The Impact of Sex and Gender on Stroke

Abstract

Women face a disproportionate burden of stroke mortality and disability. Biologic sex and sociocultural gender both contribute to differences in stroke risk factors, assessment, treatment and outcomes. There are substantial differences in the strength of association of stroke risk factors as well as female-specific risk factors. Moreover, there are differences in presentation, response to treatment and stroke outcomes in women. This review outlines current knowledge of impact of sex and gender on stroke, as well as delineates research gaps and areas for future inquiry.

Introduction

Women face a disproportionate burden of stroke mortality and disability. They have substantial differences in the strength of association of stroke risk factors as well as female-specific risk factors. Moreover, there are differences in presentation, response to treatment and stroke outcomes in women. We have undertaken this review to highlight current knowledge as well as research gaps and areas for future inquiry (see Central Illustration).

Central illustration:

The impact of sex and and gender on stroke (Biologic sex is shown in red, gender in blue, combined in purple)

IV rtPA intravenous recombinant tissue plasminogen activator. APO: adverse pregnancy outcomes

Sex/gender

Most of the published data on stroke has not explicitly separated sex and gender, and likely incorporates both influences. Sex refers to the biological characteristics of individuals including genetic, biologic and physiologic expression. In contrast, gender is a social construct that includes gender identify, expression, roles and stereotypes for female, male and gender diverse people.¹ Gender also relates to power, economic resources and healthcare access, all of which can also influence health. While neither sex nor gender are binary, most data collection in trials and cohorts has been binary, and sex and gender identity have not been collected separately. Additionally, quantification of gender traits is evolving and usually not available. Thus, many of the findings for “sex differences” may represent the combined impact of both sex and gender, and gendered exposures may contribute to observed differences between men and women.² Moreover, gender inter-relates to and intersects with race/ethnicity, sexuality, age, geography, and other attributes, so intersectional research approaches are needed. A few studies have looked the impact of gender-related characteristics on CHD outcomes³ and health status² in cis-gender women, but none are available for stroke to the best of our knowledge. Transgender data are also limited.^4–6 For the remainder of this review, we will use the term sex differences, but stress the limitations listed above.

Stroke epidemiology

In 2019, the most current year available, stroke was the second leading cause of death worldwide, as well as the second cause of disability-adjusted life years.⁷ In the US in 2019, stroke was the 3rd leading cause of death in women, compared with 5th in men.⁸ Women accounted for 57.1% of stroke deaths in 2019, with stroke accounting for 6.2% of all female deaths, while comprising 4.4% of all male deaths.⁸ In total, Approximately 55,000 more fatal strokes occur in women each year than men.⁹ In the 2015 Greater Cincinnati Northern Kentucky Stroke Study (GCNKSS), stroke case fatality in women exceeded that of men for the first time, even after adjustment for age.^{10, 11} As show in Figure 1, strokes comprise a greater percentage of deaths in women than men throughout the adult lifespan (Lichtman et al, unpublished data).

Figure 1.

Sex- and age-specific ranking, percentage, and total number of deaths attributed to cerebrovascular diseases in 2015

Data from: National Vital Statistics System (NVSS), 2015. LCWK1: Deaths, percent of total deaths, and death rates for the 15 leading causes of death in 5-year age groups, by race and sex: United States, 2015. https://www.cdc.gov/nchs/data/dvs/LCWK1_2015.pdf. Accessed October 13, 2021.

Globally, the lifetime risk of stroke (from age 25 onward) is 25.1% in women and 24.7% in men, but there is substantial regional variation.¹² There are regional differences with the highest lifetime risks for women in Eastern Europe and East Asia (36.5% and 36.3%, respectively).¹² In the US, the lifetime risk of stroke is higher in women (20–21%), than in men (14–17%), for a 55 year old individual.¹³ Stroke is more likely to be the first manifestation of cardiovascular disease in women, whereas in men coronary heart disease is more common.¹⁴ The social context for women who experience stroke is important as well. Age at onset of stroke is on average 4–6 years older in women than men^15,10 Moreover, women are more likely to be widowed, unmarried or living alone, and have a higher degree of disability in their activities of daily living than men at the time of their stroke.¹⁵

Sex is related to differences in stroke type incidence. Women have a higher prevalence and incidence of intracranial aneurysms, and substantially higher incidence of subarachnoid hemorrhage compared to men, while men have higher rates of intracerebral hemorrhage.^{7, 11} Across the lifespan, recent data from Canada shows that the risk ischemic stroke and transient ischemic attack (TIA) is higher for women than men <30 years, higher in men for midlife, and then equal above 80 years of age.⁷ In the oldest category (>85 years),^{7, 16} there is some suggestion that women have higher incidence of stroke than men.¹⁵ Significant disparities by race/ethnicity have been documented, particularly in older women. In the Northern Manhattan Study (NOMAS), Black and Hispanic women age 70 and older had a 76–77% higher risk of stroke than white women after adjusting for age, education and insurance status, a difference not seen in men.¹⁷ In data from the Women’s Health Initiative (WHI), Black women have a 47% higher risk of stroke than White women after adjustment for age, with much of the risk mediated by stroke risk factors and socioeconomic factors.¹⁸ Over the past several decades, stroke incidence has been declining, with relatively similar changes in both men and women. The Atherosclerosis Risk in Communities (ARIC) study documented substantial decline in stroke incidence among individuals age 65 year and older from 1987–2017, with similar declines in women and men.¹⁹ Similarly in the GCNKSS, incidence rates declined for both women and men from 1993/1994 to 2015.¹⁰

Impact of stroke

Given the aging of the US population and rises in life expectancy, in 2013, the American Heart Association and American Stroke Association projected that by 2030 the prevalence of stroke in the US would be 3.9%. Moreover, the economic burden of stroke was projected to rise by 129% from 2013 to 2030, with an increase in the direct economic burden of stroke-related medical costs, from $71.6 billion in 2013 to $184.1 billion in 2030.²⁰ Given the predominance of women in these older age groups, this impact will be disproportionately in women.

