Clinical and Demographic Factors Associated With Hemorrhagic Transformation Subtypes After Endovascular Thrombectomy for Acute Ischemic Stroke Authors: Nicole Khezri, Marie K. Luff, Salvador Miralbés, Bharath Naravetla, Alejandro Spiotta, Christian Loehr, Mario Martínez-Galdámez, Ryan McTaggart, Luc Defreyne, Pedro Vega, Osama O. Zaidat, Lori Lyn Price, Rishi Gupta, Markus Möhlenbruch, David S. Liebeskind Published: 2026 ABSTRACT BACKGROUND:Hemorrhagic transformation (HT) after endovascular thrombectomy (EVT) is associated with worse neurological outcomes. We sought to identify clinical and demographic factors associated with HT and HT subtypes after EVT.METHODS:Patients with anterior circulation large vessel occlusion who underwent first-pass EVT using different thrombectomy devices were enrolled in a global multicenter prospective registry. Baseline clinical characteristics were compared between individuals with and without HT and across HT subtypes 24 hours after EVT. Multivariable logistic regression models were used to determine risk factors associated with HT and HT subtypes.RESULTS:The study cohort comprised 1453 patients. Imaging evidence of HT occurred in 40% of patients, of whom 3.2% had symptomatic HT. In a multivariable regression analysis, diabetes (odds ratio [OR], 1.47 [95% CI, 1.09–1.99]; P=0.01), National Institutes of Health Stroke Scale score of 16 to 19 (OR, 1.45 [95% CI, 1.09–1.94]; P=0.01), National Institutes of Health Stroke Scale score ≥20 (OR, 1.98 [95% CI, 1.56–2.50]; P<0.0001), Alberta Stroke Program Early Computed Tomography Score ≥8 (OR, 0.31 [95% CI, 0.24–0.39]; P<0.0001), and elevated baseline serum glucose (OR, 1.56 [95% CI, 1.23–1.98]; P=0.0003) were independent risk factors for HT. Atrial fibrillation (OR, 0.75 [95% CI, 0.58–0.96]; P=0.02) was a protective factor against HT. Baseline Alberta Stroke Program Early Computed Tomography Score <8, elevated baseline serum glucose, and higher baseline National Institutes of Health Stroke Scale scores were independent predictors of hemorrhagic infarction-1 and hemorrhagic infarction-2. Diabetes was significantly associated with hemorrhagic infarction-2, and Alberta Stroke Program Early Computed Tomography Score <8 was significantly associated with subarachnoid hemorrhage.CONCLUSIONS:Given that certain risk factors are associated with specific subtypes of HT, our analysis demonstrates the importance of not only reporting symptomatic HT but also classifying HT into subtypes in clinical trials. Identifying risk factors associated with HT subtypes may improve prognostic accuracy and guide post-EVT monitoring.REGISTRATION:URL: https://www.clinicaltrials.gov; Unique identifier: NCT03845491. CLINICAL PERSPECTIVE What Is New? In a large global multicenter endovascular thrombectomy registry (n=1453), this study demonstrates that established risk factors for hemorrhagic transformation (HT) show distinct, subtype-specific associations (eg, diabetes with hemorrhagic infarction-2, Alberta Stroke Program Early Computed Tomography Score <8 with subarachnoid hemorrhage), rather than uniform effects across all HT categories. In a large global multicenter endovascular thrombectomy registry (n=1453), this study demonstrates that established risk factors for hemorrhagic transformation (HT) show distinct, subtype-specific associations (eg, diabetes with hemorrhagic infarction-2, Alberta Stroke Program Early Computed Tomography Score <8 with subarachnoid hemorrhage), rather than uniform effects across all HT categories. Beyond symptomatic HT, this analysis identifies heterogeneity in clinical outcomes across HT subtypes, highlighting that hemorrhagic infarction-2, parenchymal hematomas, and especially multiple concurrent hemorrhagic subtypes carry disproportionately worse functional outcomes and mortality compared with hemorrhagic infarction-1 and isolated subarachnoid hemorrhage. Beyond symptomatic HT, this analysis identifies heterogeneity in clinical outcomes across HT subtypes, highlighting that hemorrhagic infarction-2, parenchymal hematomas, and especially multiple concurrent