Sex Disparities in Mortality After Endovascular Therapy in Large Core Infarcts

Abstract

Background

In recent large core endovascular therapy (EVT) trials of large vessel occlusion acute ischemic stroke (AIS), treatment was associated with reduced rates of mortality. Because post‐AIS mortality can be influenced by societal and biological factors that differ between women and men, we investigate sex‐based differences in mortality outcomes following EVT in large core AIS.

Methods

From our prospectively collected multicenter registry across 4 comprehensive stroke centers in the Greater Houston area, we identified patients from 2017 to 2022 with large vessel occlusion AIS and large infarct core. Large infarct core was defined by computed tomography perfusion exceeding 70 mL (by regional cerebral blood flow measurements using automated postprocessing) or computed tomography Alberta Stroke Program Early CT [Computed Tomography] Score<6. The primary outcome of this study was the likelihood of mortality at 90 days, determined through multivariable logistic regression adjusted for EVT, sex, and EVT/sex interaction term. Secondary outcomes included 90‐day disability outcomes and intracerebral hemorrhage.

Results

Among 190 patients who met inclusion criteria, 50% were female and 45.3% received EVT. Demographic differences between the sexes were largely balanced apart from the older age of presentation for women compared with men (75 versus 67, women versus men; P<0.01). In univariable analysis, women who did not receive EVT had greater mortality (27.4% difference; P<0.001) compared with men, with comparable rates in EVT‐treated cohorts. In multivariable analysis, non‐EVT management was strongly associated with mortality in women compared with men at discharge (odds ratio [OR] 5.81, 95% CI [1.96–17.23]) and 90‐days (OR 6.77, 95% CI [2.09–21.94]). In the secondary analysis, which additionally adjusted the model for age and National Institutes of Health Stroke Scale score, these findings were unchanged. The sex/EVT interaction term showed significant interaction for mortality both at discharge and 90 days (P<0.01).

Conclusion

EVT in large core AIS populations may disparately reduce mortality in women compared to men.

Nonstandard Abbreviations and Acronyms

AIS

acute ischemic stroke

CTP

computed tomography perfusion

EVT

endovascular stroke therapy

LVO

large vessel occlusion

Clinical Perspective

What Is New?

• Endovascular therapy benefits are more pronounced in reducing mortality in women compared with men with large core infarct, suggesting a potential sex‐specific impact on mortality reduction.

What Are the Clinical Implications?

• Endovascular therapy shows greater mortality reduction in women compared with men with large core infarct, underscoring the need to consider sex‐specific impacts on treatment outcomes in stroke management.

Several studies comparing endovascular thrombectomy (EVT) to standard medical management in patients with large core infarcts have found a significant reduction in mortality in patients treated with EVT relative to medical management alone. In the TENSION (Endovascular Thrombectomy for Acute Ischaemic Stroke with Established Large Infarct) trial, mortality in the EVT arm was approximately 18% lower; LASTE (Trial of Thrombectomy for Stroke with a Large Infarct of Unrestricted Size) ¹ showed a similar approximately 20% reduction. ² Other trials demonstrated reductions in the point estimate of mortality that did not achieve statistical significance. As an outcome, however, mortality can be a challenging endpoint as cultural, societal, and other factors can influence decisions on whether to continue or withdraw care. Further exploration of the EVT effect on mortality in large core acute ischemic stroke (AIS) across key subgroups is needed. Supporting information

In a recent Centers for Disease Control and Prevention study conducted in the United States, women had a greater mortality rate from cerebrovascular disease compared with men, starting at the age of 75 and persisting across all older age groups. ³ Other cohort studies have shown that stroke‐related mortality in women surpasses men across the adult lifespan. ⁴ This known disparity in mortality rates raises questions about disproportionate treatment effects on mortality reduction in men versus women. Here, we examine whether the mortality benefit associated with EVT in large core stroke, as observed in clinical trials, can also be seen in clinical practice and whether this effect is consistent for both women and men. We hypothesize that because women may suffer greater rates of poststroke mortality, the mortality reduction with EVT in this population will be greater in women compared with men.

