Neuroimaging in Ischemic Stroke is Different between Men and Women in the DEFUSE 3 Cohort

Abstract

Background and Purpose:

Clinical deficits from ischemic stroke are more severe in women but the pathophysiological basis of this sex difference is unknown. Sex differences in core and penumbral volumes and their relation to outcome were assessed in this sub-study of the Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke (DEFUSE)-3 clinical trial.

Methods:

DEFUSE 3 randomized patients to thrombectomy or medical management who presented 6–16 hours from last known well with proximal middle cerebral artery or internal carotid artery occlusion and had target core and perfusion mismatch volumes on CT or MRI. Using univariate and adjusted regression models, the effect of sex was assessed on pre-randomization measures of core, perfusion, and mismatch volumes and hypoperfusion intensity ratio, and on core volume growth using 24-hour scans.

Results:

All patients were included in the analysis (n=182) with 90 men and 92 women. There was no sex difference in the site of baseline arterial occlusion. Adjusted by age, baseline NIHSS, baseline modified Rankin score, time to randomization, and imaging modality, women had smaller core, hypoperfusion, and penumbral volumes than men. Median (IQR) volumes for core were 8.0 mL (1.9–18.4) in women vs 12.6 mL (2.7–29.6) in men, for T_max>6 seconds 89.0 mL (63.8–131.7) vs 133.9 mL (87.0–175.4), and for mismatch 82.1mL (53.8–112.8) vs 108.2 (64.1–149.2). The hypoperfusion intensity ratio was lower in women, 0.31 (0.15–0.46) vs. 0.39 (0.26–0.57), p = 0.006, indicating better collateral circulation, which was consistent with the observed slower ischemic core growth than men within the medical group (p = 0.003).

Conclusions:

In the large vessel ischemic stroke cohort selected for DEFUSE 3, women had imaging evidence of better collateral circulation, smaller baseline core volumes, and slower ischemic core growth. These observations suggest sex differences in hemodynamic and temporal features of anterior circulation large artery occlusions.

Clinical Trial Registration:

Clinical Trial Registration-URL: http://www.clinicaltrials.gov. Unique identifier: NCT02586415

INTRODUCTION

Women present with different symptoms,^1–3 have higher lifetime risk and higher mortality from ischemic stroke than men.⁴ Disability is more common in women following an ischemic stroke and a differential response to recanalization therapy has been suggested.^5–7 Although these apparent disparities could be partially explained by uncontrolled sex-specific risk factors, physiologic differences, or selection bias, the effects on sex differences of cerebrovascular factors that predict stroke outcome and response to therapy – artery of occlusion, ischemic core volume, penumbral volume, and collateral circulation – are unknown. Controlled randomized trials of thrombectomy in stroke, selecting patients by standardized imaging as well as clinical criteria, offer a unique platform to investigate factors underlying sex differences in stroke outcomes and response to treatment.

Stroke outcomes in patients with anterior circulation large artery occlusions are improved with endovascular thrombectomy (EVT) plus standard medical therapy compared to medical therapy alone among patients with a demonstrated imaging target for thrombectomy.^{8–12,13, 14} These studies have also suggested that smaller core volumes and better collateral circulation are predictive of a greater benefit with EVT. The Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke (DEFUSE 3) trial¹³ was designed to test efficacy of endovascular therapy 6 to 16 hours after time from last known well in patients who were likely to have an ischemic penumbra and salvageable brain tissue as identified by perfusion imaging. In this study, we analyze the DEFUSE 3 sample to test whether baseline sex differences in imaging measures of cerebrovascular physiology predict clinical or physiological outcomes or a differential response to EVT.

METHODS

Data supporting findings of this study are available from the corresponding author on reasonable request.

DEFUSE 3 was a multicenter randomized trial evaluating efficacy of endovascular treatment compared to medical therapy alone for ischemic stroke patients with favorable imaging features treated in the late time-window between 6–16 hours. Ethics approval was obtained from the local institutional review board and written informed consent was obtained from patients. Full methodology and results of the study have been previously reported.¹³ In summary, volume estimates of ischemic core and penumbral regions from CT perfusion or MRI diffusion and perfusion scans were automatically calculated using RAPID software (iSchemaView). The penumbra size was estimated from the volume of tissue for which there was delayed arrival of an injected contrast agent (time to maximum of the residue function [Tmax]) greater than six seconds.