hemorrhagic subtypes carry disproportionately worse functional outcomes and mortality compared with hemorrhagic infarction-1 and isolated subarachnoid hemorrhage. What Are the Clinical Implications? Subtype-specific risk stratification can improve postendovascular thrombectomy monitoring and prognostication, supporting a shift away from using symptomatic HT alone toward granular classification of hemorrhagic subtypes to guide clinical decision-making and future trial design. Subtype-specific risk stratification can improve postendovascular thrombectomy monitoring and prognostication, supporting a shift away from using symptomatic HT alone toward granular classification of hemorrhagic subtypes to guide clinical decision-making and future trial design. Hemorrhagic transformation (HT) is the principal adverse event of endovascular thrombectomy (EVT). HT can decrease or eliminate the benefits of reperfusion after EVT, leading to worse functional outcomes. Hemorrhagic complications due to EVT are secondary to acute reperfusion injuries attributable to disrupted blood-brain barrier, failed recanalization, and procedural vascular trauma.1,2 A range of clinical, imaging, and procedural factors has been shown to predict HT after EVT. Greater baseline neurological severity, hyperglycemia, diabetes, advanced age, and larger infarct burden are consistently associated with increased risk of post-EVT HT.3–7 Baseline imaging predictors of HT include low Alberta Stroke Program Early Computed Tomography Score (ASPECTS), large ischemic core volume, and poor collateral circulation.1,8 Procedural factors, including prolonged time to reperfusion, incomplete recanalization, and increased device manipulation, have also been implicated as contributors to HT.2,9 Several classification systems exist for HT and symptomatic HT (sHT) assessment, which may explain variations in HT rate among clinical trials.8 sHT is widely used as a safety outcome in clinical trials after EVT because it is associated with increased risk of mortality, worse clinical outcome at discharge, and less favorable neurological outcomes.10 Consequently, other subtypes of intracranial hemorrhage such as hemorrhagic infarction (HI)-1, HI-2, parenchymal hematoma (PH)-1, PH-2, PH remote from infarcted tissue, intraventricular hemorrhage, subarachnoid hemorrhage (SAH), epidural hemorrhage, and subdural hemorrhage after EVT have been largely understudied.11 Although different subtypes of HT may not lead to acute neurological deterioration, they can still adversely affect functional outcomes.12,13 However, whether established predictors of HT after EVT apply uniformly across specific hemorrhagic subtypes remains incompletely understood. Identifying clinical factors associated with distinct HT subtypes may therefore improve risk stratification, procedural planning, and post-EVT monitoring. Accordingly, we investigated demographic and baseline clinical variables of patients with and without HT to identify clinically relevant predictors of HT and specifically HT subtypes after EVT in patients with large vessel occlusion in the anterior circulation. By moving beyond sICH as a singular safety end point, this study examines whether established clinical and demographic predictors of HT apply to specific HT subtypes and assesses the differential prognostic significance of individual and overlapping hemorrhagic patterns. Methods Study Design and Participants Data supporting the findings of this study may be available on request by researchers with an approved research proposal. The ASSIST registry is a global multicenter prospective consecutive registry evaluating operator EVT technique after acute, large vessel ischemic stroke in the anterior circulation. Participants enrolled in the ASSIST registry underwent first-pass ischemic stroke treatment with 1 of 3 interventional techniques: SR classic, SR combination, or direct aspiration. Participants were enrolled from January 2019 to January 2022 across 71 global sites in North America, Europe, and Asia. The study protocols of the ASSIST registry were