Search

Introduction: As acute stroke therapies expand to less well-resourced regions, it is imperative to study real-world data to assess outcomes and the quality of stroke care. The present study analyzes clinical, and imaging outcomes after acute IV thrombolysis (IV-tPA) compared to mechanical thrombectomy (MT) or combined therapy in the LASE. Methods: A retrospective analysis of consecutive acute ischemic stroke cases in 17 centers from 9 Latin American (LA) countries since 2012 was performed using weighted Euclidean matching of nearest neighbors in 3-D space of baseline NIHSS, age, and glucose. Results: 950 patients receiving only IV-tPA were matched to 127 treated with MT+IV-tPA. Matching resulted in 97 pairs well balanced for age (69.1 vs. 69.3), baseline NIHSS (17 vs. 17), and glucose (124.3 vs. 124.5), all p>0.2. 3-month mRS 0-1 (38.3% vs 29.8%, p=0.23) and mRS 0-2 (46.8% vs 41.5%, p=0.54) were non-significantly higher in the MT+IV-tPA group, with higher hemorrhage (26.6% vs 15.5%, p=0.05) and trends for higher death (15.1% vs 9.5%, p=0.40) and symptomatic hemorrhage (7.4% vs 6.2%, p=1.0). One hundred one patients receiving only MT were matched to 127 patients with MT+IV-tPA, resulting in 61 pairs; MT+IV-tPA showed trends for higher rates of 3-month mRS 0-1 (45.0% vs 35.6%, p=0.31), mRS 0-2 (48.3% vs 42.4%, p=0.54); with trends for lower rates of death (16.9% vs 20.0%, p=1.0) and symptomatic hemorrhage (8.3% vs 11.1%, p=0.75). Conclusions: In this real-world LA sample, trends for better functional outcomes were demonstrated with MT+IV-tPA compared to either treatment alone, as in non-LA populations. A trend for higher adverse events in the combined group requires further investigation.

Background: More than 70% of strokes occur in resource-poor countries. Outcomes are often not well documented. rt-PA for acute ischemic stroke was approved in 2012 for use in Costa Rica (CR). A hub and spoke model was initiated and a dataset established, the CR Stroke Registry Program (CRSRP) for conditional- and post-approval monitoring. Here, we compared CRSRP rt-PA outcomes to similarly treated subjects from the 1995 NINDS rt-PA trial and the 2019 CLOTBUST-ER control arm. Methods: Subjects were matched using a published pairing methodology and day 7-10/discharge modified Rankin Score (mRS), symptomatic intracerebral hemorrhages (SICH) and early mortality compared. A mortality model was generated from 15 randomized controlled trials (RCTs) and outcomes compared at similar baselines. SICH rates were compared with other cohorts: Get With The Guidelines (GWTG), a combined international IV thrombolysis trial pool, and 2 Ibero-American populations. Results: Of 424 CRSRP patients, 284 receiving rt-PA under 3 hrs were matched with 308 NINDS subjects. 131 non-diabetic CRSRP subjects, treated within 4.5 hrs, NIHSS 10 - 24 and Alberta Stroke Program Early CT Score (ASPECTS)>7, were matched with 300 CLOTBUST-ER subjects. Percent achieving either mRS 0-1 or 0-2 did not differ between CRSRP and either NINDS or CLOTBUST-ER (mRS 0-1: CRSRP:33.9% vs NINDS:33.6%; CRSRP:23.8% vs CLOTBUST-ER:27.0%, all p>=.05 / mRS 0-2: CRSRP:40.0% vs NINDS:41.4%; CRSRP:31.1% vs CLOTBUST-ER:36.1%, all p=>.05). Mortality was higher for CRSRP vs CLOTBUST-ER (6.6% vs 0.8%; p=0.05) but not vs NINDS (6.8% vs 4.3%; p=0.3). A predictive model (R 2 =0.39) showed neither cohort exceeded expected pooled mortality, with CLOTBUST-ER the lowest mortality. SICH rate was higher in CRSRP vs CLOTBUST-ER (7.3% vs 0.0% p=0.008) but not vs NINDS (5.7% vs 6.8% p=0.7)). SICH rates were not higher when compared with 4 international cohorts. Conclusion: Functional outcomes of Costa Rican patients receiving rt-PA compared favorably with 2 RCTs (NINDS and CLOTBUST-ER). SICH and mortality were higher than CLOTBUST-ER, although both were within expected range compared to other international cohorts. Systems of care development in order to further lower SICH and participate in the endovascular era are underway.

Background: Application of Machine Learning (ML) techniques in the prediction of Neurological Disorders has been growing rapidly. ML techniques require novel computer programming skills along with domain knowledge to produce a useful tool. We used Google Auto ML and Google ML Tables, a Cloud based front-end tools to develop, validate and deploy a prognostic model for Intracerebral Hemorrhage (ICH) without a single line of programming code. Methods: Data from Kasturba Medical College (KMC) Hospital ICH Stroke Registry between January 2015 and May 2019 was accessed. Google Auto ML Tables splits the dataset into training, validation and testing sets. 80% of the dataset was used for training, 10% for validation, and 10% for testing. Cramers V correlation statistic ( ϕ c ) between 30 variables and outcome was used to select 20 variables to predict the dichotomous outcome of good (mRS 0-3) vs bad outcome (mRS 4-6). The ML model generated was evaluated using Area Under the Curve (AUC) and Receiver Operating Characteristic (ROC) measures and by confusion/error matrix. Logistic Regression model was developed in parallel using Wizard Pro with the same set of 20 variables. Results: 1000 patients from KMC Registry were analysed. Cramers V correlation was highest for Glasgow Coma Scale (GCS) on admission ( ϕ c =0.6), followed by hematoma volume ( ϕ c =0.57). The AUC ROC of the model for good and bad outcome was 0.89 with an accuracy of 85.6%. Confusion matrix predicted 84% patients with good outcome and 87% with poor outcome. Feature importance chart showed that GCS had the highest relative importance at 25.7, followed by hematoma volume and age. A similar model created with simple logistic regression yielded a ROC of 0.87 with comparable results to the machine learning platform. Conclusion: Useful ML models can be easily developed and implemented by clinicians possessing domain expertise using Google Auto ML without computer programming experience.

