Introduction
Patient sex plays an important role not only in acute ischemic stroke (AIS) pathophysiology but also in acute stroke care. Reference Ospel, Singh, Ganesh and Goyal1,Reference Ospel, Schaafsma and Leslie-Mazwi2 Women are on average older Reference Appelros, Stegmayr and Terent3 and show a higher degree of pre-stroke disability at the time of their first stroke. Reference Willers, Lekander and Ekstrand4–Reference Caso, Paciaroni and Agnelli8 Women are also more likely to live alone at the time of stroke onset, and thus strokes in women may occur more often unwitnessed, get discovered later, and are associated with problems when obtaining consent to treatment. Reference Ospel, Schaafsma and Leslie-Mazwi2,Reference Mainz, Andersen, Valentin, Gude and Johnsen6 Delays in hospital admission, diagnosis, and treatment may be more pronounced in female stroke patients as compared to their male counterparts. Reference Hemmen9,Reference Naveed, Almasri and Kazani10 This may lead to worse outcomes in female stroke patients undergoing endovascular treatment (EVT) compared to their male counterparts, Reference Ospel, Schaafsma and Leslie-Mazwi2,Reference Deb-Chatterji, Schlemm and Flottmann11,Reference Dmytriw, Ku and Yang12 although the literature is controversial in this regard. Reference Ospel, Singh, Ganesh and Goyal1,Reference Ospel, Schaafsma and Leslie-Mazwi2
We used a comprehensive Canadian provincial dataset to compare sex differences with regard to patient baseline characteristics, treatment time metrics, and clinical outcomes in AIS patients undergoing EVT over a 5-year time period, between January 2017 and December 2022.
Methods
Patient Study Population: This study included all patients who underwent EVT between January 2017 and December 2022 in the province of Saskatchewan, Canada, with a population of 1.2 million. 13 All EVT procedures in the province occur at one comprehensive stroke center (CSC) (Royal University Hospital) because this is the only EVT-capable center in the province.
Data collection and Curation: Clinical data of all AIS patients who underwent EVT in the 5-year period in the province of Saskatchewan are captured in the Canadian OPTIMISE registry, a national registry supported by the Canadian Stroke Consortium to ensure quality control for endovascular therapy and intravenous thrombolysis among patients with AIS. Reference Sajobi, Singh and Almekhlafi14 Data for the OPTIMISE registry are captured using a web-based electronic data capture and reporting system and housed on a secure server at the Population Health Research Institute at McMaster University. These data included patient age, sex, baseline National Institutes of Health Stroke Scale, baseline Alberta Stroke Program Early CT Score, occlusion location, collateral score, time of stroke onset, witnessed vs. unwitnessed stroke onset, transport mode (direct to CSC vs. transfer from primary stroke center [PSC] vs. in-hospital stroke), intravenous thrombolysis, and detailed procedural EVT data (arterial access time, reperfusion time, final expanded Thrombolysis in Cerebral Infarction score as assessed by the operator). OPTIMISE data were supplemented with patient data from administrative data sources (patient homes’ postal code, which were derived from patients’ electronic medical records and used as a proxy for patient location at the time of stroke onset), time of imaging at the primary stroke center, and CSC and last intracranial run time [used as a surrogate for reperfusion time in case no expanded Thrombolysis in Cerebral Infarction Score 2b/3 reperfusion was achieved/documented]).
Outcomes of Interest
The co-primary outcomes of interest were
a) the time from last known well (LKW) to CSC arrival (capturing the entire pre-hospital and transfer phase prior to CSC arrival), derived from the OPTIMISE database
b) the time from CSC arrival to reperfusion (capturing the entire CSC in-hospital phase), CSC arrival time, reperfusion time (or last intracranial run time in case no reperfusion was achieved/documented) derived from the local Picture Archiving and Communications System
c) final expanded Thrombolysis in Cerebral Infarction score (capturing technical EVT success).
The secondary outcome was modified Rankin Score.