Risk factors for stroke

Since previously published guidelines and scoping reviews, additional data have advanced our knowledge of sex differences in key stroke risk factors.^21–23 See Figure 2 for an overview of risk factors.

Figure 2:

Sex differences in Stroke Risk Factors; size of figures indicates differences in strength of association between given risk factors and stroke; MHT: menopausal hormone therapy; GHT: gender-affirming hormone therapy

The lifetime risk of stroke is higher for women than men, with a 1 in 4 risk of stroke for women after age 25.¹² The change in stroke risk with age varies by sex; stroke incidence is higher in women than men for those less than 30 years,⁷ while rates are higher in men than women during mid-life and either equal by sex or higher among women beginning in the eighth decade.^{7, 10, 24} Data on female-to-male incidence rate ratios during the reproductive years conflict.²⁵ Further research is needed to determine contributors to the rising risk for women after menopause (and reversal of risk ratio relative to men) and whether factors affecting vascular dysfunction such as sex hormone levels and biologic age might be novel targets for stroke prevention. Recently published data conflict as to whether endogenous levels of age-varying sex hormones including sex hormone binding globulin, estradiol, and testosterone are independently associated with stroke.^{26, 27}

Key modifiable risk factors that appear to affect risk differentially by sex include diabetes and hypertension. Diabetes is more strongly associated with incident ischemic stroke in women compared with men, with a more pronounced difference for type 1 diabetes.^{23, 28, 29} Ischemic stroke risk also begins rising at lower fasting blood glucose levels for women than men, even after adjustment for treatment with oral medications or insulin.³⁰ More research is needed to understand the biologic mechanisms underlying these sex differences. For now, early and consistent screening for diabetes mellitus is warranted.³¹

For hypertension, although some previous data indicate no consistent sex difference in the association between increasing systolic blood pressure and stroke risk,¹⁴ recently published data from U.S. and international cohorts demonstrate a stronger association between hypertension and risk of incident total and ischemic stroke in women compared with men, adjusted for use of anti-hypertensives.^{28, 32, 33} Longitudinal data in the U.S. have revealed a faster rise in blood pressure for women than men starting in the third decade, supporting a sexual dimorphism in the contribution of blood pressure to cardiovascular disease and stroke later in life.³⁴ While data from recent large-scale observational studies indicates that stroke risk becomes significant at a lower threshold of systolic blood pressure for women than men (120 vs. 150 mm Hg, respectively), more prospective and ideally clinical trial data will be needed before sex-specific targets are integrated into guidelines.³⁵

Elevated body mass index (BMI), clinical obesity (BMI ≥30 kg/m²), and higher waist-to-hip ratio are associated with higher risk of total and ischemic stroke in both sexes, with a stronger association in women than men.²⁸ The relationship between obesity and intracerebral hemorrhage (ICH) is more complex, with some evidence for a protective effect in women, but studies also indicate this varies by sex, with data from the UK biobank demonstrating an inverse association between BMI and risk of ICH in women but a positive association in men.^{28, 36, 37}

Atrial fibrillation is a modifiable risk factor for both women and men with data demonstrating a higher risk of stroke and all-cause mortality in women with atrial fibrillation compared with men, and female sex has been incorporated as a risk factor in the CHADS2VASc2 score for decision making about anticoagulation.²³ Women older than 65 with atrial fibrillation are at an especially high risk for subsequent stroke events.^{38, 39} Sex and gender disparities in use of oral anticoagulants for stroke prevention have been previously reported despite data demonstrating similar effectiveness as well as rates of intracranial hemorrhage (ICH) between women and men. Additional surveillance is needed now that direct oral anticoagulants (DOAC) are being used more frequently.³⁹ Of note, although women were underrepresented in the four major trials comparing DOAC to warfarin (comprising 35 to 40% of participants) and formal sex comparisons were not presented, safety and efficacy for DOACs appear similar in women and men³⁹ with clinically negligible differences between DOACS.⁴⁰ Consistent with prior disparities in treatment of atrial fibrillation, newer data demonstrating lower rates of cardiac monitoring for the detection of arrhythmias following ischemic stroke in women are concerning and should be investigated further.^{23, 39} Ongoing research on the optimal time point for restarting anticoagulation in patients with atrial fibrillation and ICH should include sex-disaggregated data as well.

Stroke risk associated with lipid profiles is nuanced and differs between stroke subtypes but does not differ consistently by sex. Recent prospective data show an increased risk of ischemic stroke with increasing total cholesterol, LDL, and triglycerides as well as increased risk with increasing LDL/HDL ratio.^41–43 The relationship between lipid profiles and incident ICH is less well understood, with prospective data demonstrating an elevated risk of ICH risk with low triglyceride and LDL levels (<70) in women and elevated ICH risk associated with lower total and LDL cholesterol in both women and men in the UK biobank.^{28, 44}

Lifestyle factors including physical activity also reduce risk of stroke, although most data do not support a substantial sex difference in the impact of physical activity on stroke risk. More data are needed on the optimal level of physical activity for women compared with men given known sex differences in performance of strenuous physical activity among young and middle-aged adults as well as sex differences in activity limitations and frailty among older populations.⁴⁵

Female-specific risk factors

Female-specific risk factors to consider include those that are directly modifiable as well as factors like reproductive lifespan that are less directly modifiable but potentially important for risk prediction.