approved and upheld by the medical ethics committee of each participating institution. The ASSIST trial methods have been previously described.9 This study was conducted in accordance with the STROBE guidelines (Strengthening the Reporting of Observational Studies in Epidemiology; Supplemental Material). A blinded independent core lab identified and classified HT among 1492 individuals who were enrolled in ASSIST by comparing baseline noncontrast computed tomography to computed tomography or MRI imaging at the 24-hour visit (18–48 hours after EVT per the protocol). HT was further classified as HI-1, HI-2, PH-1, PH-2, remote from infarcted tissue, intraventricular hemorrhage, SAH, epidural hemorrhage, and subdural hemorrhage. Variable Definitions sHT was identified among hemorrhagic subtypes. sHT was defined as the presence of core lab adjudicated HT on 24-hour imaging (−6/+24 hours) and a ≥4-point change in site-reported National Institutes of Health Stroke Scale (NIHSS) score at 24 hours (−6/+24 hours) after EVT. Preprocedural blood pressure was considered elevated if systolic blood pressure was ≥140 or diastolic blood pressure was ≥90. Statistical Analysis Demographics and baseline clinical and imaging characteristics were compared between individuals with any HT and without HT 24 hours after EVT and were described for the subcategories of HT (HI-1, HI-2, SAH, and PH-1 or PH-2), including patients with multiple subcategories. The subcategories were mutually exclusive. Categorical data were expressed as percentages and continuous variables with normal distribution were presented as mean±SD. The PROC SURVEYFREQ procedure with a cluster statement was used to adjust for clustering within site for categorical variables, whereas the PROC MIXED procedure with a random intercept statement was used for continuous variables. Multiple imputation was performed for regression analyses using the following variables: age, female, hypertension, atrial fibrillation, dyslipidemia, diabetes, prior ischemic stroke, transient ischemic attack, coronary artery disease, baseline antiplatelets/anticoagulant, baseline modified Rankin Scale (mRS) score, wake up stroke, baseline ASPECTS, baseline NIHSS score, baseline systolic and diastolic blood pressure, blood glucose, creatinine, international normalized ratio, platelets, time last known well to groin puncture, and occlusion site. Outcomes were not imputed. PROC SURVEYLOGISTIC with a cluster statement was used to adjust for clustering within sites. Univariate regression analysis was used to evaluate the association between demographics and preprocedural characteristics and HT. Demographic and preprocedural clinic characteristics identified by the investigator as clinically important were evaluated by univariate regression analysis and included age, sex, medical history (hypertension, atrial fibrillation, diabetes, dyslipidemia, previous ischemic stroke or transient ischemic attack, and coronary artery disease), taking antiplatelets or anticoagulants at admission, prestroke mRS score ≥3, wake up stroke, NIHSS score at admission, ASPECTS ≥8, high blood pressure, serum glucose, serum creatinine, platelet count, and international normalized ratio ≥1.2. Variables that were statistically significant (P<0.05) in univariate analyses were candidates for multivariable regression analysis. In the multivariable regression model, these variables were entered into a backward logistic regression analysis until only variables with P<0.05 remained in the model. A sensitivity analysis was performed where all variables were forced into the multivariable regression model. PROC SURVEYLOGISTIC with a cluster statement was used to adjust for clustering within sites. Regression models with imputed data were constructed to assess factors associated with HI-1, HI-2, and SAH; patients with PH or multiple HT subtypes were excluded. A smaller subset of variables, also determined a priori, was used for these models: age, sex, ASPECTS ≥8, and NIHSS score at admission, plus any variables significant in the multivariable HT model. Backward