Methods

The data used in this study are available for collaborative groups, provided there is approval from the institutional review boards and data sharing agreements in place. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline, and the study protocol was reviewed and approved by the Institutional Review Board at UTHealth Houston.

Participants

From our prospectively collected multicenter registry encompassing 4 comprehensive stroke centers in the greater Houston area, we identified patients between 2017 and 2022. Patients were included in the study if they were found to have large vessel occlusion (LVO) AIS with large infarct core at presentation, which was defined as described in the subsequent subsection. Patients were excluded if they had incomplete data for arrival time or 90‐day follow‐up modified Rankin scale (mRS) score. Additionally, patients with posterior circulation occlusions and those with data errors were excluded.

Measurements

LVO was defined as an occlusion of the intracranial internal carotid artery, A1 or A2 segments of the anterior cerebral artery, or M1 or M2 segments of the middle cerebral artery based on radiology read of computed tomography (CT) angiogram. Infarct core was defined using CT perfusion (CTP) and CT Alberta Stroke Program Early CT Score (ASPECTS). Patients who underwent CTP had image postprocessing performed by RAPID (IschemaView, CA) or Viz.AI. Patients were defined as large core if CTP predicted infarct volume was >70 mL by relative cerebral blood flow reduction measurement, or if noncontrast head CT ASPECTS was <6. ASPECTS grading was obtained from clinical radiology reports performed by the neuroradiologist or if not available, by review of experienced neurointerventionalist blinded to clinical details. Symptomatic intracerebral hemorrhage (sICH) was defined as parenchymal hematoma type II by European Cooperative Acute Stroke Study criteria, with a worsening of the NIHSS score by at least 4 points. ⁵ Good outcome was defined by the mRS scores ranging from 0 to 2, and excellent outcome was indicated by the mRS scores ranging from 0 to 1.

Outcomes

The primary outcome was the likelihood of mortality at 90 days. Secondary outcomes included the likelihood of mortality at discharge, functional independence (mRS 0–2) at 90‐day follow‐up, mRS 0–3 at 90‐day follow‐up, mRS 0–1 at 90‐day follow‐up, ICH, and sticiICH.

Statistical Analysis

A multivariable logistic regression model was performed to assess the EVT effect and its modified effect by sex on the outcomes. To further examine the interaction between EVT and sex, we stratified our data by sex, dividing the data set into male and female groups. This approach allows for a more detailed examination of EVT's impact on outcome variables compared to non‐EVT cases by controlling the sex. Using the stratified group, we visualized and analyzed the impact of EVT compared to non‐EVT cases on the outcome variables survival at 90 days, survival at discharge, mRS score 0–3 at 90‐day, mRS score 0–2 at 90‐day, and mRS score 0–1 at 90‐day in a forest plot. In secondary model adjustment, another multivariable logistic regression model was applied, which further adjusted for age and NIHSS score.

The interaction between EVT and sex was plotted using marginal effects of sex at non‐EVT treatment versus EVT treatment from the multivariable logistic model. Univariable comparisons between categorical variables were performed using chi‐square and between continuous variables with the Wilcoxon rank‐sum test. Analyses were performed using STATA v.17 (StataCorp LLC, College Station, TX).

Results

Among 190 patients with LVO AIS who met the inclusion criteria, 95 were male (50%). Of these patients, 150 (78.94%) underwent CTP imaging to assess core volume. Among those who underwent CTP imaging, 43 (28.67%) had a core volume of <70 mL, 42 (28%) had a core volume between 70 and 99 mL, and 65 (43.33%) had a core volume of ≥100 mL. As shown in Table 1, the median age for the cohort was 71 (interquartile range [IQR], 60–82) years, the median NIHSS at presentation was 20 (IQR 16–24), and the median last known well to arrival was 344 (IQR, 133–744) minutes. Of the 86 patients treated with EVT (45.3%), 40 were female. There were no significant disparities between the sexes in the last well‐known arrival time, CTP core volume, NIHSS score, ASPECTS, EVT performance, or thrombolysis rates. Female patients were older than male patients (female, 75 years IQR [64–83]; male, 67 years IQR [59–77]; P value = 0.006). Demographic data stratified by EVT treatment and by sex for the non‐EVT cohort are available in Tables S1 and S2. EVT was associated with an increased likelihood of ICH (odds ratio [OR] = 2.65, 95% CI [1.15–6.11]). In univariable analysis, women who did not receive EVT had greater mortality (27.4% difference, p<0.001) compared with men, with comparable rates of mortality in EVT‐treated cohorts. (Table 2).