Statistical analysis: Patient baseline characteristics, workflow times, and technical and clinical outcomes for the overall patient sample and for women and men were reported using medians and interquartile ranges for continuous data and counts and frequencies for categorical data. Univariable comparisons of the above variables between female and male stroke patients were made using Fisher’s exact test for categorical variables and Wilcoxon rank sum test for continuous variables. Multivariable linear regression (LKW to CSC arrival time, CSC arrival to reperfusion time) or ordinal logistic regression (final expanded Thrombolysis in Cerebral Infarction Score, 90-day modified Rankin Score) with adjustment for age, baseline National Institutes of Health Stroke Scale and baseline Alberta Stroke Program CT Score was used to obtain adjusted effect size estimates with respective 95% confidence intervals for patient sex on outcomes. For linear regression, normal distribution of residuals was assessed visually and tested with the Shapiro–Francia–Wilk test. In case the residuals of dependent variables in linear regression were not normally distributed, logarithmic transformation of variables was attempted. The parallel regression assumption for proportional odds (i.e., ordinal logistic) regression was tested with the Brant test. Since missing data were few and balanced between sexes, and because we could not assume with certainty that data were completely missing at random, we decided against imputation for missing data.
All analyses were conducted using Stata 17.0 (Stata, LLC Corp.), and p-values < 0.05 were considered statistically significant. Visualizations were generated using Microsoft PowerBI (Version 2.91).
This study was approved by the ethics committee of the Royal University Hospital, Saskatoon (REB 23-06). The data that were used for analysis will be made available by the corresponding author upon reasonable request.
Results
Between January 2017 and December 2022, 303 patients underwent EVT in the province of Saskatchewan at the Royal University Hospital of Saskatchewan in Saskatoon, of which 144 (47.5%) were women and 159 (52.5%) were men (Fig. 1). Baseline characteristics of women and men did not differ significantly, except for age at stroke onset, which was higher in women, and intracranial occlusion location on baseline imaging, whereby intracranial internal carotid artery occlusions occurred more commonly in men and M1 segment occlusions were more frequent in women (Table 1).
Figure 1: Distribution of women and men with acute ischemic stroke across the province who underwent EVT between January 2017 and December 2022. Graph at the bottom illustrates patient age by sex, with a median age of 71 in men and 77.5 in women who underwent EVT. Note that not all dots may be seen individually since they may overlap.
Table 1: Baseline characteristics of the entire patient sample and of women vs. men
ASPECTS = Alberta stroke program early CT score; CSC = comprehensive stroke center; IQR = interquartile range; ICA = internal carotid artery; NIHSS = national institutes of health stroke scale; PSC = primary stroke center.
Note that “n” in each cell indicates the overall number of patients for which a particular variable was available.
* P-values for comparison of women vs. men derived from Wilcoxon rank sum test (continuous variables)/ Fisher’s exact test (categorical variables).
Comparison of Co-primary Outcomes in Female vs. Male Stroke Patients
In unadjusted analyses, LKW to CSC arrival times, CSC arrival to reperfusion times, and final expanded Thrombolysis in Cerebral Infarction Score did not differ between women and men (Table 2, Fig. 2).
Figure 2: Time from last known well to CSC arrival in min (A), time from CSC arrival to reperfusion in min (B), and final expanded thrombolysis in cerebral infarction score (C) in women vs. men who underwent EVT. White dots in the box plots represent means. Numbers in the bars in (C) represent patient numbers.
Table 2: Unadjusted comparison of LKW to CSC arrival times, CSC arrival to reperfusion times, and final eTICI between women and men
CSC = comprehensive stroke center; eTICI = expanded thrombolysis in cerebral infarction; IQR = interquartile range; LKW = last known well.
Note that “n” in each cell indicates the overall number of patients for which a particular variable was available.
* P-values for comparison of women vs. men derived from Wilcoxon rank sum test (time variables)/ Fisher’s exact test (final eTICI).
** In case no reperfusion could be achieved (eTICI 0-2a), the last intracranial run time was used as a surrogate for reperfusion time.
For the adjusted analysis, time from LKW to CSC arrival and time from CSC arrival to reperfusion were log-transformed to achieve a normal distribution of residuals. Log-transformed time variables resembled a normal distribution more closely compared to the non-log-transformed time variables (Suppl. Figures I–III). While the Shapiro–Francia–Wilk test indicated that the normality distribution assumption was still violated for time from LKW to CSC arrival (p < 0.001), there was no evidence that the normality distribution assumption for CSC arrival to reperfusion time was violated (p = 0.156).
Female sex was associated neither with log-transformed LKW to CSC arrival time (beta-coefficient 0.11 [95% CI –0.13–0.35]) nor with CSC arrival to reperfusion time (beta-coefficient 0.01 [95% CI –0.10–0.13]).