According to literature, intracranial volume is smaller for women on average by 10%.¹⁵ In order to include this difference in our analysis, results have also been adjusted accordingly (see supplemental data).

Outcomes:

In this sub-study we stratified study participants by sex and compared baseline and clinical characteristics as well as clinical and imaging outcomes. The clinical outcomes of interest were functional status assessed on the full modified Rankin Scale (mRS) at 90-days, functional independence defined as mRS ≤ 2, and poor outcome defined as severe disability (mRS = 5) or death. Imaging outcomes of interest were infarct volume, infarct growth and reperfusion (change in hypoperfusion lesion defined by T_max > 6 seconds delay) achieved by 24 hours post-randomization.

Statistical Analysis:

We compared baseline and clinical characteristics between women and men enrolled. Outcomes were compared between women and men within the treatment groups as well as effect of the treatment for each sex. We compared continuous and ordinal variables using Mann-Whitney U test, and utilized χ² test (exact test where necessary) to compare proportions and rates. To compare effects of treatment by sex in not adjusted analysis, we used Breslow-Day test for the binary outcomes, while we assessed interaction factor in the regression analysis for the continuous or ordinal outcomes for both not-adjusted and adjusted analyses. To adjust results for other significant factors and variables we employed logistic, ordinal, or general linear regression analyses. We adjusted for variables known to us from the main and other sub-study analyses to be independent predictors of outcomes (age, NIHSS at presentation, serum glucose level) and the baseline imbalances between female and male participants in the study. We employed backward elimination procedure to keep in the final models the variables still significant when adjusted for other factors. All statistical tests were 2-sided and significance was defined at α < 0.05. Statistical analysis was done with SAS 9.4 (SAS Institute Inc., Cary, NC) and IBM SPSS Statistics 25 (IBM, Armonk, NY).

RESULTS

Patient Characteristics

Amongst 182 patients enrolled in the DEFUSE 3 trial, 92 were women (50.5%) and 90 were men (49.5%), women represented 50% (46/92) of the population in the endovascular therapy and medical treatment arm and 51% (46/90) of those only receiving medical treatment. Comparing both sexes (Table 1), women were older by median 3 years. There were no significant differences between men and women in baseline NIHSS score (16 vs. 16, p = 0.760) or pre-stroke mRS (0 vs. 0, p = 0.539).

Table 1.

Baseline characteristics of DEFUSE 3 stratified by sex^✩

Characteristic	Women N=92	Men N=90	p
Age – yr	72 (62–81)	69 (59–78)	0.126
0–50	9 (10%)	5 (6%)	0.037
51–60	12 (13%)	25 (28%)
61+	71 (77%)	60 (67%)
NIHSS admit†	16 (11–20)	16 (12–21)	0.760
mRS baseline‡	0 (0–0)	0 (0–0)	0.539
Stroke onset witnessed	36 (39%)	30 (33%)	0.416
Symptoms present on awakening	41/56 (73%)	50/60 (83%)	0.185
Process measures hr:min
Symptom onset to arrival at enrolling hospital	9:39 (7:27–11:35)	8:49 (6:48–10:00)	0.019
Symptom onset to imaging initiation	10:45 (8:11–12:31)	9:42 (7:57–11:21)	0.032
Time to randomization	11:35 (8:38–12:58)	10:32 (8:45–11:48)	0.050
Symptom onset to femoral puncture (EVT group, N=37 & 32)	12:13 (10:25–13:16)	10:54 (9:07–12:15)	0.026
Imaging to femoral puncture	0:59 (0:39–1:29)	0:59 (0:41–1:26)	0.670
Time to recanalization (EVT group, N=37 & 32)	13:05 (11:22–14:20)	11:34 (9:01–13:08)	0.054
Imaging characteristics			-
Imaging modality			0.080
CT perfusion imaging	62 (67%)	71 (79%)
Diffusion and perfusion MRI	30 (33%)	19 (21%)
Baseline imaging measures
Core volume – mL	8.0 (1.9–18.4)	12.6 (2.7–29.6)	0.087
Perfusion lesion volume – mL§	89 (64–132)	134 (87–175)	<0.001
Penumbral volume (perfusion lesion - ischemic core) – mL	82.1 (53.8–112.8)	108.2 (64.1–149.2)	0.001
Mismatch ratio¶	11.5 (5.3–59.0)	9.7 (4.1–42.2)	0.426
Baseline vessel occluded			0.530
Proximal internal carotid artery	2 (2%)	3 (3%)
Distal internal carotid artery or T lesion	28 (30%)	35 (39%)
Middle cerebral artery, M1 branch	61 (66%)	52 (58%)
Middle cerebral artery, M2 branch	1 (1%)	0 (0%)
ASPECTS on CT (median, IQR)∥	8 (7–9)	8 (7–9)	0.620
Collateral status (Tan score = 1)✩✩	54 (75%)	51 (67%)	0.290
Hypoperfusion intensity ratio (HIR)	0.31 (0.15–0.46)	0.39 (0.26–0.57)	0.006