logistic regression was not performed for these analyses. PROC SURVEYLOGISTIC with a cluster statement was used to adjust for clustering within sites. A forest plot was generated to display logistic regression results for the 3 most common subtypes (HI-1, HI-2, and SAH). The remaining subcategories had insufficient patient numbers to perform fully adjusted analyses for the table or forest plot. Horizontal stacked bar plots were constructed to illustrate 90-day mRS score outcomes for patients with and without HT, as well as across HT subtypes. Comparisons of demographic, clinical, and imaging characteristics, along with regression analyses, were performed using SAS version 9.4 (Cary, NC), and graphs were generated in R version 4.5.1 (2025-06-13). All analyses accounted for clustering within site, and statistical significance was defined as P<0.05. Results Of 1453 patients who had core lab adjudicated imaging data available for analysis, 581 (40.0%) patients had imaging evidence of HT. There were 233 (16.0%) patients classified as HI-1 only, 160 (11.0%) patients as HI-2 only, 115 (7.9%) patients as SAH only, 16 (1.1%) patients classed as PH-1 or PH-2 only, and 56 (3.9%) of patients with multiple subcategories. Table 1 presents the demographic and preprocedural clinical characteristics of the 1453 patients in the study population. There were 47 (8.5%) patients with HT who were also determined to have core lab/NIHSS score–defined sHT. The highest frequency of sHT was seen among those with PH-1 or PH-2 (27.3%; 3/11), followed by multiple subtypes (18.9%; 10/53), HI-2 (11.9%; 18/151), SAH (9.0%; 10/111), and then HI-1 (2.7%; 6/226). Table 1.Baseline Demographic and Clinical Characteristics Baseline Demographic and Clinical Characteristics In the group without HT, the mean±SD age was 71.2 ± 14.3, and 465 (53.3%) were women. In the group with HT, the mean±SD age was 69.2 ± 14.2 and 294 (50.6%) were women. Patients with HT had a significantly lower mean age, lower baseline NIHSS score, lower incidence of atrial fibrillation, a higher incidence of diabetes, a higher baseline serum glucose, and baseline ASPECTS <8 compared with patients without HT. Table 2 summarizes the univariate and multivariable analysis of predictors in the HT group. Lower mean age and lower prestroke mRS score in the HT group were significant in univariate analysis, but not in multivariable analysis. Diabetes (odds ratio [OR], 1.47 [95% CI, 1.09–1.99]; P=0.01), baseline NIHSS score 16 to 19 (OR, 1.45 [95% CI, 1.09–1.94]; P=0.01), baseline NIHSS score ≥20 (OR, 1.98 [95% CI, 1.56–2.50]; P<0.0001), baseline ASPECTS ≥8 (OR, 0.31 [95% CI, 0.24–0.39]; P<0.0001), and baseline serum glucose ≥140 (mg/dL; OR, 1.56 [95% CI, 1.23–1.98]; P=0.0003) were independent predictors of HT. Atrial fibrillation (OR, 0.75 [95% CI, 0.58–0.96]; P=0.02) was associated with a lower risk of HT. In a sensitivity analysis with all variables in the table forced into a multivariable regression model, the results were similar, with the exception of atrial fibrillation, which became nonsignificant (P=0.07). Table 2.Univariate and Multivariable Regression Analysis of Key Demographic and Clinical Characteristics in Patients With HT Versus No HT Univariate and Multivariable Regression Analysis of Key Demographic and Clinical Characteristics in Patients With HT Versus No HT Among all analyzed baseline clinical parameters for each individual hemorrhage subtype included in the multivariable model, baseline ASPECTS ≥8 (OR, 0.37 [95% CI, 0.28–0.47]; P≤0.0001; OR, 0.16 [95% CI, 0.11–0.24]; P≤0.0001), serum glucose ≥140 (OR, 1.57 [95% CI, 1.13–2.18]; P=0.007; OR, 2.43 [95% CI, 1.61–3.68]; P<0.0001), baseline NIHSS score 16 to 19 (OR, 1.37 [95% CI, 1.04–1.81]; P=0.03; OR, 2.30 [95% CI, 1.32–4.02]; P=0.003), and baseline NIHSS score ≥20 (OR, 1.91 [95% CI, 1.28–2.85]; P=0.002; OR, 3.55 [95% CI, 2.26–5.57]; P<0.0001) were independent predictors of HI-1 and HI-2 subtypes, respectively (Table 3; Figure 1). Diabetes (OR, 1.82 [95% CI, 1.24–2.66]; P=0.002) was an independent predictor of HI-2, but not HI-1. Baseline