Introduction: The etiology and significance of diffusion weighted imaging (DWI) lesions in patients with acute intracerebral hemorrhage (ICH) remain unclear. We evaluated which factors were associated with DWI lesions, whether associated factors differed by ICH location, and whether DWI lesions were associated with functional outcomes. Methods: We pooled individual patient data from the MISTIE III trial, the ATACH-II trial, the i-DEF trial, and the ERICH study. We included only patients who underwent protocolized magnetic resonance imaging (MRI) of the brain. A poor functional outcome was defined as a modified Rankin Scale (mRS) score of 4-6 at 3-6 months. We used mixed effects logistic regression with the study database as a random effect. Results: Among 1,775 ICH patients, there were 621 (35.6%) lobar, 978 (55.9%) deep, and 148 (8.5%) infratentorial ICHs. Median time to MRI scan was 1.5 days (IQR, 1-4). DWIHLs occurred in 559 (31.5%) patients, with 190 (34.3%) in lobar ICH and 320 (57.8%) in deep ICHs. In mixed effects regression models, factors associated with DWIHLs included younger age factors associated with DWIHLs after acute ICH included younger age (OR, 0.98; 95% CI, 0.97-0.99), black race (OR, 1.59; 95% CI, 1.18-2.16), admission systolic blood pressure (SBP per 10 mm Hg, OR, 1.13; 95% CI, 1.05-1.22), cerebral microbleeds (OR, 1.71, 95% CI, 1.24-2.35), and leukoaraiosis (OR, 1.60; 95% CI, 1.14-2.25). Patients with DWIHLs had higher odds of mRS 4-6 (OR, 1.57; 95% CI, 1.24-1.99) compared to those without, after adjustment for demographics and ICH severity. In subgroup analyses, similar factors influenced DWIHLs in deep ICH. However, in lobar ICH, younger age, admission SBP, and leukoaraiosis were associated with DWIHLs. Presence of DWIHLs was independently associated with poor mRS in deep ICH but not in lobar ICH. There was no relationship between acute BP lowering and DWIHLs, regardless of location. Conclusions: In a large, heterogeneous cohort of ICH patients, our results are consistent with the hypothesis that DWIHLs represent the effects of chronic hypertensive vasculopathy and acute blood pressure elevation. Furthermore, DWIHLs portend poor prognosis after ICH, particularly in deep hemorrhages.

Introduction: Conflicting evidence has been published regarding the safety and efficacy of ultrasound-enhanced thrombolysis (sonothrombolysis) in acute ischemic stroke (AIS) patients with large vessel occlusion (LVO). Methods: We conducted an individual participant data meta-analysis of available randomized controlled trials (RCTs) comparing sonothrombolysis with or without addition of microspheres (treatment group) to intravenous thrombolysis alone (control group) in AIS patients with LVO. Results: We included 6 in total RCTs that enrolled 1077 AIS patients. A total of 138 and 134 confirmed LVO patients were randomized to treatment and control groups respectively (median age 68 years, 58% men, median baseline NIHSS score 16). Patients randomized to sonothrombolysis had increased odds of complete recanalization compared to patients receiving intravenous thrombolysis alone (40.3% vs. 22.4%; OR=2.30, 95%CI: 1.05-5.02; adjusted OR=2.33, 95%CI: 1.02-5.34). They also tended to have increased odds of any (complete or partial recanalization (66.4% vs. 53.0%; OR=1.78, 95%CI: 0.95-3.33; adjusted OR=1.85, 95%CI: 0.97-3.53). The likelihood of symptomatic intracranial hemorrhage did not differ between the two groups (7.3% vs. 3.7%, OR=2.52, 95%CI: 0.77-8.29; adjusted OR=2.55, 95%CI: 0.76-8.52). No differences in the likelihood of asymptomatic intracranial hemorrhage (adjusted OR: 1.30, 95%CI: 0.38-4.39), three-month mortality (adjusted OR: 1.23, 95%CI: 0.25-6.05), three-month favorable functional outcome (mRS-scores of 0-1; adjusted OR: 1.43, 95%CI: 0.64-3.19) and three-month functional independence (mRS-scores of 0-2; adjusted OR: 1.43, 95%CI: 0.77-2.64) were documented. Conclusion: Sonothrombolysis was associated with a two-fold increase in the odds of complete recanalization compared to intravenous thrombolysis alone in AIS patients with LVOs. Further study of the safety and efficacy of sonothrombolysis is warranted.

Background: Patients suffering from subarachnoid hemorrhage (SAH) have poor long-term outcomes. There are predictive models for ischemic and hemorrhagic stroke. However, there is paucity of models for SAH. Machine learning concepts were applied to build multi-stage Neural Networks (NN), Support Vector Machines (SVM) and Keras/Tensor Flow models to predict SAH outcomes. Methods: A database of ~800 aneurysmal SAH patients from Kasturba Medical College was utilized. Baseline variables of World Federation of Neurosurgeons 5-point scale (WFNS 1-5), age, gender, and presence/absence of hypertension and diabetes were considered in Stage 1. Stage 2 included all Stage 1 variables along with presence/absence of radiologic signs vasospasm and ischemia. Stage 3 includes earlier 2 stages and discharge Glasgow Outcome Scale (GOS 1-5). GOS at 3 months was predicted using 2-layer NN/SVM/Keras-TensorFlow models on the five point categorical scale as well as dichotomized to dead/alive and favorable (GOS 4-5) or unfavorable (GOS 1-3). Prediction accuracy of models was compared to the recorded GOS. Results: Prediction accuracy shown as percentages (See Table) for all three stages was similar for SVM, NN and Keras/TensorFlow models. Accuracy was remarkably higher with dichotomization compared to the complete five point GOS categorical scale. Conclusions: SVM, NN, and Keras-TensorFlow based machine learning models can be used to predict SAH outcomes to a high degree of accuracy. These powerful predictive models can be used to prognosticate and select patients into trials.

