Endovascular treatment versus standard medical treatment in patients with established large infarct: a cohort study

Background: Previous trials confirmed the benefit of endovascular treatment (EVT) in acute large core stroke, but the effect of EVT on outcomes in these patients based on noncontrast computed tomography (NCCT) in real-world clinical practice was unclear. The aim of this study was to explore the effect of EVT versus standard medical treatment (SMT) in patients with large ischemic core stroke defined as Alberta Stroke Program Early CT Score (ASPECTS) ≤5 based on NCCT alone. Materials and methods: Patients with acute large core stroke at 38 Chinese centers between November 2021 and February 2023 were reviewed from a prospectively maintained database. The primary outcome was favorable functional outcome [modified Rankin Scale score (mRS), 0–3] at 90 days. Safety outcomes included 48 h symptomatic intracerebral hemorrhage (sICH) and 90-day mortality. Results: Of 745 eligible patients recruited at 38 stroke centers between November 2021 and February 2023, 490 were treated with EVT+SMT and 255 with SMT alone. One hundred and eighty-one (36.9%) in the EVT group achieved favorable functional independence versus 48 (18.8%) treated with SMT only [adjusted risk ratio (RR), 1.86; 95% CI: 1.43–2.42, P<0.001; adjusted risk difference (RD), 13.77; 95% CI: 7.40–20.15, P<0.001]. The proportion of sICH was significantly higher in patients undergoing EVT (13.3 vs. 2.4%; adjusted RR, 5.17; 95% CI: 2.17–12.32, P<0.001; adjusted RD, 10.10; 95% CI: 6.12–14.09, P<0.001). No significant difference of mortality between the groups was observed (41.8 vs. 49.0%; adjusted RR, 0.91; 95% CI: 0.77–1.07, P=0.24; adjusted RD, −5.91; 95% CI: −12.91–1.09, P=0.1). Conclusion: Among patients with acute large core stroke based on NCCT in real-world, EVT is associated with better functional outcomes at 90 days despite of higher risk of sICH. Rates of procedure-related complications were relatively higher in the EVT+SMT group.


Introduction
Acute ischemic stroke with large cores accounts for ~20% of large vessel occlusion strokes but usually causes catastrophic medical condition, such as bedridden, incontinent, or even death [1] .Patients with large ischemic cores, defined by the Alberta Stroke Program Early Computed Tomography Score (ASPECTS) of 0-5 or ischemic core ≥ 50 ml, are ineligible for endovascular treatment (EVT) according to current American and European guidelines due to wide early ischemic injury and less possibility to achieve functional independence [2][3][4] .
Recently, four landmark stroke trials, Endovascular Salvage for Cerebral Ultra-acute Embolism-Japan Large Ischemic Core Trial (RESCUE-Japan LIMIT) [5] , randomized controlled trial to optimize patient's selection for endovascular treatment in acute ischemic stroke (SELECT 2) [6] , endovascular therapy in acute anterior circulation large vessel occlusive patients with a large infarct core (ANGEL-ASPECT) [7] , and endovascular thrombectomy for acute ischemic stroke with established large infarct (TENSION) [8] have confirmed the safety and efficacy of EVT combined with standard medical treatment (SMT) in patients with large ischemic burden compared with SMT-alone.The intention-to-treat population analysis of the primary outcome in the Thrombectomy for Emergent Salvage of Large Anterior Circulation Ischemic Stroke (TESLA) failed to demonstrated efficacy of EVT in patients with a large-core infarction on the basis of ASPECTS 2-5 according to noncontrast computer tomography (NCCT), but the results of secondary outcome including the proportion of mRS score of 0-3 at 90 days and rate of major neurological improvement highlighted a strong suggestion in favor of EVT [9] .Even so, a low rate of modified Rankin Scale (mRS) of 0 to 3 (31-47%) and an uncertain range of symptomatic intracerebral hemorrhage (0.6-9%) make it easy to raise a rational fear that how much of the effectiveness of randomized controlled trials confirming EVT in patients with large infarction translates into benefit of patients in real-world medical practice [1,10,11] .
The enrolled patients of previous trials were strictly screened mainly by advanced imaging with MRI or computed tomography perfusion (CTP).Advanced imaging could identify patients with large core but wide penumbra that could be salvaged through EVT [11,12] .But strict advanced imaging selection may exclude the patients that could benefit from EVT and even make delay in treatment to increase the chance of futility [13] .Besides, access to urgent MRI or CTP is not universally available in many stroke centers, especially in developing countries [14] .Conversely, NCCT is more available at stroke centers in clinical practice.Previous studies did not observed significant differences in the clinical outcomes of patients selected with NCCT compared with those selected with advanced imaging [15,16] .Therefore, the present study aimed to explore the association between EVT combined with SMT and clinical outcomes in patients with large cores according to NCCT compared to SMT-alone in real-world.