^✩Values are medians followed by interquartile range (IQR) in parentheses. Frequencies are number followed by percentage (%).

^†Scores on the National Institutes of Health Stroke Scale (NIHSS) range from 0 to 42, with higher scores indicating a higher deficit.

^‡Scores on the modified Rankin Scale (mRS) range from 0 to 6, with higher score indicating higher disability.

^§Shown is the volume of tissue for which there was delayed arrival of injected contrast agent exceeding 6 seconds (T_max > 6 s).

^¶Mismatch ratio is the volume of ischemic tissue on perfusion imaging to infarct volume.

^∥Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is a tool that is used to estimate the volume of infarcted tissue. Scores range from 0 to 10, with lower scores indicating a larger area.

^✩✩The Tan score is used to assess collateral status. Showing the number of patients with Tan score of 1, where the collateral supply is filling > 50% of the vascular territory distal to the occluded artery.

^††Hypoperfusion intensity ratio (HIR) was defined as the proportion of T_max > 6 s lesion volume with a T_max >10 s delay, with an increased HIR reflecting poor collaterals.

Overall, women had significantly longer outside-of-hospital process measures than men captured as time from symptom onset to: hospital arrival (p = 0.019), imaging initiation (p = 0.032), randomization (p = 0.050), and femoral puncture (p = 0.026). While within-hospital process times from imaging to femoral puncture were not different between men and women (p = 0.670). There were no significant differences in strokes being witnessed between women and men (39% vs. 33%, p = 0.416) or wake up strokes (73% vs. 83% of unwitnessed strokes, p = 0.185).

Baseline imaging measures differed for perfusion lesion volume, with women demonstrating smaller median [IQR] perfusion deficits than men as defined by T_max delay threshold of 6 seconds (89.0 mL [63.8–131.7] vs. 133.9 mL [87.0–175.4], p < 0.001). Women also tended to have smaller median (IQR) ischemic core volumes at randomization (8.0 mL [1.9–18.4]) than men (12.6 mL [2.7–29.6]), p = 0.087, as defined by region of reduced cerebral flow < 30% of that in normal tissue, when assessed by CT perfusion, or by the apparent diffusion coefficient (ADC) threshold of 620 × 10⁻⁶ mm²/s. The median [IQR] penumbral volume, defined as the difference between perfusion deficit volume and ischemic core volume, was smaller in women than men at baseline (82.1 mL [53.8–112.8] vs. 108.2 mL [64.1–149.2], p = 0.001).

The median [IQR] measure of collateral circulation function, the hypoperfusion intensity ratio (HIR), was lower in women (0.31 [0.15–0.46] vs. 0.39 [0.26–0.57], p = 0.006), indicating better collateral function. This measure of collateral function is related to the Tan score for collateral status within this DEFUSE 3 cohort.¹⁶ No differences were found between men and women for baseline vessel occluded, ASPECTS on CT, or collateral status as measured by Tan score.¹⁷

Sex and Baseline Imaging Measures

Sex was found to be associated with baseline imaging measures upon adjustment by age, NIHSS score at randomization, baseline mRS score, time to randomization, and imaging modality (CT or MRI). Specifically, female sex was associated with a smaller baseline perfusion deficit volume (p < 0.001), smaller penumbral volume (p < 0.001), and more favorable collateral function as measured by the HIR (p = 0.004). Sex was also approaching significance in association with acute ischemic core volume (p = 0.070).

Sex and 24-hour Imaging Measures

Penumbra salvage volumes measured at 24 hours were compared between sexes with women demonstrating a significantly smaller ischemic core growth (median [IQR]) (21.9 mL [10.7–36.7] vs. 41.9 mL [15.7–103.4], p < 0.001). Reperfusion, as measured by perfusion deficit volume decrease, was not different between women and men (76.4 mL [47.5–108.8] vs. 86.9 mL [43.7–124.7], p = 0.407).