Use of oral contraceptive pills is associated with stroke in young women; one recent systematic review and meta-analysis demonstrated an odds ratio of 2.47 for use of any type of oral contraceptive, although when broken down by formulation, combined oral contraceptives remained significant while progesterone-only pills were not associated with increased risk.⁴⁶

For menopausal hormone therapy, despite clinical trial data demonstrating increased total and ischemic stroke among women taking oral menopausal hormone therapy (either estrogen alone or estrogen combined with progestin), more recent data suggest that use of low-dose transdermal estrogen formulations may effectively treat menopausal symptoms without increasing stroke risk.^{47, 48} Recent guidelines recommend that menopausal hormone therapy may be considered for women less than 60 years old or within 10 years of the onset of menopause without other contraindications in the case of vasomotor menopausal symptoms or other menopausal symptoms.^{49, 50}

Factors affecting lifetime estrogen exposure, defined as the time from menarche to menopause, are associated with elevated stroke risk; in a recent meta-analysis, a reproductive lifespan of <30 years was associated with a 75% increase in risk compared to 36 to 38 years.⁵¹ This association seems to be primarily driven by premature <40 years old) or early (40–44 years) menopause, though early or late menarche also appears to increase risk of stroke.⁵¹ These data are consistent across recent prospective studies.^{46, 52–54}

Other reproductive health variables associated with stroke risk include parity and breast feeding. Data on parity suggests elevated stroke among women with ≥5 live births compared with 1 or 2 live-births, though in some studies, adjusting for confounding variables, particularly BMI, eliminated this association.^{55, 56} With regards to breast feeding, recent data from the Women’s Health Initiative demonstrated a 23% reduction in risk of incident stroke among women who reported ever breast feeding, and the reduction appeared to increase with longer duration of reported breast feeding.⁵⁷

Adverse pregnancy outcomes, including preterm delivery, gestational hypertension, preeclampsia and fetal growth restriction, have been consistently associated with increased long-term risk of cardiovascular disease, including stroke, in the mother.⁵⁸ Recent data from the Women’s Health Initiative showed a consistent association between history of adverse pregnancy outcomes and future atherosclerotic cardiovascular disease, including stroke.⁵⁹ Some studies have showed triple the risk of future stroke (excluding peripartum stroke) in patients with a history of preeclampsia.⁶⁰ Other adverse pregnancy outcomes have been associated with increased risk of future stroke, including fetal growth restriction (30% increased stroke risk),⁶¹ gestational hypertension (80% increased risk)⁶⁰ and preterm delivery (65% higher risk).⁶² Furthermore, some studies demonstrate an association between history of hypertensive disorders of pregnancy and future vascular cognitive impairment.^{63, 64} Putative mechanisms of these associations including adverse pregnancy outcome-induced immunological intolerance and a heightened inflammatory state, causing long term effects on the maternal vasculature;⁶⁵ alternatively, a high-risk maternal vascular phenotype may predispose to both adverse pregnancy outcomes and future cerebrovascular disease.

Understanding stroke risk factors in transgender individuals, particularly the use of gender affirming hormones, is of critical importance, but current knowledge remains limited. In a cohort study of over 4960 transgender women matched with >96,000 cisgender men and women, transgender women had an elevated stroke risk compared with ciswomen (HR 1.9 [95%CI 1.3–2.6]), with an even greater increase in risk among transwomen who initiated estrogen treatment.⁴ For those using hormones, risk appeared to be the greatest after 6 years of follow-up (HR 4.1 (95%CI 1.5–11.4) compared with cisgender women.⁴ Ischemic stroke risk for transgender men did not appear to be increased when compared to cisgender men or women.⁴ These data are largely supported by smaller studies and reviews, but there is a clear need for high quality, prospective studies with appropriate control subjects to elucidate the risks of long-term sex steroid use to better counsel trans- individuals of all ages and stages of transition.^{5, 6}

As we enter an era of increasingly personalized medicine, applying sex-specific knowledge of stroke risk factors is the first step toward more effective, more targeted stroke prevention strategies.

Acute cerebrovascular complications in pregnancy and postpartum

Incidence of maternal stroke.

Stroke, including ischemic stroke, subarachnoid hemorrhage, and ICH, is among the most feared complications of pregnancy, accounting for 8.2% of US maternal deaths yearly.⁶⁶ While stroke is rare, occurring in approximately 30 per 100,000 deliveries,⁶⁷ the incidence has increased over the last 30 years, particularly in the postpartum period.^{42, 68–70} Some groups are at far higher risk; in patients with preeclampsia, stroke may occur in up to 1 in 500 pregnancies.⁷¹

Risk factors.

Hypertensive disorders of pregnancy, including chronic or gestational hypertension, preeclampsia/eclampsia, or the HELLP (hemolysis, elevated liver enzymes, and low platelets) syndrome, are among the most important maternal stroke risk factors.^72–74 Other notable risk factors include migraine,⁷⁵ thrombophilias,⁷⁶ congenital or acquired heart disease,⁷⁷ and infections.⁷⁸ Migraine confers increased risk of hypertensive disorders of pregnancy and maternal stroke, with some studies showing as high as 15-fold higher stroke risk in pregnant patients with migraine.^{69, 75} A recent retrospective study of 3 million births in California found a 6.8-fold higher risk of stroke in pregnant patients with migraine; a mediation analysis found approximately one quarter of the increased risk was mediated by higher risk of hypertensive disorders of pregnancy.⁷⁹ As with other adverse pregnancy outcomes, Black patients experience disproportionately high pregnancy-related stroke rates, independent of socioeconomic status and vascular risk factors.^{70, 80–82}

Timing.