Female sex was not associated with final expanded Thrombolysis in Cerebral Infarction Score (adjusted common odds ratio 1.15 [95% CI 0.74–1.81]). The Brant test indicated that the proportional odds assumption was not violated (p = 0.184).
Comparison of Secondary Outcomes in Female vs. Male Stroke Patients
In univariable analysis, mRS at 90 days did not differ significantly between women and men (Suppl. Table I, p = 0.715). In multivariable analysis (after adjustment for age, baseline National Institutes of Health Stroke Scale and baseline Alberta Stroke Program Early CT Score), female sex was not associated with 90-day modified Rankin Score (adjusted common odds ratio 1.15 [95%CI 0.75–1.77], p[Brant] = 0.768).
Discussion
In this Canadian province-wide study that included patients undergoing EVT over a 5-year time period, female EVT patients were on average older at the time of stroke onset, but no sex differences in EVT time metrics in the pre-hospital and in-hospital phase, reperfusion quality, or clinical outcomes were seen.
Numerous studies have found that women are on average 4–5 years older than men when they suffer their first stroke, Reference Appelros, Stegmayr and Terent3 and our study confirmed this “age discrepancy.” Prior research has also shown that women are more commonly disabled at the time of their first stroke, are more likely to be widowed, and live alone more often. Reference Willers, Lekander and Ekstrand4–Reference Caso, Paciaroni and Agnelli8 Furthermore, strokes in women may occur more often unwitnessed, and by the time the patient’s symptoms are discovered and emergency services are called, the window of opportunity for intravenous thrombolysis and/or EVT may have already passed. Reference Mainz, Andersen, Valentin, Gude and Johnsen6
Prior studies on the topic show conflicting results on whether sex differences in post-EVT outcomes exist or not: many studies did not find any sex differences in functional outcomes after EVT. Reference Bala, Casetta and Nannoni15,Reference Ouyang, Shajahan and Liu16 As an example, similar to our study, Sheth et al. report that women, despite being on average older at stroke onset and having higher rates of atrial fibrillation, achieve similar functional outcomes compared to men. Reference Sheth, Lee and Warach17 On the other hand, one large meta-analysis by Dmytriw et al. Reference Dmytriw, Ku and Yang12 and a nationwide registry study from Germany Reference Deb-Chatterji, Schlemm and Flottmann11 found worse post-EVT outcomes in women, whereby the latter study showed that this discrepancy is mostly, if not completely, explained by confounding factors such as older age at stroke onset.
However, stroke care often does differ between women and men: in some studies, women were 13% less likely to receive intravenous thrombolysis than men in Western countries, Reference Ospel, Schaafsma and Leslie-Mazwi2,Reference Foerch, Misselwitz and Humpich18 particularly in hospitals without specialized stroke unit care. Reference Perez-Sanchez, Barragan-Prieto and Ortega-Quintanilla19 With regard to EVT, some studies reported undertreatment of women in the acute phase, while others suggested the opposite. Reference Mainz, Andersen, Valentin, Gude and Johnsen6,Reference Uchida, Yoshimura, Sakai, Yamagami and Morimoto7,Reference Smith, Saver and Cox20 Due to their worse pre-stroke functional status, female AIS patients may also be at a greater risk of “withdrawal bias” during the acute and post-acute phase of stroke, which may aggravate undertreatment even further in a vicious cycle. It is currently not entirely clear to what extent outcomes in female EVT patients can attributed to their baseline status vs. undertreatment and delayed treatment. Reference Holloway, Arnold and Creutzfeldt21
One reason for the similar outcomes between women and men undergoing EVT in Saskatoon, despite the slightly older age of women, may be that Saskatchewan has a uniform acute stroke pathway, whereby patients are directly triaged to stroke centers (PSC or CSC) if an acute stroke is suspected by the pre-hospital team, and hospitals that neither provide thrombolysis nor EVT are bypassed. Although one would intuitively think that this is the case in most jurisdictions, this workflow is not explicitly established in some geographies.