Within Sex Differences based on Treatment Arms

Differences in baseline clinical features such as pre-stroke mRS and NIHSS at randomization were not significant between treatment arms within each sex. Baseline imaging measures were uniform with the exception of the ASPECTS score. The median ASPECTS score for women in the EVT plus medical therapy group was significantly higher than that of the medical control group (8 vs. 7, p = 0.006).

Baseline Sex Differences within Treatment Arms

The baseline characteristics within each treatment arm were well balanced between the sexes with only a few differences, as seen in Tables 2 and 3. A lower percentage of women in the EVT arm with unwitnessed event experienced wakeup strokes than men (70% vs. 90%, p = 0.042). Women within the EVT treatment arm also experienced a longer time from symptom onset to randomization than men (11 hours, 43 mins. vs. 10 hours, 9 mins, p = 0.019). Those women within the medical treatment arm were more likely to undergo a CT than an MRI as baseline imaging with 84% of women and 59% of men getting CT (p = 0.008).

Table 2.

Baseline characteristics of the patient features based on sex^✩

Characteristic	Women			Men
	EVT+Medical N=46	Medical Alone N=46	p	EVT+Medical N=46	Medical Alone N=44	p
Age – yr	73 (65–80)	71 (59–82)	0.522	68 (58–76)	72 (60–79)	0.450
0–50	4 (9%)	5 (11%)	0.767	3 (7%)	2 (5%)	0.717
51–60	5 (11%)	7 (15%)		14 (30%)	11 (25%)
61+	37 (80%)	34 (74%)		29 (63%)	31 (70%)
NIHSS admit†	16 (10–20)	15 (12–20)	0.521	15.5 (11–21)	17.5 (12–21.5)	0.484
mRS baseline‡	0 (0–0)	0 (0–0)	0.560	0 (0–0)	0 (0–0)	0.057
Stroke onset witnessed	16 (35%)	20 (44%)	0.393	15 (33%)	15 (34%)	0.882
Wakeup stroke (unwitnessed)	21/30 (70%)	20/26 (77%)	0.560	28/31 (90%)	22/29 (76%)	0.175
Process measures hr:min
Time to randomization	11:43 (9:34–12:53)	10:57 (8:22–13:08)	0.587	10:09 (8:11–11:48)	10:42 (9:00–12:12)	0.355
Imaging characteristics
Baseline imaging measures
Core volume – mL	7.9 (2.0–22.9)	8.4 (1.8–17.8)	0.919	12.8 (2.7–26.9)	12.3 (3.3–29.9)	0.997
Perfusion lesion volume – mL§	89 (69–129)	100 (60–157)	0.984	134 (96–177)	134 (84–174)	0.878
Penumbral volume (perfusion lesion-ischemic core) – mL	80 (62–112)	88 (51–117)	0.859	109 (73–154)	105 (62–145)	0.707
Mismatch ratio¶	11.8 (5.2–54.9)	10.9 (5.4–5029)	0.852	11.1 (4.0–30.8)	9.3 (4.6–48.0)	0.868
ASPECTS on CT (median, IQR)∥	8 (8–9)	7 (7–8)	0.006	7 (7–8)	8 (7–9)	0.107
Collateral status (Tan score = 1)✩✩	27 (71%)	27 (79%)	0.414	25 (69%)	26 (65%)	0.681
Hypoperfusion intensity ratio (HIR)††	0.29 (0.11–0.43)	0.33 (0.16–0.48)	0.477	0.40 (0.23–0.57)	0.37 (0.27–0.58)	0.875

^✩Values are medians followed by interquartile range (IQR) in parentheses. Frequencies are number followed by percentage (%)

^†Scores on the National Institutes of Health Stroke Scale (NIHSS) range from 0 to 42, with higher scores indicating a higher deficit.

^‡Scores on the modified Rankin Scale (mRS) range from 0 to 6, with higher score indicating higher deficit.

^§Shown is the volume of tissue for which there was delayed arrival of injected contrast agent exceeding 6 seconds (Tmax > 6 s)

^¶Mismatch ratio is the volume of ischemic tissue on perfusion imaging to infarct volume

^∥Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is a tool that is used to estimate the volume of infarcted tissue. Scores range from 0 to 10, with lower scores indicating a larger area.