While stroke can occur at any time during pregnancy or puerperium, the postpartum period confers the highest per-day risk of stroke, with the highest risk in the immediate peripartum and up to two weeks postpartum.^{67, 69, 83, 84}

Stroke mechanisms and pathophysiology.

Maternal stroke due to atherosclerotic disease is extremely rare. In contrast to the general population, where 87% of strokes are ischemic,⁹ up to half of maternal strokes are due to subarachnoid or ICH.⁴² Common stroke mechanisms seen in pregnant and postpartum patients include cardioembolism, carotid or vertebral artery dissection, infarction or hemorrhage due to cerebral venous sinus or cortical vein thrombosis, reversible cerebral vasoconstriction syndrome causing vasospasm-related ischemia or convexity subarachnoid hemorrhage, hypertensive ICH, and rupture of vascular anomalies such as intracranial aneurysms, arteriovenous malformations, or moyamoya vasculopathy.^{67, 84–88} The unique and complex physiology of pregnancy and its complications, particularly hypertensive disorders of pregnancy, contributes to an increased risk of all of these acute cerebrovascular disorders (Figure 3).

Figure 3:

Panel 1: Stroke in pregnancy or puerperium can occur due to pathological mechanisms throughout the cerebrovascular tree, including cardioembolism (A), cervical artery dissection (B), venous sinus thrombosis (C), rupture of vascular malformations (D), hypertensive intraparenchymal hemorrhage (E), or reversible cerebral vasoconstriction syndrome causing intracranial vasospasm and/or convexity subarachnoid hemorrhage (F).

Panel 2: Physiological changes of pregnancy affecting the cardiovascular, hematological, and immune systems may all increase stroke risk. Pathological complications of pregnancy such as peripartum cardiomyopathy or preeclampsia exacerbate this risk.

Panel 3: At the microvascular level, pathophysiological changes associated with preeclampsia may contribute to increased stroke risk, including neuroinflammation, leakiness of the blood-brain barrier, abnormal cerebral autoregulation, and sympathetic hyperreactivity.

Created with BioRender.com

The pregnancy “stress test.”

Pregnancy puts enormous strain on the maternal cardiovascular system, with cardiac output increasing by up to 50% by the end of gestation in response to fetal demand.⁸⁹ In addition, pregnancy-related hormonal changes lead to systemic vasodilation, increased vascular compliance and venous stasis.⁹⁰ Hypercoagulability is also part of the adaptive physiology of pregnancy, including increased levels of von Willebrand factor, factor V, factor VIII, and fibrinogen, acquired resistance to activated protein C, and placentally-produced anti-fibrinolytics (plasminogen activator inhibitors 1 and 2). Thus, even normal pregnancy is characterized by Virchow’s triad of venous stasis, hypercoagulability, and delivery-associated vascular damage, increasing the risk of thromboembolism.⁹¹ This risk is magnified in patients with preexisting cardiac conditions or comorbidities that put them at risk for arrhythmias or cardiomyopathy.

Immunomodulation and inflammation.

Inflammatory cytokines play crucial roles in implantation and parturition.⁹² In the first trimester, a highly inflammatory milieu predominates, with increased levels of pro-inflammatory T helper (Th)-1 activity, interleukin-6 (IL-6), C-reactive protein (CRP) and tumor necrosis factor alpha (TNFα), all of which affect endothelial function and may increase thrombotic risk.^{93, 94} The second trimester is characterized by immune quiescence, but near delivery, upregulation of NF-κB results in a robust inflammatory cytokine response and induction of labor.⁹⁵ Dysregulation of this normal immunomodulatory response, with resulting inflammation-induced endothelial dysfunction, has been implicated in the pathophysiology of preterm birth⁹⁶ and preeclampsia,⁹⁷ and may be a driver of maternal stroke risk.

Hypertensive disorders of pregnancy and maternal stroke.

Preeclampsia and related hypertensive disorders of pregnancy affect nearly 1 in 10 US pregnancies.⁹⁸ While the complex pathophysiology of preeclampsia is beyond the scope of this review, placental ischemia plays a critical role, leading to widespread maternal endothelial dysfunction as well as a procoagulant, pro-inflammatory, anti-angiogenic, and hyperreactive sympathetic state.^{99, 100} In addition to being highly associated with peripartum cardiomyopathy,¹⁰¹ preeclampsia has direct effects on the cerebral vasculature, including increased blood-brain barrier permeability, abnormal arteriolar remodeling, and sympathetic vasomotor activation.^102–106 Thus, the preeclamptic milieu may predispose to stroke via both indirect (hypercoagulability, inflammation, angiogenic imbalance, and endothelial dysfunction) and direct mechanisms (peripartum cardiomyopathy-induced cardioembolism, venous sinus thrombosis, arterial dissection, reversible cerebral vasoconstriction syndrome, and hypertensive ICH). Interestingly, the use of low dose aspirin during pregnancy reduces the incidence of preeclampsia in high-risk patients by more than 50%,¹⁰⁷ and is recommended by the US Preventive Services Task Force after 12 weeks gestation in patients at high risk of preeclampsia.¹⁰⁸ Exactly how aspirin reduces preeclampsia risk is unknown, with multiple mechanisms proposed.^109–111 In addition, it is not known whether people who have experienced preeclampsia should continue taking aspirin after pregnancy as primary prevention for long-term cardiovascular and cerebrovascular disease; however, some observational evidence has suggested a possible benefit of aspirin use for stroke risk reduction in this population.⁷⁸

Outcomes.