Table 1.

Demographic Full Cohort

	Total (N = 190)	Men (N = 95)	Women (N = 95)	P value
EVT	86 (45.3%)	46 (48.4%)	40 (42.1%)	0.38
Age	71 (60–82)	67 (59–77)	75 (64–83)	0.006
Race or ethnicity
NH‐White	87 (45.8%)	40 (42.1%)	47 (49.5%)	0.059
NH‐Black	45 (23.7%)	22 (23.2%)	23 (24.2%)
NH‐Asian	11 (5.8%)	8 (8.4%)	3 (3.2%)
Hispanic	41 (21.6%)	19 (20.0%)	22 (23.2%)
Other *	6 (3.2%)	6 (6.3%)	0 (0.0%)
LKW/onset to arrival (min)	344 (133–744)	355 (162–686)	304 (121–774)	0.95
Arrival to puncture (min)	96.5 (65–124)	95 (64–126)	97 (69.5–123)	0.68
CTP core volume (mL)	94.5 (54–120)	90 (38–113)	95.5 (62–123.5)	0.17
Final TICI score
0	6 (7.5%)	2 (4.8%)	4 (10.5%)	0.24
1	3 (3.8%)	0 (0.0%)	3 (7.9%)
2a	3 (3.8%)	2 (4.8%)	1 (2.6%)
2b	23 (28.7%)	11 (26.2%)	12 (31.6%)
2c	9 (11.2%)	4 (9.5%)	5 (13.2%)
3	36 (45.0%)	23 (54.8%)	13 (34.2%)
tPA	78 (41.1%)	42 (44.2%)	36 (37.9%)	0.38
LKW/onset to tPA (min)	112 (73.5–163)	120.5 (84.5–192.5)	109 (64.5–152.5)	0.27
CTA occlusion location
ICA	64 (33.9%)	30 (31.6%)	34 (36.2%)	0.61
MCA	119 (63.0%)	61 (64.2%)	58 (61.7%)
Other	6 (3.2%)	4 (4.2%)	2 (2.1%)
NIHSS	20 (16–24)	19 (16–23)	21 (17–25)	0.13
ASPECTS	5 (3–5)	5 (3–5)	5 (3–5)	0.97
Patients’ history
Atrial fibrillation	47 (24.7%)	20 (21.1%)	27 (28.4%)	0.24
Hypertension	124 (65.3%)	56 (58.9%)	68 (71.6%)	0.067
Congestive heart failure	16 (8.4%)	8 (8.4%)	8 (8.4%)	1.00
Diabetes	54 (28.4%)	22 (23.2%)	32 (33.7%)	0.11
Prior stroke	45 (23.7%)	22 (23.2%)	23 (24.2%)	0.86
Smoking	27 (14.2%)	18 (18.9%)	9 (9.5%)	0.061
Coronary artery disease	28 (14.7%)	11 (11.6%)	17 (17.9%)	0.22
Hyperlipidemia	71 (37.4%)	39 (41.1%)	32 (33.7%)	0.29
Peripheral artery disease	4 (2.1%)	2 (2.1%)	2 (2.1%)	1.00

Data are presented as median (interquartile range [IQR]) for continuous measures, and n (%) for categorical measures. ASPECTS indicates Alberta Stroke Program Early CT Score; CTA, computed tomography angiography; CTP, computed tomography angiography perfusion; EVT, endovascular thrombectomy; ICA, internal carotid artery; LKW, NH, Non‐Hispanic; NIHSS, National Institutes of Health Stroke Scale; TICI, thrombolysis in cerebral infarction; and tPA, tissue plasminogen activator.

