Sex Differences in Outcome After Thrombectomy for Acute Ischemic Stroke are Explained by Confounding Factors

Multiple randomized controlled trials (RCTs) have demonstrated the efficacy of endovascular treatment (EVT) among patients with acute ischemic stroke caused by large vessel occlusions. Several studies have reported worse outcome for females after acute ischemic stroke compared to males [1‐5]; however, conflicting results were reported on sex differences in outcome after stroke thrombectomy.

Post hoc analyses of RCTs on stroke thrombectomy within the HERMES collaboration suggested that sex does not influence functional outcome after EVT and demonstrated a similar proportion of both women and men, achieving functional independence 90 days after intervention [6]. Similar results were obtained in a pooled analysis of patients from different trial cohorts, which did not find a gender difference in outcome at 90 days, although females were older and had higher rates of atrial fibrillation [7]. Conversely, an analysis of the MR CLEAN trial reported that women were less likely to benefit from EVT with a lower proportion of independence before stroke and higher rates of serious adverse events and mortality in women than in men [8]. Thus, it is still largely unknown whether outcome after stroke thrombectomy differs between women and men. Of note, comparable studies in nonclinical trial populations are scarce. Identifying sex differences in outcome after EVT in a real-world setting may lead to targeted approaches for patient selection or sex-specific rehabilitation therapies.

The objective of our study was to determine whether sex is associated with outcome of stroke thrombectomy in clinical practice.

In this large prospective multicenter study of stroke patients treated by thrombectomy in clinical practice, we observed a worse functional outcome and higher mortality at 90 days in women; however, after adjustment for crucial confounders there was no association of sex with independent outcome.

These study results are consistent with previous studies, which did not find sex differences in outcome 90 days after stroke thrombectomy [6, 7]. In the HERMES pooled analysis of RCTs of stroke thrombectomy both women and men experienced comparable functional outcome and treatment effect of EVT was similar[6], despite women being older, similar as in our analysis. In contrast, in the MR CLEAN trial lower rates of functional independence and higher mortality were observed in women, together with a significant sex by treatment interaction with no significant treatment benefit in women [8]; however, these findings are solitary and have to be interpreted with caution, as the analysis was a post hoc subgroup analysis and the authors themselves suggested that the results may be explained by a play of chance.

Our analysis adds to the available studies of sex differences in outcome after stroke thrombectomy, as it is based on an independent large prospective registry of EVT in clinical practice. Trial populations are different from real-world patient cohorts resulting from the strict inclusion and exclusion criteria of clinical trials which are meant to ensure enrolment of those patients that are most likely to benefit from the investigational treatment. With respect to sex differences, women are frequently underrepresented in clinical trials [13], while in the real world stroke affects women and men in equal proportions. Moreover, stroke patients in trials are usually younger and have fewer comorbidities and less prestroke functional impairments than patients treated in clinical practice [1, 14, 15]. Comparing baseline characteristics of our study population to that from the HERMES analysis, some of these observations are confirmed. In HERMES, women were slightly underrepresented (47%), although in high-income countries women experience anterior circulation strokes more often [16]. In contrast, in our cohort 50.5% of patients were female and a higher proportion of women suffered from ischemic stroke of the anterior circulation. The analysis of the HERMES trial also showed similar rates of atrial fibrillation between women and men [6], while the prevalence of atrial fibrillation is usually higher in women [17], as it was in our study cohort. Moreover, the NIHSS score on admission was lower in women compared to men in the HERMES analysis, whereas the baseline NIHSS score did not differ by sex in our patient cohort. Of note, the MR CLEAN trial had less strict inclusion criteria than other RCTs of stroke thrombectomy, resulting in a patient cohort more comparable to real-world populations than other trial cohorts. Presumably, this might explain the similar outcome results with respect to the worse outcome in unadjusted analysis in MR CLEAN as compared to our cohort.

Similar to our analysis, a recent single center study of 279 patients in a real-world setting reported that females were less likely to be independent at 90 days after stroke thrombectomy than males [18]. In contrast, another single center study of stroke thrombectomy in clinical practice comprising 145 patients, found no sex differences in clinical outcome after EVT in acute ischemic stroke patients [19]. Small patient numbers as well as differences in baseline characteristics may account for the variability in the reported outcome results. In both studies mentioned above, women were older than men; however, the latter study included females with an equal prestroke disability compared to men, in contrast to our multicenter patient cohort in which prestroke disability differed significantly by sex with a larger proportion of women exhibiting functional disability before stroke. Notably, these parameters, i.e. age and prestroke functional impairment, are known predictors of outcome after EVT [20, 21].

After adjustment for confounders of interest (in both binary logistic and ordinal regression analyses), there was no significant association of sex and functional outcome in our study. Age and prestroke mRS were also identified as predictors of functional outcome in our patient cohort and showed significant differences between women and men. Thus, both variables may at least partially explain the observed sex differences in outcome. Similar findings were reported by the INSTRUCT collaboration [22].

The reason for the differences in age and prestroke functional status between women and men remains elusive. One interpretation is that women take care of more preventive efforts and have healthier lifestyles resulting in a longer stroke-free life. Thus, women are older when they suffer from first-ever stroke [22]. In consequence, more age-related diseases may occur accompanied by a higher prestroke disability [23] resulting in a less capacity to recover from stroke compared to men [22]. Moreover, the facts that a worse prestroke functional status in women may lead to less clear symptoms at onset time and that women usually take stroke symptoms less seriously, together may cause a late arrival to hospital and result in longer workflow times.

Notwithstanding, it is notable that other pivotal factors, such as the baseline NIHSS score, ASPECT score, the rates of prior IVT and successful reperfusion, in fact, did predict outcome in this patient cohort but showed no sex differences. Furthermore, in-hospital variables, such as the NIHSS and mRS score 24 h after intervention and at discharge as well as procedure times, the rate of adverse events and the rate of any intracerebral hemorrhage, were comparable between women and men. Thus, we cannot exclude that other variables which we did not collect might also add to sex differences in stroke outcome.

Our study has limitations. It lacks a patient group that received best medical treatment instead of EVT and, thus, we are not able to evaluate the effectiveness of EVT between women and men in clinical practice. Moreover, another important limitation is that some of the variables collected in the study had many missing values, in particular, the workflow times symptom onset-to-groin puncture and symptom onset-to-recanalization. However, this was considered in the statistical analysis. In addition, we cannot exclude a patient selection bias, since EVT was most likely performed in patients in whom a treatment benefit was assumed. Furthermore, we do not have data on other important, e.g. prestroke or postdischarge variables, which might have an impact on outcome in females.

In a prospective multicenter registry study of endovascular stroke treatment in clinical practice, we observed a worse functional outcome in women, which appears to be explained by differences in baseline characteristics, such as age and prestroke modified Rankin scale, between women and men. As a consequence, sex was not an independent predictor of functional independence at 90 days after adjustment for confounders.