Maternal outcomes after stroke depend on the stroke mechanism. Pregnant patients with acute ischemic strokes treated with reperfusion therapy (thrombolysis or mechanical thrombectomy) have similarly good outcomes compared to non-pregnant patients, with few complications,^{73, 112, 113} prompting both the American Heart Association/American Stroke Association and Canada Heart and Stroke to recommend consideration of these therapies in pregnant patients with disabling stroke symptoms.^{114, 115} Outcomes after maternal hemorrhagic stroke, particularly ICH, are substantially worse, often leading to death or severe disability.^{74, 84, 116–120} Patients with hypertensive disorders of pregnancy also have worse outcomes after stroke, compared to those without, regardless of stroke mechanism.^{72, 121, 122}

Gaps in knowledge and future directions.

Despite being a major cause of severe maternal morbidity and maternal mortality, maternal stroke continues to be poorly understood, especially in the postpartum period. Many epidemiological studies are based on large administrative datasets using billing data, while studies that include detailed chart review and validation are small and may suffer from selection bias. Our understanding of the intriguing relationship between migraine, preeclampsia, and maternal stroke remains limited. There is an urgent need for more translational, clinical, and rigorous population-based studies, particularly prospective or registry-based studies, to better characterize the mechanisms and predictors of maternal stroke. Robust collaboration across disciplines, specifically between researchers in cerebrovascular and cardiovascular disease, neurobiology, maternal-fetal medicine, and placentology, will be needed to address these gaps in knowledge and reduce maternal cerebrovascular morbidity and mortality.

Stroke presentation and diagnosis

Early recognition of stroke symptoms is the first step in ensuring rapid investigations to confirm the diagnosis, understand prognosis, and institute treatment. A systematic review of 15,721 patients found that about 9% of cerebrovascular events are initially not diagnosed as such in the emergency department and a large population-based cross-sectional study on this issue found that men had 25% lower odds of misdiagnosis than women.^{123, 124} Misdiagnosis leads to missed opportunities for treatment, rehabilitation, and secondary prevention to improve stroke outcomes and reduce disability.

Cerebrovascular diseases have a wide spectrum of disease severity, ranging from TIA to severe strokes with disabling motor, visual, or language deficits among other symptoms. Diagnosing TIA or non-disabling stroke is particularly challenging because symptoms may be vague and signs on physical examination are often subtle.¹²⁵ A Japanese study using data from a nationwide clinical registry found that 16% of patients with patients with TIA report non-focal symptoms, particularly in posterior circulation events.¹²⁶

A common teaching pearl in clinical practice is that women are more likely than men to report “atypical” stroke symptoms. Indeed, prior work has shown that women are more likely than men to report non-focal neurological symptoms, such as headaches, fatigue, cognitive changes, generalized malaise or weakness.^127–129 However, the absolute difference in symptoms by sex was small and these studies focused on individual symptoms even though patients with TIA or stroke commonly report a constellation of symptoms. A recent study found that even though non-focal symptoms were common in patients presenting with TIA or minor stroke, over 90% of patients reported some form of focal symptoms, defined as deficits involving motor, sensory, vision, or speech (aphasia or dysarthria) function, and importantly, there were no sex differences in presenting symptoms.¹³⁰ Thus, it is important to recognize that non-focal neurologic symptoms, when reported with focal ones, should not discourage clinicians from investigating further for stroke.

Another potential reason why the more recent study did not find sex differences in stroke symptoms is because the authors included all patients with suspected TIA or minor stroke referred to stroke neurologists.¹³⁰ Most prior work on the topic of stroke symptoms first identified patients with TIA or stroke and subsequently compared the symptoms reported by women and men.^127–129 Given that misdiagnosis is more common in women than men, this approach could lead to selection bias. Investigators in cardiology took a similar approach to examine the nature of chest pain in 1,941 patients with suspected acute coronary syndrome presenting to an emergency department, regardless of final diagnosis.¹³¹ Although it is often said that women experience “atypical” chest pain, the authors found that using their approach, “typical” chest pain was in fact more common in women than men, and it also had greater predictive value for myocardial infarction in women.

In addition, sex differences in comorbid conditions may influence the clinical diagnosis. In the Diagnosis of Uncertain Origin Benign Transient Events (DOUBT) study, a prospective cohort study of patients with transient or minor neurological events, women were more likely than men to report a history of migraine (33% women and 17% men) and recent stressors (19% women and 12% men), which are two conditions that may lower the clinical suspicion of stroke.¹³² Indeed, migraine aura and somatoform disorder were among the most common stroke mimic diagnosis in DOUBT. On the other hand, there are known sex-specific vascular risk factors, such as hypertensive disorders of pregnancy and premature menopause, yet these conditions are not routinely included in the history-taking for a vascular event.^{52, 100} Further studies are needed to understand whether there are differences in how women describe their symptoms compared to men, provider-level biases when evaluating symptoms described by women and men, and whether sex-specific cardiovascular disease guidelines may reduce gaps in care.¹³³

Even with a detailed history physical examination, ambiguous clinical scenarios are unavoidable. In a prospective multi-national cohort study of 1,028 patients with acute low-risk transient or minor stroke symptoms, the authors found that routine use of MRI brain within 8 days of symptom onset led to a revision in diagnosis in 30% of study participants, either from stroke mimic to TIA/stroke or vice versa.¹³² A subsequent sex-stratified analysis of this cohort found that women were more likely than men to be initially diagnosed as stroke mimic, but among 496 patients initially diagnosed with stroke mimic, men and women were equally likely to have acute infarct on MRI brain (5% of women and 8% of men, adjusted odds ratio 0.56, 95% confidence interval [0.26, 1.21]).¹³⁴ Researchers from the Netherlands similarly showed that up to a quarter of patients clinically diagnosed with non-ischemic “transient neurological attacks” have an acute infarct on MRI.¹³⁵ In the context of these studies, a brain MRI was mandated by the study protocol, but in clinical practice, advanced imaging is frequently only pursued when there is a clinical suspicion of stroke. Furthermore, prior studies have shown that women, even after being diagnosed with stroke, were less likely than men to be investigated with the standard diagnostic tests and women admitted to hospital with stroke are less likely to receive defect-free care.^{135, 136} As shown in Figure 4, sex differences in investigations for suspected stroke can perpetuate the higher risk of misdiagnosis in women.