^*“Other” category includes Native American and unclassified.

Table 2.

Univariable Outcome of (Non‐EVT) and (EVT)

(a) Univariable outcome (non‐EVT)
	Total (N = 104)	Men (N = 49)	Women (N = 55)	P value
Mortality 90 d	83 (79.8%)	32 (65.3%)	51 (92.7%)	<0.001
Discharge mortality	81 (77.9%)	31 (63.3%)	50 (90.9%)	<0.001
ICH	30 (28.8%)	17 (34.7%)	13 (23.6%)	0.21
mRS score 0–3	11 (10.6%)	8 (16.3%)	3 (5.5%)	0.072
mRS score 0–2	8 (7.7%)	6 (12.2%)	2 (3.6%)	0.10
mRS score 0–1	6 (5.8%)	5 (10.2%)	1 (1.8%)	0.067

(b) Univariable outcome (EVT)
	Total (N = 86)	Male (N = 46)	Female (N = 40)	P value
Mortality 90 d	36 (42%)	23 (50%)	13 (32%)	0.10
Discharge mortality	31 (36%)	20 (43%)	11 (28%)	0.12
sICH	10 (12%)	6 (13%)	4 (10%)	0.66
ICH	53 (62%)	27 (59%)	26 (65%)	0.55
mRS score 0–3	25 (29%)	14 (30%)	11 (28%)	0.76
mRS score 0–2	11 (13%)	6 (13%)	5 (12%)	0.94
mRS score 0–1	8 (9%)	4 (9%)	4 (10%)	0.84

EVT indicates endovascular thrombectomy; ICH, intracerebral hemorrhage; mRS, modified Rankin Scale; and sICH, symptomatic intracerebral hemorrhage.

In multivariable logistic regression model including EVT performance, sex, and their interaction term, we observed a significantly greater likelihood of mortality at 90‐day follow‐up (OR = 6.77, 95% CI [2.09–21.94]) and discharge (OR = 5.81 95% CI [1.95–17.23]) among female patients who did not receive EVT treatment compared with male patients (Table 3). Furthermore, we identified a significant interaction between EVT and sex in mortality at discharge and 90‐day follow‐up. The interaction between sex and EVT was linked to a reduced likelihood of patient mortality at both 90 days (β = −2.64, SE = 0.75) and discharge (β = −2.47, SE = 0.72) with EVT treatment (Table 3). A similar observation was seen in the logistic interaction plot, adjusted for age and NIHSS score, which highlights the mortality difference between EVT‐treated and non‐EVT patients, with a more pronounced treatment effect on women (Figure 1).

Table 3.

Multivariable Logistic Regression

Variables	Coefficient	SE	OR	95% CI
Mortality 90 d
EVT	−0.63	0.42	0.53	[0.23–1.21]
Women	1.91	0.60	6.77	[2.09–21.94]
Interaction between EVT and sex	−2.64	0.75	0.07	[0.02–0.31]
Discharge mortality
EVT	−0.81	0.42	0.45	[0.12–1.02]
Women	1.76	0.55	5.81	[1.96–17.23]
Interaction between EVT and sex	−2.47	0.722	0.08	[0.02–0.35]
mRS score 0–3
EVT	0.81	0.50	2.24	[0.84–6.00]
Women	−1.22	0.71	0.30	[0.07–1.19]
Interaction between EVT and sex	1.08	0.85	2.93	[0.55–15.65]
mRS score 0–2
EVT	0.07	0.62	1.10	[0.32–3.61]
Women	−1.31	0.84	0.27	[0.05–1.41]
Interaction between EVT and sex	1.26	1.06	3.52	[0.44–28.26]
mRS score 0–1
EVT	−0.18	0.70	0.84	[0.21–3.33]
Women	−1.81	1.11	0.16	[0.02–1.45]
Interaction between EVT and sex	1.97	1.34	7.16	[0.52–98.76]
Intracerebral hemorrhage
EVT	0.98	0.42	2.67	[1.17–6.14]
Women	−0.54	0.44	0.58	[0.25–1.37]
Interaction between EVT and sex	0.81	0.62	2.24	[0.66–7.63]

EVT indicates endovascular thrombectomy; mRS, modified Rankin Scale; and OR, odds ratio.