Study cohort and patients
This study was a subanalysis of a prospective multicenter corhort study and patients treated between 1 November 2021 and 8 February 2023.The registry was an ongoing, prospective, observational, nationwide registry including all patients with acute large vessel occlusion within 24 h from the point that they were last known well and undergoing standard treatment in China (registered at the https://www.chictr.org.cn/).The study protocol was approved by ethics committee of the leading center and the local committees of the participating hospitals gave approval as well.All patients or their legally authorized representatives provided signed, informed consent.
The inclusion criteria for this analysis were as follows: (1) an age at least 18 years old; (2) acute ischemic stroke due to anterior circulation large vessel occlusion, defined as occlusion of the internal carotid artery (ICA) or the M1 segment or M2 segment of the middle cerebral artery; (3) large ischemic core on NCCT (defined as an ASPECTS of 0-5); (4) within 24 h of stroke onset or last known within 24 h (the time metric of time last known well within 24 h was used instead if the presentation time was unavailable).Patients were excluded from the study in the case of (1) prestroke mRS > 2; (2) lack of follow-up information on 90-day outcomes; (3) serious or terminal illness that was not related to acute ischemic stroke.

Treatments
Patients were divided into the SMT-alone and EVT plus SMT group.The SMT-alone group received SMT including intravenous thrombolysis (IVT, the dose of alteplase was 0.9 mg/kg for Alteplase and 0.25 mg/kg for Tenecteplase), antiplatelet drugs, anticoagulation drugs, or combination of these treatments according to the guidelines for the management of acute ischemic stroke [17] .EVT included stent retrievers, aspiration, balloon angioplasty, stenting, intraarterial thrombolysis, mechanical fragmentation, or any combinations of these approaches.The decision to perform EVT + SMT or SMT alone was left to the discretion of the local physicians.Decisions to perform decompressive hemicraniectomy in patients with severe brain swelling were made in accordance with local practices.

Data collection
Patients' baseline demographic characteristics, stroke risk factors, laboratory findings, stroke severity [based on the National Institutes of Health Stroke Scale (NIHSS) [18] ], collateral status [based on the American Society of Interventional and Therapeutic Neuroradiology/Society of Interventional Radiology collateral grading system (ASITN/SIR) [19] ], time from symptom onset or last known well to imaging, groin puncture and recanalization, EVT technique, complications, reperfusion grades, presumed HIGHLIGHTS • Our analysis found that the use of endovascular treatment resulted in better functional outcomes at 90 days despite of higher risk of symptomatic intracranial hemorrhage and complications for patients with large infarctions defined as Alberta Stroke Program Early Computed Tomography Score (ASPECTS) of 0 to 5 based on noncontrast CT selection.
• Several landmark studies have demonstrated the efficacy of endovascular treatment for patients with large infarction for selected.Even so, a low rate of modified Rankin Scale (mRS) of 0-3 (31-47%) and an uncertain range of symptomatic intracerebral hemorrhage (0.6-9%) make it easy to raise a rational fear that how much of the effectiveness of randomized controlled trials confirming EVT in patients with large infarction translates into the benefit of patients in real-world medical practice.• The data that support the findings of this study are available from the corresponding author upon reasonable request.
stroke causative mechanism [based on the Trial of ORG10172 in Acute Stroke Treatment (TOAST) classification [20] ], location of occlusion, and baseline core infarct determined by the NCCTbased ASPECTS were recorded.