^✩✩The Tan score is used to assess collateral status. Showing the number of patients with Tan score of 1, where the collateral supply is filling ≤ 50% but > 0%.

^††Hypoperfusion intensity ratio (HIR) was defined as the proportion of Tmax >6 s lesion volume with a Tmax >10 s delay, with and increased HIR reflecting poor collaterals.

Table 3.

Baseline characteristics of the patient features based on treatment^✩

Characteristic	EVT+Medical			Medical Alone
	Women N=46	Men N=46	p	Women N=46	Men N=44	p
Imaging characteristics
Imaging modality			0.810			0.008
CT perfusion imaging	35 (76%)	34 (74%)		37 (84%)	27 (59%)
Diffusion and perfusion MRI	11 (24%)	12 (26%)		7 (16%)	19 (41%)
Baseline imaging measures
Core volume – mL	7.9 (2.0–22.9)	12.8 (2.7–26.9)	0.183	8.4 (1.8–17.8)	12.3 (3.3–29.9)	0.274
Perfusion lesion volume – mL†	88.9 (69.2–128.8)	134.0 (95.7–176.9)	0.003	100.1 (60.0–156.5)	133.9 (83.6–173.5)	0.015
Penumbral volume (perfusion lesion - ischemic core) – mL	79.7 (62.4–111.7)	109.3 (72.6–154.1)	0.009	88.0 (51.4–117.2)	105.4 (61.7–145.0)	0.047
Mismatch ratio‡	11.8 (5.2–54.9)	11.1 (4.0–30.8)	0.530	10.9 (5.4–5029)	9.3 (4.6–48.0)	0.618
ASPECTS on CT (median, IQR)§	8 (8–9)	7 (7–8)	0.017	7 (7–8)	8 (7–9)	0.053
Collateral status (Tan score = 1)¶	27 (71%)	25 (69%)	0.880	27 (79%)	26 (65%)	0.171
Hypoperfusion intensity ratio (HIR)∥	0.29 (0.11–0.43)	0.40 (0.23–0.57)	0.029	0.33 (0.16–0.48)	0.37 (0.27–0.58)	0.044

^✩Values are medians followed by interquartile range (IQR) in parentheses. Frequencies are number followed by percentage (%)

^†Shown is the volume of tissue for which there was delayed arrival of injected contrast agent exceeding 6 seconds (Tmax > 6 s)

^‡Mismatch ratio is the volume of ischemic tissue on perfusion imaging to infarct volume

^§Alberta Stroke Program Early Computed Tomography Score (ASPECTS) is a tool that is used to estimate the volume of infarcted tissue. Scores range from 0 to 10, with lower scores indicating a larger area.

^¶The Tan score is used to assess collateral status. Showing the number of patients with Tan score of 1, where the collateral supply is filling ≤ 50% but > 0%.

^∥Hypoperfusion intensity ratio (HIR) was defined as the proportion of Tmax >6 s lesion volume with a Tmax >10 s delay, with and increased HIR reflecting poor collaterals.

In regards to baseline imaging measures, women within both treatment arms demonstrated significantly smaller perfusion lesion volumes (EVT: p = 0.003, Medical: p = 0.015), smaller penumbral volumes (EVT: p = 0.009, Medical: p = 0.047), and better collateral function as measured by HIR (EVT: p = 0.029, Medical: p = 0.044) than men.

Sex Differences in Outcomes within Treatment Arms

Overall, there were no significant sex by treatment interactions. Of patients included, women within the EVT treatment arm had worse mRS at 90 days than men (mRS median [IQR], 3.5 [2–5] vs. 2 [1–4], p = 0.046) despite similar imaging outcomes at 24-hours as well as similar NIHSS (p = 0.555) and mRS (p = 0.135) at discharge (Table 4). This potential sex difference was persistent even with patients achieving recanalization as indicated by a TICI score of 2b or 3 (p = 0.045 for sex difference). Within this group of patients receiving EVT intervention and achieving recanalization, less women than men seemed to reach functional independence at 90 days (38% vs. 67%, p = 0.016).

Table 4.