Figure 1.

Sex‐differences forest plot between EVT versus non‐EVT. The forest plot illustrates the odds ratios (ORs) comparing patients who underwent EVT with those who did not (non‐EVT). In this analysis, the vertical red dashed line represents the OR value 1, signifying a favorable difference between EVT and non‐EVT treatments. EVT indicates endovascular thrombectomy; and mRS, modified Rankin Scale.

The results were similar in a secondary adjusted analysis that additionally adjusted for age and NIHSS score in multivariable logistic regression. The odds of both 90‐day mortality (OR = 4.66, CI [1.25–17.35]; P‐value = 0.022) and mortality at discharge (OR = 3.82, 95% CI [1.13–12.92]) significantly increased for women compared with men among those not treated with EVT (Table 3). Furthermore, we observed statistical significance in the interaction between EVT and sex associated with mortality at 90 days (P value = 0.004) and discharge (P value = 0.007).

After stratifying the data set by sex, we observed a greater treatment effect on mortality in women compared with men at both discharge, 90‐day follow‐up, and mRS score 0–3 following EVT treatment (Figure 2). However when stratified by sex, there was a greater likelihood of ICH in women (OR = 6.00, 95% CI [2.44–14.75]) following non‐EVT treatment compared to men (OR = 2.67, 95% CI [1.17–6.14]). Although the results did not reach statistical significance, the forest plot indicates a trend where women exhibited a greater treatment effect on mRS scores 0–2 and 0–1 at the 90‐day follow‐up following EVT treatment compared with men (Figure 1).

Figure 2.

Interaction plots between EVT and sex for mortality. The logistic interaction plot, adjusted for age and NIHSS score, illustrates a sex‐based difference in patients’ mortality probabilities at discharge (B) and during the 90‐day follow‐up period (A) based on whether they received EVT. EVT indicates endovascular thrombectomy.

Discussion

In this study of nearly 200 patients with LVO AIS and large core infarct volumes within the Greater Houston area, the mortality reduction associated with EVT was more pronounced in women compared with men. Women had a remarkably high rate of mortality without EVT (over 90%), which fell substantially by approximately 60% with treatment. A significantly smaller improvement was seen with treatment in men (∼15.3% decrease).

Prior studies have demonstrated differences in treatment and post‐EVT outcomes based on sex, influenced by factors such as baseline status, imaging results, and social implications. ⁶ In the CRISP (Computed Tomographic Perfusion to Predict Response to Recanalization in Ischemic Stroke) study, differences in infarct growth between the sexes were associated with better functional outcomes in women, and they experienced smaller final infarcts following treatment. Increased poststroke mortality in women relative to men has been observed in multiple cohort studies. A study conducted in the United States found that stroke ranked as the third leading cause of death for women and fifth for men. ⁷ , ⁸ In the same study, among the distribution of the top 10 leading causes of death among women and men, stroke accounted for 6.2% of deaths for women but only 4.3% of deaths for men. ⁸ In the Greater Cincinnati/Northern Kentucky stroke study, the percentage of stroke‐related mortality in women surpassed that in men across the entire adult lifespan. ⁴ In a global systematic review encompassing data from 19 different countries, women experienced a 5% greater mortality rate compared with men. ⁹ Other studies, however, have suggested that these differences may be largely powered by the older age of onset for stroke in women relative to men. In an Australian multicenter stroke study spanning across 39 hospitals from 2010 to 2014, initial observations indicated higher 1‐year mortality rates in women compared with men. After adjusting for factors such as age, stroke severity, and early aspirin administration, no statistically significant difference in mortality between men and women was observed after 1 year. ¹⁰ Our findings are consistent with these former studies, demonstrating that women generally experience strokes at a later age compared with men. In our secondary analysis, accounting for age and NIHSS score, the robustness of our results remained unchanged.