Background and Purpose— Observational data suggest that antiplatelet therapy after intracerebral hemorrhage (ICH) alleviates thromboembolic risk without increasing the risk of recurrent ICH. Given the paucity of data on the relationship between antiplatelet therapy after ICH and functional outcomes, we aimed to study this association in a multicenter cohort. Methods— We meta-analyzed data from (1) the Massachusetts General Hospital ICH registry (n=1854), (2) the Virtual International Stroke Trials Archive database (n=762), and (3) the Yale stroke registry (n=185). Our exposure was antiplatelet therapy after ICH, which was modeled as a time-varying covariate. Our primary outcomes were all-cause mortality and a composite of major disability or death (modified Rankin Scale score 4–6). We used Cox proportional regression analyses to estimate the hazard ratio of death or poor functional outcome as a function of antiplatelet therapy and random-effects meta-analysis to pool the estimated HRs across studies. Additional analyses stratified by hematoma location (lobar and deep ICH) were performed. Results— We included a total of 2801 ICH patients, of whom 288 (10.3%) were started on antiplatelet medications after ICH. Median times to antiplatelet therapy ranged from 7 to 39 days. Antiplatelet therapy after ICH was not associated with mortality (hazard ratio, 0.85; 95% CI, 0.66–1.09), or death or major disability (hazard ratio, 0.83; 95% CI, 0.59–1.16) compared with patients not started on antiplatelet therapy. Similar results were obtained in additional analyses stratified by hematoma location. Conclusions— Antiplatelet therapy after ICH appeared safe and was not associated with all-cause mortality or functional outcome, regardless of hematoma location. Randomized clinical trials are needed to determine the effects and harms of antiplatelet therapy after ICH.

Background: A predictive outcome model for intracranial hemorrhage (ICH) from patient level data would be of interest in planning clinical trials. Despite decades of effort and several different types of models attempted, there is no universally accepted model. Here, we sought to apply Information Theory precepts to select the optimum set of baseline variables that can predict outcomes using data from the control arm of ATACH-II, a recently concluded ICH trial that tested the effects of control of systolic BP (SBP). Methods: Sixty-three logistic regression models of mortality and good outcome (modified Rankin Score of 0-3 at 90 days) were generated based on all possible combination of 6 baseline variables: NIH Stroke Scale (NIHSS), Glasgow Coma Scale (GCS), age, hematoma volume (Hem Vol), intraventricular hematoma volume (IVH Vol), and SBP. Akaike Information Criterion (AIC) a measure of information content in a model was calculated. Those with the lowest AIC are considered best. Results: ATACH-II had 405 subjects with non-lobar hemorrhages that reported all six baseline variables and long-term outcomes. The best model of good outcome was based on NIHSS, age, Hem Vol, IVH Vol (R 2 = 0.33; p <2 = 0.18; p < Conclusion: We show that Information Theory constructs can be used to select the best set of variables to model outcomes in ICH. While we found generally better predictive ability for good outcome compared to mortality, the best model predicted only 57% of the outcome variance. While additional types of data fitting might yield additional predictive value, the relatively modest results seen here suggest that there remain additional factors related to outcome not typically measured in ICH trials. Additional data points from other trails are being pooled to see if predictive ability improves.

Background and Aims: Spontaneous subarachnoid hemorrhage (sSAH) is a rare stroke event with a relatively high rate of mortality. There are a number of risk factors, including gender and race, for sSAH. However, it is unclear whether these risk factors are associated with the age of mortality due to sSAH. Methods: We utilized the CDC Wonder dataset (wonder.cdc.gov) “Underlying Cause of Mortality 1999-2016” to query for sSAH mortality in named arterial distributions. Age, race and other demographic data from the entire United States was available. Statistical differences between populations of different races including Caucasian African American, Asian, or Native American or between genders were assessed using a Chi-squared test of independence and unpaired Student’s t tests. Results: Our analysis detected an association between race and age range of mortality due to sSAH (χ 2 = 414.00, df = 51, p -value = 1.6E-58). Further analyses revealed a younger age of mortality due to sSAH in males as compared to females, with an average age of mortality at 53.3 and 60.0 years old respectively ( p < 0.0001). Notably, race differences were greater in females than in males, whereby black females exhibited the youngest age of mortality due to sSAH with an average age of mortality at 51.8 years as compared to 61.6 for non-black females ( p < 0.0001). Conclusion: These findings suggest that underlying differences in race and gender have an effect on age of mortality due to sSAH.

Introduction: Many randomized clinical trials (RCTs) have been conducted for subarachnoid hemorrhage (SAH), yet no treatment definitively improves outcome. Predictive ability of early phase trials is uncertain because outcome is strongly related to baseline severity factors, yet imbalances are common in these trials. We previously developed a method for ischemic stroke to identify promising treatments at early phase using a pooled model constructed from the control arms of RCTs and comparing trial outcomes at their own baselines. We applied this method to SAH. Methods: SAH predictive models for functional outcome and mortality were developed based on published variables (eg, WFNS, Fisher, age). The outcome model is the best fit 3-d surface bounded on either side by p=±0.1 prediction interval (PI) surfaces. These surfaces incorporate statistical variability to assess whether a treatment differs from expected outcome of a pooled sample. Treatment arms from RCTs and single arm trials for SAH are then visually compared against the pooled controlled arm to see if they surpassed the prediction surfaces at their own baseline conditions. Results: The best model fit was for good outcome (modified Rankin Score 0-2) based on WFNS grade 4-5% and age (26 RCTs, 5200 subjects; R 2 =0.54; p Conclusions: Outcome models based on percentage of high grade WFNS and age were successfully developed. This approach may be useful to prioritize treatments worthy of further study.