Imaging outcomes based on treatment^✩

Characteristic	EVT + Medical			EVT + Medical: Achieving TICI 2b-3			Medical Alone
	Women N=46	Men N=46	p	Women N=37	Men N=33	p	Women N=46	Men N=44	p
Ischemic core growth – mL	20.0 (6.1–33.4)	37.1 (13.2–103.4)	0.064	16.2 (5.2–33.4)	36.2 (13.2–74.5)	0.209	23.7 (13.2–42.0)	47.6 (25.3–106.2)	0.003
Perfusion deficit volume decrease (reperfusion)	88.9 (66.5–129.2)	108.9 (86.3–145.0)	0.177	91.4 (77.0–129.2)	119.9 (86.9–176.9)	0.055	54.5 (40.4–88.0)	47.7 (8.7–94.2)	0.367

^✩Values are medians followed by interquartile range (IQR) in parentheses

Assessment of the medical treatment group showed an absence of sex differences in both ordinal mRS and functional independence at 90 days despite women demonstrating slower ischemic core growth between baseline and 24 hours than men (23.7 mL [13.2–42.0] vs. 47.6 mL [25.3–106.2], p = 0.003) with medical treatment alone.

DISCUSSION

Advanced neuroimaging provides a reflection of tissue status and can be used to individualize therapeutic decisions according to the unique evolution of patient’s ischemic tissue. We have examined sex as a biological variable in the DEFUSE 3 trial with a focus on neuroimaging measures. We showed that sex differences exist in baseline imaging characteristics, process measures, clinical outcomes, and imaging outcomes within this group of patients with a large vessel occlusion selected based on a favorable imaging profile, functional independence prior to stroke, and presentation between 6–16 hours from last known well.

Baseline imaging features were different between sexes with women presenting with smaller perfusion deficits and smaller penumbra volumes. Women also demonstrated better collateral function when compared to men, which supports the smaller volumes of hypoperfusion. Our findings of superior collateral function in women are in contrast with previous preclinical structural and functional evaluation of pial collaterals in Wistar rats¹⁸ and mice.¹⁹ It is also relevant to consider that women within the DEFUSE 3 cohort had longer symptom onset to arrival to hospital times and consequently later imaging for women while within hospital processes were equivalent between sexes.

The primary efficacy outcome was met in the DEFUSE 3 trial; EVT plus standard medical therapy led to a more favorable distribution of disability scores on the modified Rankin scale at 90 days compared to standard medical therapy alone.¹³ Analysis of DEFUSE 3 data indicates there is no sex difference for treatment effect upon ordinal analysis of the mRS at 90-days (OR [95% CI], men: 3.7 [1.7–7.8], women: 2.7 [1.3–5.7], p = 0.557), although the difference in mRS distributions at 90 days between the EVT and medical groups for women was small (p = 0.0504), no significant interaction was found between sex and treatment for efficacy. Of note, sex-based analysis of the randomized clinical trial: intra-arterial therapy (IAT) for acute ischemic stroke in the Netherlands (MR CLEAN),⁸ demonstrated an interaction between sex and IAT for efficacy. The treatment effect for men was demonstrated by a significant improvement with IAT but a beneficial treatment effect was not found for women.²⁰ On the other hand, several other studies report no sex differences in outcome recanalization or functional outcomes after EVT, as was found in the current analysis of the DEFUSE 3 cohort. Although baseline characteristics were not initially reported according to sex in these cited studies, which complicated interpretation of results,^{10, 12, 21, 22} recent analyses of the HERMES²³ and SWIFT, STAR, and SWIFT PRIME²³ data focused on sex as a variable. Chalos, et al. did not find an interaction of sex and endovascular treatment effect within their cohort (n=1762) although higher baseline collateral score and smaller final infarct volumes were found in women. Analysis of patients treated with the Solitaire stent retriever by Sheth et al. did not find radiologic differences between men and women but did find that women had more years of disability-adjusted life years after thrombectomy.²⁴

The secondary outcome for the DEFUSE 3 trial was also met, with a higher percentage of patients in the EVT therapy group achieving functional independence at 90 days when compared with medical therapy alone.¹³ This efficacy of EVT in achieving functional independence was experienced by both men and women with the adjusted odds ratio for women achieving functional independence being 7.6 (95% CI 3.1–18.6) and that of men being 8.3 (95% CI 2.4–28.2), p = 0.927 for the difference between these treatment odds ratios. We found a difference in the likelihood of functional independence between sexes in the DEFUSE 3 cohort with 16/46 women (35%) and 25/46 men (54%) receiving EVT being functionally independent (mRS ≤ 2) at 90 days, p = 0.059 for difference, while not statistically significant, may represent a clinically-important difference. Further adjustment for age, NIHSS at randomization and blood glucose resulted in an adjusted OR (95%CI) of 2.6 (0.93–7.5) for men vs women achieving functional independence. These results correspond with a recent prospective study with females being less likely to achieve independence at 90 days.²⁵