One potential explanation for the differential effects of reperfusion that we observe may be biologic. ¹¹ In a murine middle cerebral artery occlusion model, infarct volumes were greater in older female mice relative to young female mice, a difference attributed to decreasing levels of estrogen. ¹² However, a contrasting study in postmenopausal murine models found that estrogen treatment increased infarct volume, associated with the downregulation of insulin‐like growth factor 1, a neuroprotective regulator that decreases with age. ¹³ , ¹⁴ A similar observation was found in human studies showing an age‐related decrease in insulin‐like growth factor 1 levels coinciding with a decline in estrogen levels, which furthermore underscore the role of aging‐related biological mechanism in shaping stroke core infarct volume and emphasizing the significance of neuroprotective hormones in women. ¹⁵ , ¹⁶

Sex disparities emerge in multiple medical diseases in end‐of‐life care. Prior studies have shown increased likelihood of written advanced directives limiting care in women compared with men. ¹⁷ A study from a US‐based cohort of patients with advanced cancer found women were less likely to receive life‐prolonging medical treatments, and another study in Asia showed the same finding as well as lower rates of intubation and mechanical ventilation in women. ¹⁸ Women have been shown to have greater rates of care withholding or withdrawing and lower rates of aggressive ICU care in end‐of‐life scenarios. Similar disparities in decision making around terminal illness could have contributed to the striking differences in mortality observed between women and men in our study. ¹⁹ , ²⁰

In our study, the mortality rate in patients with untreated large core AIS was close to 80%, and >90% in women. This rate is substantially greater than those of the recently published trials and may reflect differences between clinical practice and clinical trials, in which there may be more reluctance to withdraw care. ² , ²¹ , ²² , ²³ Consistent with this concept, our observed mortality rate in EVT‐treated patients of approximately 40% is largely consistent with those of the randomized trials. Similarly, this rate is compatible with results from nonrandomized cohorts published prior to the recent randomized trials of large core EVT. A European cohort observed 73% mRS score 5–6 in a medical management arm. ²⁴ In our study, we also observed a higher rate of hypertension among women compared to men in our cohort; however, this difference did not reach statistical significance (P value = 0.067), possibly due to our small sample size. Greater rates of pre‐morbid hypertension could potentially provide insight into why women may experience a higher mortality rate in large core AIS compared to men. ²⁵ , ²⁶

This study has several limitations. Although our study underscores the existence of sex‐specific differences in mortality rate, it did not fully elucidate the underlying mechanisms driving these disparities, neither at discharge nor during the 90‐day follow‐up period. To address these gaps, in our future study, we aim to expand sample sizes for greater statistical power. Additionally, we plan to delve into societal factors that could potentially influence mortality outcomes, such as social isolation. Furthermore, exploring other biological or hospital factors not accounted for in the study could provide further insights into mortality differences between sexes. Furthermore, although multicenter, the cohort is drawn from a single region with a retrospective and nonrandomized design. Future research with multi‐institution, prospective, and randomized designs are needed to validate and extend our findings. Given the potential biases associated with a uniform physician group in a single region, this limitation may limit the generalizability of our findings.

In this study of patients with large core infarcts and LVO AIS, EVT was associated with a significant reduction in mortality in women compared with men. This finding suggests that the benefit of EVT extends more prominently to women with large core infarcts, highlighting a potential sex‐specific impact on mortality reduction in this cohort.

Sources of Funding

Dr Sheth reports funding from the National Institutes of Health (R01NS121154).

Disclosures

Dr Sheth reports grant support from the National Institutes of Health (U18EB029353) as well as Viz.AI for unrelated projects. He also reports consulting fees from Viz.AI, Penumbra, and Imperative Care outside of the submitted work. Sunil Sheth is an Associate Editor for S:VIN and was not involved in the handling or final disposition of this article. Disclosures provided by Sunil Sheth in compliance with American Heart Association's annual Journal Editor Disclosure Questionnaire are available at https://www.ahajournals.org/editor-coi-disclosures.

Supporting information

Table S1. Demographic Stratified by EVT

Table S2. Demographic Stratified by Sex (Non‐EVT Only)