.