Figure 4. Case study: 54-year-old woman with a history of migraine with aura presents with 15 minutes of right arm heaviness and hand clumsiness.

This occurred during a stressful work meeting. She needed to concentrate to speak at the meeting, but she does not think her colleagues noticed anything. She experienced a headache afterwards. She presents to medical attention.

In summary, any sex differences in stroke symptoms, if present, are likely to be small and we should move away from the male-centric narrative that women overall are more likely to report “atypical” symptoms of cardiovascular disease compared to men. Instead, the focus should be on improving diagnosis through establishing standardized investigations for suspected acute cerebrovascular events, implementing strategies to reduce sex differences in investigations and care, and routinely inquiring about sex-specific risk factors for stroke.

Acute treatment of stroke

Ischemic stroke

Intravenous thrombolysis with recombinant tissue plasminogen activator (IV rtPA) and endovascular thrombectomy are evidence-based treatments for acute ischemic stroke that improve outcomes in carefully selected individuals. However, there continues to be controversy as to whether sex differences exist in access, efficacy, and safety of these treatments.

A meta-analysis of 24 studies, published between 2008 and 2018, reported on sex-specific usage rates of IV rtPA use in ischemic stroke.¹³⁷ These studies were a mix of hospital-based, registry-based and administrative data studies. While there was some heterogeneity between studies, the analysis found that women had a 13% lower odds of receiving IV rtPA treatment compared to men (summary unadjusted OR 0.87 [95% CI, 0.82–0.93]). These sex differences in IV rtPA use in ischemic stroke differed across regions and were observed in European (summary unadjusted OR was 0.82 [95% CI, 0.78–0.85]) and United States-based (0.85 [95% CI, 0.75–0.96]) studies but not in studies from Asia (0.98 [95% CI, 0.94–1.03]) and Germany.^137,138 Thus, further research into the causes of lower rtPA treatment in women, and effective strategies to improve equity, is needed. Given the high incidence of ischemic stroke in women, treatment differences in IV rtPA use are likely to translate into a large number of untreated women. In addition, because women have worse stroke outcomes than men,¹³⁹ but receive comparable treatment benefits,¹³⁷ this missed treatment opportunity could have significant consequences for disability status in women.

Secondary analysis of large, international, randomized controlled trials comparing IV rtPA to placebo in ischemic stroke have not shown any sex differences in clinical outcome.¹⁴⁰ Being post-hoc analysis, these studies were not specifically designed to test sex differences in efficacy; however, the findings are reassuring for both women and men. Symptomatic ICH is one of the most serious adverse effects of IV rtPA. In the Safe Implementation of Treatments in Stroke-International Stroke Thrombolysis Register (SITS-ISTR),¹⁴¹ a total of 45,079 patients treated with IV rtPA were recorded from 2002 to 2011. In the multivariate analysis, after adjustment for confounding variables such as age, stroke severity and time to treatment), men had 19% higher odds of symptomatic ICH (OR of 1.25 [95% CI 1.04–1.51] with women as referent).

Several studies have suggested higher rates of endovascular thrombectomy in women with ischemic stroke than men. Sex differences in ischemic stroke treatment were explored in a nationwide administrative database in Germany where over 1.11 million patients were hospitalized for first or recurrent ischemic stroke from 2013 to 2017. During the 5-year study period, women had an overall 26% (95% CI, 22%–30%) higher chance of receiving endovascular thrombectomy. This higher endovascular thrombectomy utilization in women compared to men was significantly higher across all age groups.¹⁴² Similar findings were seen in a retrospective, serial, cross-sectional study of over 4 million patients in the United States, where women were 20% more likely to be treated with endovascular thrombectomy compared with men in the period 2016 to 2017.¹⁴³ The reasons for this sex disparity in endovascular thrombectomy use is not known but women with atrial fibrillation are less often treated for stroke prevention which may lead to the great number of large vessel cardioembolic strokes in women and higher need for endovascular thrombectomy.

Whether there are sex differences in endovascular thrombectomy efficacy is controversial. In the Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN trial)¹⁴⁴ of 500 patients (42% women), the treatment effect (modified Rankin score at 90 days) was null in women (OR 0.99 [95% CI 0.60–1.66]) compared to protective in men (OR 2.39 [95% CI 1.55–3.68]). Women in the intervention group were also more likely to die, as well as have serious adverse events, than men in MR CLEAN. In a later meta-analysis of individual patient data from five randomized trials of endovascular thrombectomy after large-vessel ischemic stroke (including MR CLEAN), no sex differences were found in the benefit of the intervention on 90-day functional outcomes, nor on serious adverse events.¹⁴⁵ Additionally, in the combined cohorts SWIFT (Solitaire Flow Restoration With the Intention for Thrombectomy), STAR (Solitaire Flow Restoration Thrombectomy for Acute Revascularization), and SWIFT PRIME (Solitaire Flow Restoration With the Intention for Thrombectomy as Primary Endovascular Treatment) where 389 eligible patients were treated with endovascular stroke thrombectomy (55% women), women were found to have had more disability-adjusted life years after endovascular thrombectomy compared to men, 10.6 versus 8.5 years (P<0.001). These findings diminished with increasing age.¹⁴⁶

In summary, women are more likely to receive endovascular thrombectomy, although it is not clear whether this is due to differences in stroke types and presentations. There are no sex differences in efficacy and safety; however, women may gain more optimal life years after this procedure. More research on optimizing the use of revascularization treatment in men and women is required.