Introduction: Many stroke patients do not fully respond to intravenous rt-PA, possibly due to factors that increase clot resistance or impede reperfusion. GPIb receptors on the platelet surface initiate vessel occlusion through an interaction with subendothelium-bound von Willebrand Factor and the interaction transduces signals resulting in further activation. An increase in GPIb availability may result in macro or micro-vascular thrombosis. We report the effect of rt-PA on a potential compensatory conformational response of platelet GPIb. Methods: 5 subjects with large artery stroke and treated with rt-PA within 4.5 hours of onset were enrolled after consent. Blood was obtained immediately before and 1.3 hours post rt-PA bolus. Blood was incubated with 3 monoclonal antibodies (mABs: AN51, 6D1 and SZ2) and control antibody. AN51 and 6D1 are conformation sensitive and SZ2 is relatively less conformation sensitive. Flow cytometry was performed to assess fluorescent intensity of platelets pre and post-treatment and corrected for non-specific staining with control IgG. Blood from healthy donors subjected to appropriate concentrations of rt-PA in vitro was tested for conformational changes with AN51 and 6D1. Results: Compared to pre-treatment, 4 of 5 patients showed increased GPIb or binding with AN51 and 6D1 post treatment, with an overall median percentage increase with AN51 of 40.6% and 60.3% with 6D1. SZ2 showed a smaller increase (30.7%). The 1 subject who showed no change had the only hemorrhagic transformation. Blood from healthy donors showed an increase of 21.4% with AN51 and 41.3% with 6D1 after rt-PA. Conclusion: These results indicate for the first time rt-PA induced conformation changes on platelet GPIb in acute ischemic stroke in vivo despite that GPIb is not the direct target. These changes were also seen with rt-PA in vitro. It is possible that up-regulation of GPIb receptors plays a protective role in preventing hemorrhages but also may mediate increased propensity for clot resistance or downstream thrombosis. If confirmed in a larger ongoing sample, measuring GPIb binding may provide a rational approach to post-rt-PA antiplatelet therapy and identify a risk for hemorrhagic transformation.

Introduction: Failures to translate pre-clinical results have been discouraging. We have contended that stroke is too heterogeneous with respect to factors influencing outcome to expect small studies to be balanced. It is not only difficult to control for biological and methodological variability but efforts to improve homogeneity, such as minimizing physiological variability, may render results less applicable to humans. Here, we report a predictive outcome model in experimental stroke which incorporates baseline variability and provides statistical thresholds a treatment must exceed to be efficacious in a broad population. Methods: We generated a mathematical model to predict outcome using transient MCA occlusion in 23 unfasted rats. To create baseline variability, we varied occlusion times from 90-120 min, altered baseline glucose with streptozotocin, and assessed neurological outcome 3 days later with a modified Bederson Score (BS; 0-6 functional measure, 7 death). Statistical surfaces in 3 dimensions were generated using Jacobian matrices flanking the model to provide a screening threshold (1 SD) for comparing new therapies against this model. Results: We successfully generated an outcome model from occlusion time, glucose and BS (Fig; R 2 =.49, p=.0003; middle surface is the model surrounded by ±SD surfaces). Outcome was sensitive to change in glucose and time, suggesting small imbalances in these factors between groups may influence outcome, and hence the perceived efficacy of a new therapeutic intervention. At normoglycemia and 90 mins, the lower surface overlapped with no deficit, indicating it would be difficult to reliably demonstrate benefit under those conditions. Conclusions: These results indicate it is feasible to incorporate biological variability to generate more clinically relevant conditions. The method will be tested with other stroke models and modifiers towards a generalized model to screen for therapies worthy of further study.

Introduction: While oxidative stress mediates many of the pathologies in stroke, antioxidant therapies have been unsuccessful. Current antioxidants are limited either by a narrow range of radicals quenched or by low capacity. Some can yield toxic intermediates. We developed a new class of antioxidant, highly modified carbon nanoparticles called PEGylated hydrophilic carbon clusters (PEG-HCCs). These 40 nM particles quench superoxide by a high capacity novel catalytic mechanism (PNAS 2015: 2343-8), are effective against - OH and recouple nitric oxide synthase. Here we report the effect of PEG-HCCs administered during reperfusion after transient MCA suture occlusion in the rat under varying glucose levels. Hyperglycemia occurs in 1/3 of stroke patients and worsens outcome in part by accelerating release of radicals during reperfusion. Methods: Variation in glucose was induced by streptozocin IP 2 days prior to MCAO. At 90 mins, 23 rats were injected IV with PBS or PEG-HCCs (4 mg/kg) and the suture removed. A second dose was given 2 hours later (dosing based on half-life and in-vitro efficacy). At 3 days, infarct size corrected for edema and neurological function (Bederson Score, BS) was obtained. An outcome curve was developed using a range of baseline glucose in controls and the effect of PEG-HCCs compared at rat’s baseline glucose values. Results: We found a strong sigmoidal relationship between infarct size and glucose level in control rats (Fig. R 2 =.93; p=.0001; blue curve: control arm ± 95% confidence intervals). All PEG-HCC treated rats showed infarct size below the confidence interval (red stars; p Conclusions: This novel, broadly active, high capacity antioxidant was able to improve outcome in a severe test of pre-clinical stroke at clinically relevant times. Studies are ongoing to assess the treatment window and longer term outcomes.

Background and Purpose— Sex and race reportedly influence outcome after recombinant tissue-type plasminogen activator (rtPA). It is, however, unclear whether baseline imbalances (eg, stroke severity) or lack of response to thrombolysis is responsible. We applied balancing methods to test the hypothesis that race and sex influence outcome after rtPA independent of baseline conditions. Methods— We mapped group outcomes from the National Institute of Neurological Disorders and Stroke (NINDS) dataset based on race and sex onto a surrogate-control function to assess differences from expected outcomes at their respective National Institutes of Health Stroke Scale and age. Outcomes were also compared for subjects matched individually on key baseline factors using NINDS and 2 recent datasets from southeastern United States. Results— At similar National Institutes of Health Stroke Scale and age, 90-day good outcomes (modified Rankin Score, 0–2) in NINDS were similarly improved after rtPA for white men and women. There was a strong trend for improvement in black men. Conversely, black women treated with rtPA showed response rates no different from the controls. After baseline matching, there were nonsignificant trends in outcomes except for significantly fewer good outcomes in black versus matched white women (37% versus 63%; P =0.027). Pooling the 3 datasets showed a similar trend for poorer short-term outcome for black women ( P =0.054; modified Rankin Score, 0–1). Conclusions— Matching for key baseline factors indicated that race and sex influence outcome most strikingly in black women who demonstrated poorest outcomes after rtPA. This finding supports the hypothesis that poor response to rtPA, rather than differences in baseline conditions, contributes to the worse outcome. This finding requires prospective confirmation.