ASPECTS ≥8 was the only significant predictor of SAH subtype (OR, 0.55 [95% CI, 0.35–0.86]; P=0.009). Table 3.Multivariable Effect of Clinical Factors on HI-1, HI-2, and SAH Multivariable Effect of Clinical Factors on HI-1, HI-2, and SAH Multivariable predictors of hemorrhagic transformation subtypes (hemorrhagic infarction [HI]-1, HI-2, and subarachnoid hemorrhage [SAH]). ASPECTS indicates Alberta Stroke Program Early Computed Tomography Score; NIHSS, National Institutes of Health Stroke Scale; and OR, odds ratio. HT and HT subtypes were associated with 90-day functional outcomes. Among patients without HT (n=727), 71.1% (517/727) achieved functional independence (mRS score 0–2), whereas 10% (73/727) experienced severe disability or death (mRS score 5–6; Figure 2). In contrast, patients with sHT (n=35) had markedly worse outcomes, with only 2.9% (1/35) achieving independence and nearly half (49%, 17/35) expiring by 90 days. Patients with asymptomatic HT (n=416) demonstrated an intermediate pattern, with 44% (185/416) achieving independence and 16% (66/416) mortality. Ninety-day modified Rankin Scale (mRS) score outcomes by presence and type of hemorrhagic transformation (HT). aHT indicates asymptomatic hemorrhagic transformation; and sHT, symptomatic hemorrhagic transformation. When stratified by hemorrhagic subtype (n=472 total), functional independence was most frequent in HI-1 (48.4%, 93/192) and isolated SAH (53.3%, 48/90), whereas HI-2 and PHs were associated with poor recovery (Figure 3). In HI-2 (n=129), only 27% (35/129) achieved independence, whereas 12% (16/129) were severely disabled and 22% (29/129) expired. PHs (PH-1/PH-2, n=12) demonstrated the poorest outcomes overall, with just 17% (2/12) of patients independent and 33% (4/12) mortality. Notably, patients with >1 hemorrhagic subtype (n=48) had similarly poor outcomes, with only 21% (10/48) achieving independence and 31% (15/48) mortality, suggesting this group is prognostically distinct. Ninety-day modified Rankin Scale (mRS) score outcomes by hemorrhagic transformation (HT) subtypes. HI indicates hemorrhagic infarction; PH, parenchymal hematoma; and SAH, subarachnoid hemorrhage. Discussion Whereas previous trials have established predictive factors associated with sHT and any HT, our analysis looked at whether those same clinical and demographic factors were associated with specific HT subtypes. In unadjusted analyses, younger age, diabetes, elevated baseline NIHSS score, baseline ASPECTS <8, and admission glucose ≥140 mg/dL were associated with increased risk of HT after EVT in patients with acute ischemic stroke. In adjusted analyses, diabetes, higher NIHSS score (≥16–20), baseline ASPECTS <8, and elevated admission glucose remained independent predictors of HT. Atrial fibrillation was found to be protective against HT. Elevated baseline serum glucose and higher baseline NIHSS scores were independent predictors of HI-1 and HI-2. Diabetes was an independent predictor of HI-2, and baseline ASPECTS <8 score was an independent predictor of HI-1, HI-2, and SAH. Younger age was associated with HT in unadjusted analysis but was not significant when adjusted for potential confounders in our analysis. There is no consensus regarding the impact of age on HT following intravenous thrombolysis or EVT. The ECASS (European Cooperative Acute Stroke Study) 1 trial demonstrated that advanced age was associated with increased risk of PH.14 Secondary analysis of the ATLANTIS (Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke), ECASS, and NINDS (National Institute of Neurological Disorders and Stroke) trials, also found higher PH rates with increasing age.15 In the Multicenter rt-PA Acute Stroke Survey, increased age was associated with HT and sHT (according to NINDS trial criteria), but did not remain significant when adjusted for potential confounders.16 A systematic review comparing patients ≥80 versus <80 years treated with intravenous r-tPA (recombinant tissue-type plasminogen activator) found similar rates of sHT between both age groups.17 In addition, the NINDS tPA trial