Imaging assessment
The imaging core laboratory evaluated the findings on baseline NCCT for the ASPECTS, baseline imaging (computed tomographic angiography, magnetic resonance angiography, or digital subtraction angiography) for the occlusion site, angiographic outcomes on digital subtraction angiography imaging for technical efficacy outcomes regarding reperfusion, and the follow-up computed tomography within 48 h for the presence of intracranial hemorrhage.Successful reperfusion was defined as the restoration of blood flow to greater than 50% (2b to 3) of the involved territory, as assessed with the use of the modified Treatment in Cerebral Ischemia classification [mTICI, scores range from 0 (no flow) to 3 (normal flow) [21] ].Baseline imaging, reperfusion grades, and postprocedural imaging were independently evaluated by the imaging core laboratory who were blind to the treatment groups and clinical outcomes.

Clinical outcomes
The primary outcome was favorable functional outcome, defined as a mRS of 0-3 at 90 days, which was recorded during a followup visit or telephone encounter at 90 days after stroke by local physicians or registered nurse.Secondary outcomes included ordinal score on mRS at 90 days, functional independence (defined as mRS of 0-2), the proportion of mRS 0-4, successful reperfusion.Safety outcomes included the incidence of death within 90 days, symptomatic intracranial hemorrhage (SICH) within 48 h according to the Heidelberg Bleeding Classification [22] , and any intracranial hemorrhage within 48 h.

Statistical analysis
Baseline characteristics, treatment profiles, time metrics were reported according to the treatment arms.Data were presented as medians [interquartile ranges (IQRs)] or numbers with percentages, unless otherwise indicated.Univariate analysis was performed using the Kruskal-Wallis test, χ 2 test, or Fisher exact test, as appropriate.Missing baseline covariates were imputed using the hot deck methods in the covariate adjusted analysis based on the covariate distributions.Only a small number of patients needed the hot deck imputation; therefore, the techniques recommended in (24) for a variance estimate that incorporates the additional variance from the missing information was not implemented.
For efficacy and safety outcomes assessment between patients treated with EVT + SMT and those with SMT-alone, propensity score matching (PSM) methods were used to balance prognostic important factors.The propensity score was estimated using a multivariable logistic regression model, with the treatment received as the dependent variable and age, history of hypertension, hyperlipidemia, diabetes, baseline ASPECTS, baseline NIHSS, systolic blood pressure, diastolic blood pressure, intravenous thrombolysis, ASITN/SIR, stroke mechanism, occlusion sites, time from last known well to imaging as covariates.We performed a 1:1 matching based on the nearest neighbor matching with a 0.2 caliper.
The multivariable models were adjusted for age, history of hypertension, hyperlipidemia, diabetes, baseline ASPECTS, baseline NIHSS, intravenous thrombolysis, ASITN/SIR, stroke mechanism, occlusion sites, time from last known well to imaging, systolic blood pressure, diastolic blood pressure.The generalized linear models were used as the primary analysis.Models with robust error estimators with the Poisson distribution and log link function were used to estimate the risk ratio (RR), and with the Gaussian distribution and identity link function were used to estimate the risk difference (RD).For the comparison of the distributions of the mRS scores at 90 days, ordered logistical regression was used to estimate the common odds ratio.Besides, two assumption-free method, the Wilcoxon-Mann-Whitney generalized odds ratio and win ratio approaches was used for the comparison of the distribution of the mRS scores for sensitivity analysis [23,24] .Besides, generalized linear mixed models were used take into account of center effect and pair effect in sensitivity.Generalized estimating equation were also used as sensitivity analysis to account center-effect.
In the inverse probability of treatment weighting (IPTW) cohort, the treatment effect was estimated with the inversed probability-weighted regression adjustment model, which used the inversed propensity score to weight each subject, and adjusted for the weighted regression coefficients to compute the averages of treatment-level predicted outcomes.Using the doubly robust estimation to reduce the bias and be less sensitive to misspecification [25] .The primary analysis of the primary outcome were based on the IPTW analysis.
We further investigated the heterogeneity in treatment effect size for the primary outcome within the following subgroups: age ( ≤ 75 vs.> 75 years old), sex (female vs. male), baseline NIHSS score ( ≤ 17 vs.> 17), ASPECTS ( ≤ 2 vs. > 2), IVT (no vs. yes), occlusion site, stroke causative mechanism, time from last known well to imaging ( ≤ 360 vs. > 360 min).A multiplicative term was entered into regression models to estimate the significance of the interaction with the treatment assignment.
In addition, an instrumental variable analysis (IVA) was performed to evaluate the association of treatment allocation with clinical outcomes.The center-level preference for EVT, which is defined as the proportion of EVT for all patients at a particular center, was used as the instrument.A two-stage residual inclusion approach was employed: in the first stage, an expectation of treatment allocation based on co-variables and instrumental variable was estimated, and the co-variables were the same as in the other adjusted model; then, in the second stage, outcomes were predicted based on original treatment allocation, covariates, and residuals from the first-stage regression.
All statistical tests were two-sided, with P-values <0.05 considered statistically significant.Statistical analyses were conducted in SAS 9.4 and STATA 17.All the work has been reported in line with the strengthening the reporting of cohort, cross-sectional, and case-control studies in surgery (STROCSS) criteria [26] (Supplemental Digital Content 1, http://links.lww.com/JS9/C498).