Background: While stent-retriever randomized clinical trials (RCTs) were uniformly positive, their designs differed. Here, we assessed whether trials also differed in the magnitude of improvement and if factors could be identified that distinguished the best performers. Because % utilization of rt-PA in the control arms varied, we developed a multi-dimensional outcome model from the control arms of all IV tPA RCTs treating various % of subjects and their baseline stroke severity (NIHSS). This model was used to compare each trial’s functional outcome (mRS < 2) and mortality at their own baseline without need for statistical manipulation. We assessed factors that potentially distinguished the best performers. Methods: The model was developed from 54 RCT control arms representing >11,000 subjects treating between 0-100% with IV rt-PA. 3-D non-linear functions were fit to the 90-day mortality and mRS 0-2. This model includes ±95% statistical interval surfaces to assess whether a trial’s outcomes surpass the variability inherent in a sample of RCT subjects (Neurology 85:274-83, 2015). 6 stent retriever RCTs were compared. Results: The models showed excellent fit: mRS 0-2 (r 2 =0.83, p<2 =0.65, p< Conclusion: A new outcome model based on stroke severity and % IV rt-PA permitted analysis of stent-retriever therapy compared to a large population. Best outcomes were related to shorter treatment time and higher utilization of rt-PA, suggesting a treatment interaction between these modalities. Factors influencing mortality are being analyzed.

Background and Purpose— Preclinical and retrospective clinical data indicate that glyburide, a selective inhibitor of sulfonylurea receptor 1-transient receptor potential melastatin 4, is effective in preventing edema and improving outcome after focal ischemia. We assessed the feasibility of recruiting and treating patients with severe stroke while obtaining preliminary information on the safety and tolerability of RP-1127 (glyburide for injection). Methods— We studied 10 patients with acute ischemic stroke, with baseline diffusion-weighted imaging lesion volumes of 82 to 210 cm 3 , whether treated with intravenous recombinant tissue-type plasminogen activator, age 18 to 80 years, and time to RP-1127 ≤10 hours. Results— Recruitment was completed within 10 months. The mean age was 50.5 years, and baseline diffusion-weighted image lesion volume was 102±23 cm 3 . There were no serious adverse events related to drug and no symptomatic hypoglycemia. The increase in ipsilateral hemisphere volume was 50±33 cm 3 . The proportion of 90-day modified Rankin Scale ≤4 was 90% (40% modified Rankin Scale, ≤3). Conclusions— RP-1127 at a dose of 3 mg/d was well tolerated and did not require any dose reductions. A clinical trial of RP-1127 is feasible. Clinical Trial Registration— URL: http://www.clinicaltrials.gov . Unique identifier: NCT01268683.

BACKGROUND: Outcome following ischemic stroke is strongly related to baseline severity. Differences in baseline balance between arms seen even in larger trials complicate analysis. We developed a method (pPREDICTS) that generates a pooled outcome model from baseline factors taken from randomized clinical trials (RCTs). It visualizes variability to assess whether an individual study treatment arm differs sufficiently from expected outcome to be considered positive if tested in a larger population, without need for any statistical corrections. Here we assessed whether intracerebral hemorrhage (ICH) could be similarly modeled and how blood pressure (BP) interventions fared considering between-trial variance. METHODS: pPREDICTS generates an outcomes and provides multi-dimensional ±p=.05 statistical surfaces to visualize the variance among these trials. ICH RCTs were selected that provided baseline NIHSS, age and 90 day outcomes. Individual results from 4 BP trials were tested against these models to determine whether outcomes surpassed the ±95% prediction interval surfaces. Mortality and mRS0-2, the most frequently reported outcomes were modeled. RESULTS: Outcome functions derived from 12 ICH RCTs representing 2766 subjects were generated (middle surface ± p=.05 surfaces; Fig 1a: Mortality: r 2 =0.84; Fig 1b:mRS0-2:r 2 =0.79; p≤0.005). Control management (green) and treatment arm (red) outcomes of ATACH, ICH-ADAPT, INTERACT and INTERACT2 were plotted (see Figs). No mortality or functional outcome different from control was seen as outcomes are mostly distributed along the middle surface. However, tier 3 of ATACH, BP goal of SBP 110-140 showed increased mortality (>+95%) and decreased mRS 0-2 outcome ( CONCLUSION: Outcome models for mortality and mRS 0-2 were successfully developed. No benefit was seen from any BP intervention. We suspect small differences in outcomes reported as positive are more likely random variation.

BACKGROUND: Although a phase 3 trial using cardiac dose of the GP IIb/IIIa inhibitor abciximab was halted, we previously reported that a 30% lower total intravenous (IV) abciximab dose achieved optimal platelet inhibition (92%) and showed benefit in patients ineligible for tPA when compared using a rigorous matching process to historical controls (Mandava et al, TSR 2010). Given lack of proven alternatives, we were interested in abciximab specifically in the posterior circulation. METHODS: 0.2 mg/kg IV bolus followed by 0.05 ug/kg/hr infusion over 12 hrs of abciximab was administered with suspected large vessel posterior circulation occlusion after informed consent and presentation of alternatives. Patients were ineligible for tPA mostly from time window (24 hrs, later changed to 12 hrs) or elevated INR < 2.5. A weight based heparin protocol was used in most patients to achieve PTT 46-70. Ninety day follow up was obtained. To assess benefit, patients’ baseline characteristics were compared to the NINDS control arms subjects restricted to large vessel and cardioembolic subtypes. Matching was performed for 5 hour baseline NINDS NIHSS, age, glucose, and posterior stroke. An excess of posterior strokes in the abciximab group permitted 2:1 matching with outlier elimination. McNemars test compared 90 day mRS 0-1/0-2, mortality, and hemorrhage RESULTS: 26 subjects were treated in 3 university centers. Post-match baseline variables were slightly more severe in the abciximab group and they were treated later (Table). 90-day functional outcomes were higher in the abciximab group (mRS 0-1:54.5% vs 23.1%, mRS 0-2: 63.6% vs 30.8%; p CONCLUSION: Low dose IV abciximab in extended time window for large vessel posterior stroke shows promise. Given lack of proven alternatives, these results need to be verified prospectively.