reported no significant association between age and sHT.18 In a large observational study from the ETIS (Endovascular Treatment for Ischemic Stroke) registry, age (per 1 year increase) was an independent predictor of PH. The MR CLEAN (Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands) demonstrated that increased age was significantly related to sHT in univariate analysis, but not in multivariable regression analysis.3 Conflicting evidence regarding age as a predictor for HT may be related to use of different definitions of HT. Increased risk of sHT in older individuals could possibly be related to numerous comorbidities. Elevated glucose at admission was an independent risk factor of HT, HI-1, and HI-2, whereas diabetes was an independent risk factor of HT and HI-2 after EVT in our analysis. Diabetes and hyperglycemia at admission have been associated with increased risk of sHT after intravenous thrombolysis, intra-arterial thrombolysis, and EVT.4,19,20 Meta-analysis of HT predictors after EVT also demonstrated that diabetes was a predictor of any HT.5 Failure to decrease blood glucose in the first 48 hours after EVT has shown to lead poor outcomes at 90 days.6 Hyperglycemia accelerates damage to the blood-brain barrier, resulting in edema and hemorrhagic transformation with reperfusion.21 Baseline neurological severity and ASPECTS <8 were independent risk factors for HT, HI-1, and HI-2 in our analysis. Stroke severity, evaluated by NIHSS score, and early detection of major infarct changes on brain computed tomography scan were major predictors of sHT after IV thrombolysis and EVT.7,22 A higher NIHSS score correlates with larger infarct size.23 Although high blood pressure has been associated with increased incidence of sHT after reperfusion therapy, there was no significant correlation between blood pressure and HT in our analysis.8,24 It is possible that many factors influenced systolic and diastolic blood pressure at admission or that blood pressure at hospital arrival was lower compared with initial blood pressure after acute ischemic stroke. Baseline ASPECTS <8, previously linked to any HT and sHT, is also an independent predictor of HI-1, HI-2, and SAH in our analysis. This reflects the greater infarct burden and poorer outcomes typically associated with baseline ASPECTS <8. Importantly, recent trials demonstrate that patients with ASPECTS <6 and large core volumes display functional improvement after EVT compared with medical management, highlighting the clinical relevance of stratifying HT risk in this group.25,26 Although cardioembolic stroke has previously been shown to increase the risk of sHT after EVT, atrial fibrillation was shown to decrease the risk of HT.27 This may be explained by differences in clot composition and procedural efficiency. Thrombi originating from atrial fibrillation are typically fibrin- and erythrocyte-rich, making them more amenable to EVT.28 Reduced vascular manipulation and shorter ischemia-to-reperfusion intervals likely decrease endothelial injury and reperfusion-related damage, both key drivers of HT. In addition, atrial fibrillation–related strokes are less frequently associated with underlying intracranial atherosclerosis, which reduces the need for adjunctive interventions such as angioplasty, stenting, or rescue techniques that can exacerbate microvascular trauma and increase postreperfusion hemorrhage risk.29 Together, these factors may explain the observed protective effect of atrial fibrillation against HT in our cohort. The risk of sHT was highest in PHs (27.3%) followed by patients with multiple hemorrhagic subtypes (18.9%), HI-2 (11.9%), SAH (9.0%), and HI-1 (2.7%). These findings reinforce prior evidence that PHs are strongly associated with neurological deterioration and poor outcomes but also highlight that patients with multiple hemorrhagic subtypes represent an under-recognized high-risk group. The presence and subtype of hemorrhagic transformation were associated with long-term functional outcomes after ischemic stroke. Patients without HT were most likely to regain