The imaging-based outcome infarct growth was slower in women at 24 hours in the DEFUSE 3 cohort as a whole and specifically within the medical group. This slower growth could be explained by better collateral function, as measured by the HIR, in women, a sex difference that was present in both treatment arms. Additionally, women had smaller perfusion deficit and smaller penumbral volumes at randomization translating into less potential for penumbral salvage. Despite better collateral function and smaller perfusion deficits, however, women did not experience better outcomes. Additional analyses incorporating the known difference in intracranial volume¹⁵ are included in the supplementary data and this normalization did not change the presented results. It is possible that these collaterals eventually fail to support the hypoperfused region resulting in an eventual increase in the perfusion deficit volume and eventual infarction, albeit delayed compared men. It is possible that the final infarct volume is equal between the sexes, but without a larger cohort and/or imaging measures within the subacute period, the DEFUSE 3 cohort is unable to address this issue and further study is warranted with appropriate serial imaging data.

A comprehensive pooled analysis of recent EVT trials (HERMES collaboration) did not find a sex difference in revascularization as reflected by proportions of men and women achieving TICI 2b/3 status after EVT.²⁶ The present evaluation of the DEFUSE 3 cohort also found no difference in proportions of patients achieving revascularization defined by TICI 2b/3 (78% women vs. 70% men, p = 0.407). There was a trend for women to have a longer time to recanalization, although not significant, p = 0.054. Interestingly, no differences in neuro-imaging outcomes (mTICI score and CT final infarct volumes at 5–7 days) were found in secondary analysis of MR CLEAN.²⁰ Conversely, a recent analysis of the Recovery by Endovascular Salvage for Cerebral Ultra-Acute Embolism)-Japan Registry 2 patient database²⁷ found that women with acute ischemic stroke with large vessel occlusion showed poor functional outcome compared to men. This sex difference was partially explained by a lower proportion of women receiving EVT.

We further examined baseline characteristics of the EVT group of DEFUSE 3 for imbalance between men and women. Women presented with more favorable ASPECTS scores than men (8 [8–9] vs. 7 [7–8], p = 0.017). The ASPECTS score has been used for patient selection in some of the recent large stent retriever thrombectomy trials with a recent meta-analysis including a central reading of all pre-treatment scans from these five trials demonstrating a clear benefit of thrombectomy in patients with ASPECTS >5.²⁶ Although evidence of effect modification has been found between favorable ASPECTS (score 8–10) and good clinical outcome with intra-arterial thrombolysis,^{28, 29} ASPECTS was not able to identify subjects that would benefit from endovascular therapy in a recent analysis of the IMS-III trial.³⁰

As discussed, several studies have assessed sex differences in EVT therapy with only that by Uchida, et al.²⁷ designed to specifically address sex differences and did not have adjustments made for factors that might affect treatment outcomes, not enough patients to power sex-based studies, or not enough patients included that are over 75 years of age. Studies addressing sex as a modifier of outcome have been inconclusive.^31–33 We have demonstrated baseline differences between men and women with women having smaller infarct cores and perfusion deficits. Women in DEFUSE 3 did not achieve functional independence more frequently than men did although the women had better collateral function and slower infarct growth. These results demonstrate the need to further assess the impact of sex on stroke presentation evolution in future trials.

SUMMARY/CONCLUSIONS

Sex differences exist in presentation and evolution of imaging measures within the imaging-selected acute ischemic stroke population cohort for DEFUSE 3. Women presented with smaller ischemic cores, smaller perfusion deficit volumes, smaller mismatch volume, and better collateral function. Women had a longer time to randomization resulting in more time for tissue infarct evolution prior to baseline imaging. Though sex is a factor associated with outcomes, no interaction between sex and treatment was found for efficacy. Follow up imaging indicated slower ischemic core growth in women compared to men within the medical treatment alone group supporting better collateral function in women. These findings suggest sex differences in hemodynamic and temporal features of anterior circulation large artery occlusions.