BACKGROUND: While intravenous thrombolysis (IVT) remains the only established therapy for ischemic stroke including beyond 3 hours, there is considerable room for improvement in IVT outcomes. Non-randomized series suggested better outcomes for add-on therapies; however randomized clinical trials (RCTs) were often equivocal. One reason for these discrepancies is that imbalances in important baseline factors between treatment and control arms affect all but the largest trials and can produce misleading results. We developed techniques to identify potentially efficacious therapies by generating a pooled outcome model. Here we tested RCTs attempting to improve upon IVT against this model at each treatment arm’s baseline NIHSS and age. METHODS: We generated outcome models (mortality, mRS 0-1, mRS 0-2) pooled from the thrombolytic arms of all IV tPA RCTs with the novel feature of multi-dimensional statistical intervals. A function was iteratively fitted and individual results of IV, add-on and endovascular arms tested against this model to determine whether outcomes surpassed the + p< .05 surface (see mRS0-2 Fig.). RESULTS: Functions derived from 19 IVT RCT thrombolytic arms representing 2748 subjects were generated (Mortality: r 2 =0.51; mRS0-1: r 2 =0.79; 0-2: r 2 =0.85). No treatment increased mortality. Three hour IVT/ultrasound with (TUCSON) or without (CLOTBUST) microspheres, the neuroprotectant Cerebrolysin and IA tPA alone (SYNTHESIS) had better functional outcomes, as did CTP-guided 6 hr. IV tenecteplase. Notably, 4.5 hour SYNTHESIS-EXPANSION (IA tPA) and IMS III (IV tPA + endovascular) did not. No extended window device showed benefit. CONCLUSION: We found that several treatments show the potential to improve outcome compared to our pooled IVT model, including 3 hour ultrasound, IA tPA and 6 hr. tenecteplase. Consistent with the cardiac experience, there remains little evidence that the 3 hour window can be extended other than by an IV thrombolytic.

Background and Objective: Thrombin inhibitors such as Argatroban in combination with full-dose IV-tPA are currently being tested in a series of clinical trials for acute ischemic stroke. A single-arm, phase IIa study (ARTSS-1) of 65 patients treated with low-dose Argatroban + IV-tPA in either the 3 or 4.5 hour treatment window demonstrated an overall sICH rate of 4.6% and 40% complete recanalization within 2-hours. The objective of the study was to assess safety in a matched group of IV-tPA alone patients. Methods: pPAIRS © performs a post-hoc Euclidean matching of subjects on factors of choice and excludes outliers when no close match is found. We applied these techniques to compare outcomes of Argatroban+tPA treated subjects from ARTSS-1 with tPA treated subjects from NINDS. Matching factors were NIHSS, age, glucose, race, gender, hemisphere, and subtype. Outcomes were sICH, death and mRS 0-1, 0-2 at day 7 and 90. Results: Of the 65 Argatroban+TPA patients, 55 and 44 had good NINDS matches at day 7 and 90, respectively. Rates of sICH were identical at both day 7 and 90 between ARTSS-1 and NINDS (3.6% and 2.3%). Mortality was not significantly different (5.5% vs. 3.6% at day 7; p=1.00; and 11.4% in each group at day 90). There was a non-significant trend for higher percent of mRS 0-1 (36.4% vs. 32.0%, p=0.814) and 0-2 (47.7% vs. 43.2%, p=0.823) at day 90 in the ARTSS-1 patients that was not present at day 7. The mRS 0-2 trend remained after excluding ARTSS-1 patients treated with tPA between 3-4.5 hours (see table). Conclusion: There is no evidence for significantly increased mortality or symptomatic hemorrhage after addition of low dose Argatroban compared to a NINDS population with similar baseline characteristics particularly in those tPA-treated within 3 hours. A non-significant trend emerged at 90 days for 14% relative improvement in functional outcomes. A larger, Phase IIb, randomized trial (ARTSS-2) is ongoing to determine whether these trends reflect a true treatment effect.

Background: Proven options to treat ischemic stroke patients ineligible for IV t-PA are limited. Our focus is on treatments to spare the penumbral microcirculation. Evidence indicates that Glycoprotein (GP) IIb/IIIa inhibitors when given with heparin, as used in coronary intervention, increases macro- and distal micro-vascular patency in the heart and animal stroke models. We hypothesized that the cardiac dose of IV abciximab (ABX) used in the halted AbESTT trial was too high leading to hemorrhage, while exclusion of heparin reduced efficacy. We previously showed that a 30% lower ABX dose achieved target GP IIb/IIIa platelet inhibition of 92% in stroke patients and hypothesized a lower dose when given with short term heparin would be safe and effective. Methods: After consent, patients with non-lacunar stroke ineligible for t-PA were offered low dose IV ABX (bolus:0.2 mg/kg & infusion:0.05 ug/kg/hr) along with IV heparin (36 hrs) for ant. circulation strokes within 6 hrs and post. within 12 hrs of onset. In place of a control arm, we adapted our pPAIRS matching strategy to obtain best matches from non-lacunar placebo patients in the NINDS dataset at the 2 hr post-randomization NIHSS. Age, glucose and gender were other factors that were matched. Results: Forty five patients have been treated. t-PA ineligibility was mostly due to time of onset, recent surgery or anticoagulation. Mean time to treatment was 5.2 h (Ant:4.1h; Post:9.3h). Matching with NINDS placebo subjects was excellent: median NIHSS: ABX: 18 vs NINDS: 17; age: ABX: 66.6 ± 13 vs NINDS: 66.3 ± 11; and glucose: ABX:148 ± 58 vs NINDS:144 ± 52; all p>0.05. There was 1 symptomatic hemorrhage in ABX patients (2.2%) vs none in NINDS; p=1.0. Ninety day mRS 0-1 and 0-2 in the ABX arm (42.2% and 55.6%) were higher than in NINDS matched placebo subjects (15.6% and 31.1%); all p.05. Conclusion: Low dose ABX appeared safe and potentially effective when compared with NINDS placebo subjects assessed 2 hrs after their initial presentation so as to provide a comparable baseline. Since low dose ABX achieves sufficient platelet inhibition and the combination with short term heparin appeared well tolerated, we suggest these findings deserve confirmation in a prospective RCT.