independence, whereas sHT and PHs were strongly associated with severe disability and mortality. These results reaffirm prior observations that HI-2 and PHs are associated with poor prognoses, while HI-1 and SAH are often compatible with meaningful recovery. This study also highlights the prognostic relevance of grouping patients into multiple hemorrhagic subtypes. Outcomes in this subgroup were comparable to those of HI-2 and PHs, with high rates of disability and mortality. Previous studies have typically analyzed HT subtypes as isolated entities, thereby overlooking the additive risk of overlapping hemorrhagic patterns. By identifying multiple-subtype HT as a distinct prognostic category, our findings suggest that future clinical studies and therapeutic trials should explicitly account for these patients. This refinement may improve prognostic accuracy, guide post-EVT monitoring, and inform patient counseling regarding long-term outcomes. Limitations The main limitation of our study was the relatively small number of patients with PH-1 and PH-2. Consequently, these groups were excluded from the regression models, limiting our ability to identify independent predictors and outcomes specific to these HT subtypes. This exclusion may introduce bias, as PH represents a clinically severe subtype of HT with potentially unique risk factors and prognostic implications. Another limitation is the exclusion of patients with multiple HT subtypes from subtype-specific modeling. The challenge lies in the large number of possible subtype combinations, which makes systematic categorization and meaningful statistical modeling difficult. Future studies should explore alternative approaches to better account for overlapping hemorrhagic subtypes. Conclusions In this analysis of a global, multi-institutional data set, we identified key predictors not only of HT but also of specific HT subtypes in patients with acute ischemic stroke undergoing EVT. Greater severity of pretreatment neurological deficit, baseline ASPECTS ≥8, and higher glucose levels at hospital arrival were independent predictors of HT, HI-1, and HI-2. Diabetes independently predicted HT and HI-2, whereas baseline ASPECTS <8 was the only independent predictor of SAH. Notably, atrial fibrillation emerged as a protective factor against HT. Functional independence was most likely achieved in patients with HI-1 and SAH, whereas HI-2, PHs, and multiple hemorrhagic subtypes were associated with poorer long-term outcomes. Early identification of these HT predictors may enhance prognostication and guide targeted post-EVT monitoring. ARTICLE INFORMATION Disclosures Dr Loehr is a consultant for Penumbra, Phenox, and Stryker. Dr Defreyne is a consultant for Stryker. Dr Zaidat received personal fees for serving as a consultant for Cerenovus, Medtronic, Stryker, and Penumbra outside the submitted work; having a patent for system and device for engulfing thrombi (20190239911). L.L. Price is an employee of Stryker and holds Stryker stock. Dr Gupta serves as Principal Investigator (PI) for the ASSIST registry (Stryker), PI for the RECCLAIM II Study (Zoll), Clinical Events Committee (CEC) for the MIND Trial (Penumbra), Data Safety Monitoring Board (DSMB) Membrane Study (Cerenovus), ELEVATE Study (Medtronic) consultant and stock options for Vesalio and Rapid Medical. Dr Martínez-Galdámez received research grants from Acandis, Balt, Medtronic, Microvention, Phenox, and Stryker (industry payments are made to the research fund of the institution) and receives payment or honoraria for lectures from Balt, Medtronic, and Stryker (industry payments are made to the research fund of the institution). Dr Liebeskind is a consultant for the imaging core lab for Cerenovus, Genentech, Medtronic, Rapid Medical, and Stryker. Dr Zaidat and Dr Liebeskind serve on the editorial board of Stroke: Vascular and Interventional Neurology. Editorial board members are not involved in the handling or final disposition of submissions. The other authors report no conflicts. Supplemental Material STROBE Checklist Supplementary Material