Additionally, some studies in which women suffered worse 90-day functional post-EVT outcomes than men were conducted in jurisdictions with private healthcare systems, Reference Demel, Reeves and Xu22,Reference Dmytriw, Ku and Yang23 where access to stroke rehabilitation programs depends on the individual patient’s financial situation and employment. Since women may work more often part-time, and women’s wages are on average lower Reference Mandel and Semyonov24 in almost any country, one may hypothesize that many women simply cannot afford stroke care, which may, in part, explain the observed discrepancies. Canada, on the other hand, has a publicly funded healthcare system where access to acute stroke care and post-stroke rehabilitation is, at least in theory, relatively equal for both sexes irrespective of the patient’s financial situation. That being said, few studies that reported inferior post-EVT outcomes in women were conducted in countries with publicly funded healthcare systems such as Germany. Reference Deb-Chatterji, Schlemm and Flottmann11
In the current study, women did have a higher age at stroke onset and thus a poorer baseline functional status. However, the NIHSS scores were similar to male patients, and none of the time metrics captured in the pre-hospital an in-hospital phase differed between women and men, neither did technical EVT success rates or clinical outcomes at 90 days. It seems that the timely high-quality stroke care provided to female EVT patients could “offset” the higher age at stroke onset. Thus, EVT should be provided to female EVT patients even if their baseline characteristics are more unfavorable compared to their male counterparts, since it seems that they can nevertheless achieve similar good outcomes compared to men. This also means that sex-specific undertreatment and withdrawal bias should be investigated and addressed whenever such sex discrepancies are identified.
Limitations
This study has several limitations. First, only patients undergoing EVT were included, and we were therefore unable to investigate potential biases with regard to patient selection for EVT. In other words, we could not exclude the possibility that EVT selection criteria for women were more stringent or lenient compared to men. It is also possible that the time of presentation/time since LKW differed between women and men in such a way that women presented later and were thus more often excluded from EVT, or vice versa. Therefore, we will aim to investigate sex differences a more inclusive patient sample that included both patients treated and not treated with EVT in a future study. Second, we performed analyses stratified by patient sex rather than gender. While sex refers to biological factors, the term gender encompasses social roles, behaviors, and expressions Reference Moleiro and Pinto25 and is arguably the more important concept in the setting acute stroke care. In older generations, which account for the majority of EVT patients, however, the number of transgender individuals is thought to be low (although it is unclear if there are generational differences in disclosure/self-identification) and sex can likely be reasonably used as a proxy for gender. Reference Chan26 Third, our data are a from a single Canadian province with a single CSC and the results may therefore not be generalizable to other geographic regions or countries. Lastly, we are limited by data availability; our patient sample included all patients treated with EVT, but we cannot reliably report the total number of strokes having occurred in the province during this timeframe. As such, we cannot definitively comment on the percentage of male and female patients undergoing EVT among all eligible patients.
Conclusion
Women undergoing EVT in the province of Saskatchewan between 2017 and 2022 were on average older compared to men at the time of their stroke onset. However, despite this age discrepancy, women were treated just as fast and achieved similar technical and clinical EVT outcomes compared to male patients.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/cjn.2023.286.
Acknowledgments
OPTIMISE Investigators, Canadian Stroke Consortium.
Author contribution
Nima Kashani and Johanna M. Ospel are contributed equally.
JO, NK, GH, RW, MK, LP: data curation, data analysis and interpretation, drafting, and critical revision of the manuscript and figures. Remaining authors: data interpretation and critical revision of the manuscript.
Funding
None.
Competing interests
Aravind Ganesh reports grants from CIHR, the Canadian Cardiovascular Society, Alberta Innovates, Campus Alberta Neuroscience, Government of Canada – INOVAIT program and New Frontiers in Research Fund, Microvention, Alzheimer Society of Canada, Alzheimer Society of Alberta and Northwest Territories, Heart and Stroke Foundation, Panmure House, Brain Canada, MSI Foundation, consulting fees from MD Analytics, MyMedicalPanel, Figure 1, CTC Communications Corp, Atheneum, DeepBench, Research on Mind, Creative Research Designs, AlphaSights, 42mr, honoraria from Figure 1, Alexion, Biogen, Servier Canada, expert testimony payment from Grosso Harper Law, travel support from AAN, Association of Indian Neurologists in America, American Heart Association, University of Calgary, patents pending US 17/317,771, editorial board membership Neurology:Clinical Practice, Neurology, Stroke, Frontiers in Neurology, stock options from SnapDx and Collavidence Inc., participation in several clinical trials. Johanna Ospel is a consultant to Nicolab. Michael Kelly reports consulting fees from Cerenovus, Imperative Care, and honoraria from Penumbra. The remaining authors have nothing to disclose.