Introduction: African American (AA) women have been reported to have higher mortality after stroke. Our prior analysis of the NINDS dataset also showed poorer outcomes after tPA for AA compared to Caucasian (C) females. Small numbers of AA patients available in a well characterized dataset and large differences in baseline factors hamper definitive conclusions about race-based outcomes. We developed analytical techniques for small samples that explicitly consider baseline factors rather than adjust or control for them. We applied these techniques to two stroke datasets and analyzed them separately and together with NINDS. Methods: pPREDICTS uses the placebo arms of all eligible stroke RCTs to generate a pseudo-control outcome model. The novel feature is generation of multi-dimensional statistical intervals to compare whether an individual trial’s outcome differed from this pooled sample at the trial’s specific baseline NIHSS and age. pPAIRS performs post-hoc Euclidean matching of subjects based on chosen factors and excludes outliers when no close match is found, generating a more homogenous population. We applied these techniques to assess outcomes of tpa-treated females from the NINDS, Tulane, and University of Alabama-Birmingham datasets. Matching factors were baseline NIHSS, age, glucose, race and stroke subtype. Outcomes were mRS 0-1, 0-2 and mortality at discharge or 90 days. Results: Pooled results from individual sites are reported here. Comparison of outcomes from each site with the pPREDICTS model did not show better outcomes for the newer datasets compared to NINDS. mRS 0-2 for tPA-treated AA females was near the p=.05 threshold for poorer outcomes compared to pooled controls. After outlier exclusion, 80 matched AA and C females were found by pPAIRS. After matching, median baseline NIHSS were similar: 12.5 for AA’s and 11.5 for C (p = 0.64). Final mRS 0-1 and mRS 0-2 for AA females were lower: 22.5% and 30% vs. 40% (p = 0.03) and 46.3% for C (p = 0.05). Mortality was 10% in AA and 7% in C (p=0.6). Conclusion: We confirmed in this larger pooled sample that AA females have a poorer functional outcome following tPA after baseline matching. Factors associated with these poor outcomes are being assessed and will be presented. A prospective study is warranted.

Background: There remains debate as to the most appropriate functional outcome measurement in acute stroke trials. One approach is to use the whole range of ordinal/categorical variables of the modified Rankin Score (mRS), often termed “Shift Analysis”. Here we consider one of its main theoretical bases, Shannon’s information theory. Noise in Shannon’s model assumes the signal transmitted is uniform. However, van Swieten (Stroke 1988), and others show that the errors (noise), in the mRS are predominantly in the mid-range values of 2-4. In order to test the applicability of Shannon’s concept to shift analysis, we used entropy models to quantify errors with the primary noise considered in this range. Methods: We identified 16 randomized stroke trials that reported the full range of mRS. A custom Matlab® program was written to simulate Shannon’s model both when the whole range of mRS was used and with commonly used dichotomous cut-points of mRS0-1 or mRS0-2 and mRS 6. Each stroke trial’s mRS distribution was multiplied with the “confusion matrix” from Table 1 of van Swieten’s mRS inter-rater variability. This matrix is equivalent to a band-limited noise in the signal-processing sense. The output of this step produced a misclassification/error percentage for each mRS category both from the full range of mRS or with specified cut-points. Error rates were compared by ANOVA. Results: If the full range of mRS is considered, error rates ranged from 22 to 32% (Mean ± SD: 26% ± 2.8). If mRS 0-1 and mRS 6 were cut-points, then error ranged from 2.6 to 13.4% (7.5% ± 2.5). Error rates for mRS 0-2 and mRS 6 were 6.9 to 14.3% (8.8% ± 1.8; overall p Conclusion: We show from an information theory perspective that when the uncertainty in the midrange mRS is considered, there are relatively high error rates if the full range is used. Lower error rates were found with specified cut off dichotomization. Federov et al (2009) contend there is loss of information with dichotomization. However, this view does not consider the non-random distribution of noise and it assumes a Poisson distribution of mRS values. As seen in the left figure , none of the 16 trials had a Poisson distribution (compare to right panel modeling different Poisson distributions). We conclude that while use of entire mRS range is conceptually appealing, even if these were Poisson distributions, any potential gain of information with considering the complete scale is accompanied by a decrease in fidelity, hence more uncertainty regarding the validity of the outcome compared to dichotomization.

Background: Stroke outcome is dependent on baseline factors such as NIHSS and age. Relationships between these variables and outcomes are often non-linear and imbalances can influence outcomes, particularly in subgroup analysis with smaller number of subjects. Balance in baseline variables factors are typically compared by Wilcoxon rank sum, t-test or ANOVA. Because of non-linearity, these tests may be insensitive to important differences in the distribution of these factors and if multiple factors are considered simultaneously. We adapted a multi-dimensional extension of Kolmogorov-Smirnov (KS) test proposed by Fasano and Franceschini (FF) to compare population distributions. The FF algorithm provides a method to calculate KS Distance (KSD) between two distributions in multiple dimensions and a probability value can be obtained. We hypothesized that the FF algorithm would be more sensitive than traditional statistical tests to determine whether baseline factors differed among two trial arms. We further show that matching for baseline variables (nearest neighbor Euclidean matching, pPAIRS©; Mandava Kent Stroke 2010) improves the KSD, indicating closer matched populations. Methods: The NINDS database was used for this study ( ntis.gov ). The subgroup of rt-PA and placebo treated normoglycemic subjects with large artery stroke was analyzed. Median and mean NIHSS and age were compared and KSD and a p value were calculated using a custom program, pPOPULATION© written in Matlab®. rt-PA and placebo subjects were then matched using pPAIRS© and outliers eliminated. KSD and p value for the post-matched groups were calculated. Results: The left half of the table shows the pre-match comparisons. Baseline variables were not different using usual tests. A KSD value of 0.283, however yielded p=0.008, suggesting that the population distributions are indeed different when two variables are considered simultaneously. Right half of the table shows the post-match comparisons of baseline variables. The KSD value, 0.217, is lower and is associated with a p value = 0.175, indicating that the post-matched distributions are similar. a Wilcoxon Rank-Sum; b Student t-test; Conclusion: We demonstrate here a new application of a 2d version of the KS distance to verify the similarity of stroke populations and show that it is more sensitive than traditional difference testing. This finding is important because baseline imbalances are critical for accurate assessment of outcome. This algorithm can be further extended to additional dimensions (e.g.: glucose). Its relative advantages over other methods will be discussed.