Patients cohort and baseline characteristics
Totally, 745 eligible patients were eligible and consented from the prospective study between November 2021 and February 2023, from 38 stroke centers across China.A total of 255 patients received SMT alone, while 490 treated with EVT plus SMT. Figure 1 shows a flowchart of patients enrolled in this study.(Power were analyzed in Figure S1, Supplemental Digital Content 2, http://links.lww.com/JS9/C499).

Sensitivity analysis
Using the IVA model in sensitivity analysis (Supplementary Table S4 in the Supplement, Supplemental Digital Content 2, http:// links.lww.com/JS9/C499), the Wald F-statistic for center proportion of EVT plus SMT was 217.51, suggesting a strong instrument.There was a significant association between EVT plus

Discussion
Our results suggest that, in the real-word practice, EVT may improve clinical functional outcomes in patients with large vessel occlusion presenting with large ischemic core (ASPECTS ≤ 5 on NCCT) within 24 h of symptom onset or last known well despite of high risk of symptomatic intracerebral hemorrhage.
Several observational studies have investigated the effect of imaging modality (NCCT vs. CTP/MRI) on the selection of EVT in AIS patients [15,16,[27][28][29] .However, the results were inconsistent, with some indicating similar functional outcomes between the two imaging modalities [15,16,[27][28][29] , while others showed improved outcomes in patients selected by advanced imaging paradigms.Moreover, these studies were based on patients with mild or moderate infarction, whether the result can be extended to patients with large core infarction remain unclear.Previous trials of EVT in patients with large core stroke mainly based on advanced imaging [5][6][7] .However, none of the participants of the previous studies were enrolled based on NCCT alone.Nearly 86% of patients included in the RESCUE-Japan LIMIT with an ASPECTS value were based on MRI, which have been showed to be more sensitive to identify ischemic regions but overestimating ischemic core volumes compared with NCCT.Moreover, MRI-based ASPECTS was reported 1 scale lower than that measured by NCCT [30] .Most of the enrolled patients in the ANGLE-ASPECT trial were screened by CTP.Advanced imaging selection is beneficial to improve clinical outcomes of patients with large core, but this selection may make delay in treatment and deny many patients who could benefit from EVT.In these trials, nearly only three patients of 10 large core patients with EVT are functional independent, as NCCT is available at all stroke centers, how about the effect of EVT on clinical outcomes in patients with large core evaluated by NCCT alone?In the EVT group of our studies, favorable outcome occurred in 36.9% of the patients.This result was slightly lower than that of the SELECT 2 trial, which mostly used more generalizable imaging triage methods (NCCT).This can be explained that our study enrolled patients with ASPECTS 0-5, but only patients with ASPECTS 3-5 were enrolled in the SELECT trial, as low ASPECTS rating on NCCT predicts poor outcome after reperfusion [31] .In a secondary analysis of the RESCUE-Japan LIMIT, EVT was not associated with improved clinical outcomes at 90 days in patients with large core stroke and ASPECTS 3 or less [32] .