The most recent guidelines on the management of carotid artery stenosis recommend carotid endarterectomy in symptomatic patients with ≥ 50% stenosis and in asymptomatic patients ≥60% stenosis.^{147, 148} Carotid artery stenting can be considered for symptomatic patients with stenosis of ≥ 50% at high risk for carotid endarterectomy for anatomic or medical reasons¹⁴⁷ or for patients < 70 years old with symptomatic ≥ 50% carotid stenosis.¹⁴⁸ These recommendations warrant a review on the impact of sex on these surgical and endovascular treatments of carotid stenosis. Not only are these guidelines based on randomized trials performed over a decade ago, before best medical therapy or technological advances were yet to be refined, but the trials have also underrepresented women (<35% women enrolled)^{149, 150} and reported higher periprocedural risk of for any stroke, myocardial infarction, or death in women undergoing carotid artery stenting compared to carotid endarterectomy. In a pooled analysis of symptomatic patients in large, randomized trials, sex differences were examined in the carotid artery stenting -to- carotid endarterectomy risk for any stroke or death during the 120-day periprocedural period and stroke.¹⁵¹ The authors found that there was significant heterogeneity in the 4 completed revascularization trials and pooling of the studies was not appropriate. No clear conclusions can be drawn regarding carotid artery stenting versus carotid endarterectomy. These findings further strengthen the need to adequately power trials to allow reliable sex-disaggregated analysis and conclusions about the treatment for carotid artery stenosis in women.

Intracerebral hemorrhage

Unlike ischemic stroke, there are no medical treatments definitively proven in randomized clinical trials to improve outcomes for ICH.¹⁵² The primary goals of acute management are to stabilize the patient to ensure they survive the initial insult by managing blood pressure, reversing anticoagulant-associated hemorrhage and neurosurgical intervention. The Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage trials 1 and 2, randomized patients to intensive (target <140mmHg within 1 hour) or guideline-recommended blood pressure lowering treatment. The interventions showed no differences in the primary outcome of death or major disability. In a post-hoc analysis of the 3233 patients in the pooled studies,¹⁵³ the 37% women were significantly older, had higher stroke severity and smaller hematoma volumes compared to men. There was no effect of sex on the combined endpoint of death or major disability, hematoma growth or effect of randomized blood pressure lowering treatment.

Nearly twenty percent of all ICH cases are associated with anticoagulant use.¹⁵² Discontinuing the offending agent and reversing anticoagulation is key to improving outcomes. Idarucizumab¹⁵⁴ is a fragment of humanized monoclonal antibody that binds dabigatran and rapidly reverses its anticoagulant activity. At this time, only one trial on the reversal of direct oral anticoagulation has included anticoagulant-associated ICH. This multicenter, prospective, single cohort study enrolled 503 patients in 39 countries with 53 (17.6%) having ICH. The median maximum percentage reversal within 4 hours after the administration of idarucizumab was 100%. The reversal occurred independent of age, sex, renal function, and dabigatran concentration at baseline; however, sex-specific outcomes for patients with ICH were not reported.

Sex differences in ICH treatment effects have been less often reported compared to ischemic stroke—particularly for access to and surgical treatment outcomes after ICH. Disaggregating results by sex, especially in interventional trials, is vital for generalizability of results to both sexes.

Stroke Outcomes

As the US population ages, it is estimated that by 2050, there will be an excess of 68,000 stroke deaths in women as compared with men, with the greatest excess seen in women 85 years of age or older.¹⁵⁵ Studies have reported variable results for stroke case fatality by sex, with many providing little evidence of a difference, some showing higher case fatality, and others reporting lower case fatality for women.^{9, 155–161} Overall, these studies suggest that baseline differences in age, stroke characteristics, and cerebrovascular risk factors account for much of the observed sex differences in case fatality. Age-adjusted mortality rates are commonly reported, but because sex differences are strongly modified by age, they can mask the complex relation of sex differences across age groups and can hide the burden of stroke mortality for elderly women.

As shown in Figure 1, using the most recent national data on stroke death rates per 100,000 population from the Centers for Disease Control and Prevention (CDC), Lichtman and colleagues graphed the total number of cerebrovascular deaths, the percentage of total deaths due to cerebrovascular disease, and the ranking of cause of death for cerebrovascular diseases for women and men. The total number of deaths due to cerebrovascular diseases for women exceeded those for men starting from the age of 75 years and remained substantially higher for women across all older age groups. Cerebrovascular deaths comprised a greater proportion of total deaths for women as compared with men for all age groups and was ranked higher as a leading cause of death for women than men until the age of 70 years.

Studies have generally reported poorer functional recovery and lower quality of life after stroke for women compared to men.^{21, 155, 160–169} Moreover, women fared worse in the measured dimensions of pain/discomfort, anxiety/depression, fatigue, and mobility than men.^{15, 21, 155, 160, 161, 169, 170} Individual-level factors such as advanced age, poorer pre-stroke function, comorbidities, lower social support, and an increased likelihood of being widowed have been hypothesized as contributors to sex and gender differences in functional outcomes and quality of life, yet adjustment for these factors does not fully explain the observed differences in outcomes for women and men.^{155, 160, 169} There remains a need to explore the role of access to clinical care and rehabilitation services as well as social, mental health, and socioeconomic factors that may be contributing to observed sex and gender differences in these outcomes. Few studies have addressed sex differences in cognitive impairment after stroke. A focused review included 5 studies that found mixed results, with some showing a higher risk among men and others showing a higher risk among women.¹⁶⁹ None of the studies were population-based. Similarly, there is a paucity of population-based studies that report sex differences in recurrent stroke events. A study of 8900 adult ischemic stroke patients in a rural Pennsylvania population did not find a sex difference in 5-year survival after controlling for age and comorbid conditions.¹⁵⁸