Introduction: Although intra-arterial/endovascular (IA) stroke therapies are widely employed, there has been a paucity of randomized clinical trials (RCT) to support their use. In lieu of RCT’s, case series are often compared to each other or to other RCT’s. However, stroke outcome is highly dependent on baseline factors such as severity and age. Because those factors have a complex relationship, comparisons are often unreliable. Even with the inclusion of a control arm, imbalances often complicate valid conclusions. In order to provide meaningful comparisons, we developed pPREDICTS©, that generates a surrogate control outcome function based on the baseline factors and outcomes of the placebo arms of published RCT’s (Mandava & Kent Stroke 2009). pPREDICTS© now includes 24 RCT’s representing 7400 subjects and explains more than 80% of the variance of mortality and functional outcome (mRS). In this study, we applied pPREDICTS© to published IA series and calculated benefit/risk for each based on differences from predicted outcome at comparable baseline NIHSS and age. We performed the same analysis with intravenous (IV) thrombolytic RCT’s in order to gauge relative efficacy. Methods: We identified 51 IA and 22 IV studies that included baseline NIHSS, age, time window, and outcomes (mRS 0-2 and mortality) and with N > 10. Proportions of subjects that died or achieved mRS 0-2 were subdivided by time window. Publication bias, the apparent greater benefit in those trials with smaller N, was corrected by ‘trim and fill’ using the pPREDICTS© outcome as the surrogate baseline (Mandava & Kent Neurology 2007). Data is represented as mean + 95% confidence interval. Results: The IA series had a higher baseline NIHSS than the IV trials (17.7+3.2 vs 12.5+3.2; pFigure ) up to 6 hour time period. However, beyond 6 hours there was excess mortality seen with IA therapy. Conclusion: Unlike other systematic or meta-analysis, by comparing individual treatment arms to a continuous placebo outcome model at baseline factors similar to their own we avoided the need to adjust for imbalances. Decreasing, but potential benefit of both IA and IV therapies was found up to 6 hours, but with higher mortality for IA at > 6 hrs. Analysis will be presented of which clinical factors predicted the best outcomes.

Background and Purpose— Outcome from stroke is highly dependent on baseline conditions. Patients with stroke have a wide range of severities, ages, and etiologies and it has proven difficult to achieve randomization of key variables in clinical trials. We present a new post hoc approach to achieve balance among selected variables. To illustrate the approach, we rebalanced the National Institute of Neurological Diseases and Stroke Recombinant Tissue Plasminogen Activator trial, in which the contribution of baseline imbalances continues to be debated. Methods— We selected baseline stroke severity (National Institutes of Health Stroke Scale), age, and glucose as matching criteria. The closest matched placebo and treated subjects were identified based on nearness to each other in 3-dimensional Euclidean space. Matching was performed within the quintiles of National Institutes of Health Stroke Scale that have been previously used to assess balance. Subjects who could not be matched were eliminated. Outcomes were assessed using the original specified National Institute of Neurological Diseases and Stroke trial measures. Results— We successfully matched the 2 arms resulting in nearly identical baseline characteristics and distribution among quintiles. Despite fewer subjects after outlier elimination, the primary outcome measures remained significantly improved. After rebalancing, the magnitude of benefit was reduced by 13% to 23%. Benefit was apparent mostly in the large vessel occlusion subtype. Conclusion— This study demonstrated the feasibility of rebalancing individual subjects within a randomized trial. After rebalancing and outlier elimination, recombinant tissue plasminogen activator continued to demonstrate improved outcome. That the apparent treatment effect was reduced suggests that imbalances contributed to the magnitude of the original National Institute of Neurological Diseases and Stroke outcomes. This method could in theory be applied to any data set to find matched subjects for outcome or other analyses.

Background and Purpose— Many early phase trials in stroke have not been subsequently confirmed. Randomization balance in baseline factors that influence outcome are difficult to achieve and may be partly responsible for misleading early results. We hypothesized that comparison with an outcome function derived from a large number of pooled control arms would mitigate these randomization problems and provide a reliable predictor for decision-making before proceeding to later phase trials. We developed such a model and added a novel feature of generation of multidimensional 95% prediction surfaces by which individual studies could be compared. We performed a proof-of-principle study with published clinical trials, determining whether our method correctly identified known outcomes. Methods— The control arms from all randomized, controlled trials for acute stroke with ≥10 subjects, including baseline National Institute of Health Stroke Scale, age, and 3-month outcomes published between 1994 and May 2008, were identified. A Matlab program (PPREDICTS) was written to generate outcome functions based on these parameters. Published treatment trials were compared with these 95% intervals to determine whether it successfully identified positive and negative trials. Results— Models of mortality and functional outcome were successfully generated (mortality: R 2 =0.69; functional outcome, modified Rankin Scale 0 to 2: R 2 =0.81; both P Conclusion— This proof-of-principle analysis confirmed all positive and negative clinical stroke trial results and identified some promising therapies. The use of a pooled standard treatment group function combined with statistical bounds may improve selection of early studies for further study. This method may be applicable to any condition in which baseline factors influence outcome and at any stage of the development process.

To the Editor: We read with interest Rha and Saver’s1 review of relationships between recanalization and outcome and commend them on their comprehensive review of pertinent literature. However, we disagree with their main conclusion that recanalization may substitute for measurement of clinical outcome. The central finding that leads Rha and Saver1 to their conclusion is that those patients who recanalized fared better than those that did not. We contend that this conclusion is valid only if a fixed dose of agent was used and the groups were comparable at baseline. In nearly every cohort there will be individuals that do not respond to a specific therapy regardless of dose and so are subject primarily to the toxicity of the intervention as dose is raised …