Although, EVT is associated with improved clinical outcomes in our study, death occurred in more than 40% of patients despite of EVT, and there is no significant difference between the two groups.It still remains a great challenge for both relatives of patients and neurointerventionists to decide whether to perform EVT considering a high chance of death and high cost.In the RESCUE-Japan LIMIT and ANGLE-ASPECT trial, ~20% of death within 90 days were reported, which was much less than that of our study.This could be explained that advance imaging selection excluded those patients with more opportunity to achieve poor outcome or even death.However, mRS of 5 occurred in 37 (7.6%) patients in the EVT group and 32 (12.5%) patients in the SMT group in our study, which suggests that EVT may decrease the opportunity of outcome of bedridden and incontinent.To some degree, EVT may improve the quality of lives among the survivors.However, the EVT was associated higher risk of complications as sICH.In our study, the rate of SICH was 13.3% in the EVT group, which was significantly higher than the SMT group.Previous study reported 11.2% of SICH in patients with ASPECTS 2 to 5 after EVT [10] .In the recent clinical trials, SICH occurred in 0.6-9% patients treated with EVT, which is much less than that of our study [5][6][7] .This could be explained as followed.First, patients with low ASPECTS are at higher risk of SICH [33] .In our study, 27.6% of patients in the EVT group presented with ASPECTS 0 to 2. All of the previous trials excluded those patients with low ASPECTS (0-2) due to high risk of SICH.A prespecified secondary analysis of the RESCUE-Japan LIMIT trial showed that SICH occurred in 10.7% patients among those with ASPECTS 0-3 after EVT [32] .Second, more patients with large artery atherosclerotic thrombosis were included in our study, which predicts a lower chance of successful reperfusion and a high number of thrombectomy passes [34,35] .In addition, these patients usually need to be treated with antithrombotic therapy.These may increase the risk of intracerebral hemorrhage.Third, despite the proportion of IVT (24.9%) in our study was comparable with previous trials (20.8-28.7%), it is also an important predictor of SICH.

Limitations
The strengths of our study included the large-scale, prospective, multicenter design.This study also has several limitations.First, it has all the inherent limitations of a nonrandomized study.PSM or multivariable analyses can never adjust completely for systematic differences between treatment groups.Second, only Chinese patients were included, which may limit the generalizability.

Conclusions
In patients with large cores on NCCT, EVT resulted in reasonable rates of favorable functional outcomes despite of higher risk of symptomatic intracerebral hemorrhage.Future clinical trials aimed at addressing the efficacy and safety of EVT in patients with large cores based on NCCT are warranted and under way.

Figure 2 .
Figure 2. Distribution of the Modified Rankin Scale score at 90 days.
shows baseline characteristics of the eligible patients.

Table 1
Baseline characteristics of the patients.
a Data on blood pressure on admission were missing for eight patients in the EVT group.b Data on glucose were missing for 12 patients in EVT group and eight patients in SMT group.c Data on ASTIN/SIR grade were missing for two patients in the SMT group.d Data on last seen well to imaging time were missing for seven patients in the SMT group.e Data on last seen well to puncture time were missing for five patients in the EVT group.f Data on last seen well to recanalization time were missing for six patients in the EVT group.

Table 2
Primary and secondary efficacy outcomes.