Depression is common following a stroke, affecting up to one-third of stroke survivors.¹⁷¹ Across the adult lifespan, depression is nearly twice as common in women than men.¹⁷² Studies have found a higher prevalence of poststroke depression in women than men.^{21, 155, 169, 170, 173–175} A systematic review of the prevalence of poststroke depression in 45 publications between 1982 and 2006 found the prevalence of PSD was 78% higher among women as compared with men.¹⁷⁵ A more recent review of sex differences identified 10 papers that reported poststroke depression.¹⁶⁹ In the 8 papers that reported unadjusted associations between sex and PSD, women had a higher likelihood of poststroke depression (OR range 1.27–3.15 and HR 3.52) or a higher mean number of poststroke depression symptoms. Women remained significantly more likely to have higher prevalence, incidence, or increased number of symptoms of depression than men after multivariable adjustment for factors such as age, stroke severity, and activity limitation. The majority of studies to date have not been population-based or specifically designed to examine sex differences in poststroke depression. A study of 786 patients from the population-based Brain Attack Surveillance in Corpus Christi Project assessed sex differences in poststroke depression at 90 days after first-ever stroke using the 8-item Patient Health Questionnaire.¹⁷⁴ The study did not find a statistically significant sex difference in poststroke depression after accounting for attrition. However, the study reported significant effect modification of the association between sex and poststroke depression such that women on medication for depression at the time of stroke were significantly protected against poststroke depression, whereas women with a history of depression who were not on medication showed an elevated risk compared with men. The results highlight that depression among stroke survivors is heterogeneous and multifactorial with sex differences throughout the disease course before and after a stroke. Because untreated depression after stroke can lead to a poorer prognosis, reduced quality of life, and increased mortality, enhanced screening and the development of interventions to prevent poststroke depression are needed, particularly for women and those with pre-stroke depression not undergoing treatment for depression at the time of the stroke.^{169, 171, 174, 175}

In summary, research shows that women experience worse outcomes after stroke than men regarding mortality, quality of life, poststroke depression, and activity limitations. Differences are partly explained by women’s poorer health at the time of the index stroke event, advanced age, and greater stroke severity. Data on how aspects of access and quality of clinical care, pre-stroke health, mental health, social isolation, loneliness, social support, or socioeconomic status may be contributing to worse outcomes for women are limited.^{169, 176} Rigorous population-based studies are needed to explore these potential factors for sex differences in outcomes, including designs that can study the bidirectional nature between sex and these factors.

Conclusion: Gaps in research

Despite substantial progress in examining sex differences in stroke as well as specific factors influencing risk and outcomes in women, significant research gaps remain (Table). Overall, sex and gender have often been combined in stroke research; thus, delineating the impact of biologic sex from the social impact of gender-related factors needs to continue. With regard to risk factors, the effects of gender-affirming hormone therapy on risk of stroke need to better quantified, particularly for transgender women. While a number of female-specific risk factors have been noted, no currently available stroke prediction models incorporate female-specific risk factors. Despite the higher long-term risk of stroke among individuals with a history of adverse pregnancy outcomes, including hypertensive disorders of pregnancy, preterm delivery and fetal growth retardation, we lack specific screening and treatment guidelines to reduce this risk. To reduce misdiagnosis in women, research is needed to understand whether women describe their symptoms differently, whether provider-level biases influence the interpretation of symptoms described by women and men, and whether strategies to reduce sex differences in investigations and care may reduce gaps in diagnosis. With regard to treatment, although women comprise more than half of strokes, the inclusion of women in clinical trials of stroke treatments has been lower.^{177, 178} Methods to achieve better representation of women in stroke clinical trials should include carefully scrutinizing inclusion criteria and their impact on gender balance in trials, avoiding age-based exclusion criteria, and increasing the number of women who lead stroke clinical trials.¹⁷⁹ Further, it is essential that clinical trials and registries present sex disaggregated data and stratified results. Given older ages and higher disability prior to stroke and worse outcomes after stroke for women, the role of access to clinical care and rehabilitation services as well as social, mental health, and socioeconomic factors that may contribute to these observed sex differences are needed. Better knowledge of the impact of sex and gender on stroke will improve outcomes for all individuals.

Table.

Research gaps to eliminate disparities in stroke by sex and gender.

Research area	Research gaps
Stroke epidemiology	• Delineation of the separate effects of biologic sex and sociocultural gender
Stroke risk factors	• Delineating the separate effects of biologic sex and sociocultural gender • Prediction models with female-specific risk factors • CVD prevention for women with a history of adverse pregnancy outcomes: screening and treatment guidelines, as well as understanding biologic mechanisms
Stroke diagnosis	• Research on strategies to reduce misdiagnosis in women, as well as in investigation and care
Stroke treatment	• Inclusion of women in clinical trials consistent with their stroke incidence • Increasing the number of women who lead clinical trials • Presentation of sex disaggregated data and stratified results from all clinical trials • Investigation of sex differences in outcomes with endovascular therapy, carotid artery stenting and carotid endarterectomy
Stroke outcomes	• Improved understanding of the factors contributing to worse outcomes after stroke in women • Interventions to improve outcomes including disability and post-stroke depression

Nonstandard Abbreviations and Acronyms:

ICH

intracerebral hemorrhage

BMI

body mass index

IV rtPA

intravenous thrombolysis with recombinant tissue plasminogen activator

TIA

transient ischemic attack