Emergency Administration of Abciximab for Treatment of Patients With Acute Ischemic Stroke: Results of an International Phase III Trial
Abciximab in Emergency Treatment of Stroke Trial (AbESTT-II)
Background and Purpose— A previous randomized, placebo-controlled, double-blind study suggested that abciximab may be safe and effective in treatment of acute ischemic stroke. The current phase 3 study was planned to test the relative efficacy and safety of abciximab in patients with acute ischemic stroke with planned treatment within 5 hours since symptoms onset.
Methods— An international, randomized, placebo-controlled, double-blind phase 3 trial tested intravenous administration of abciximab in 2 study cohorts using stratification variables of time since onset and stroke severity. The planned enrollment was 1800 patients. The primary cohort enrolled those patients who could be treated within 5 hours of onset of stroke. A companion cohort enrolled patients that were treated 5 to 6 hours after stroke as well as a smaller cohort of patients who could be treated within 3 hours of stroke present on awakening. The primary efficacy measure was the dichotomous modified Rankin Scale score at 3 months as adjusted to the baseline severity of stroke among subjects in the primary cohort. The primary safety outcome was the rate of symptomatic or fatal intracranial hemorrhage that occurred within 5 days of stroke.
Results— The trial was terminated prematurely after 808 patients in all cohorts were enrolled by recommendation of an independent safety and efficacy monitoring board due to an unfavorable benefit-risk profile. At 3 months, approximately 33% of patients assigned placebo (72/218) and 32% of patients assigned abciximab (71/221; P=0.944) in the primary cohort were judged to have a favorable response to treatment. The distributions of outcomes on the modified Rankin Scale were similar between the treated and control groups. Within 5 days of enrollment, ≈5.5% of abciximab-treated and 0.5% of placebo-treated patients in the primary cohort had symptomatic or fatal intracranial hemorrhage (P=0.002). The trial also did not demonstrate an improvement in outcomes with abciximab among patients in the companion and wake-up cohorts. Although the number of patients was small, an increased rate of hemorrhage was noted within 5 days among patients in the wake-up population who received abciximab (13.6% versus 5% for placebo).
Conclusions— This trial did not demonstrate either safety or efficacy of intravenous administration of abciximab for the treatment of patients with acute ischemic stroke regardless of end point or population studied. There was an increased rate of symptomatic or fatal intracranial hemorrhage in the primary and wake-up cohorts.
Inhibition of platelet aggregation is a major component of treatment strategies to prevent ischemic stroke. Platelet activation is a key step in thromboembolism and may be involved in the propagation of an arterial thrombus. Changes in platelet function and interactions with inflammatory cells or endothelial cells may affect perfusion, particularly in the microcirculation.1–4 Abciximab, a chimeric mouse/human monoclonal antibody with high affinity for the platelet glycoprotein IIa/IIIb receptor, is indicated as an adjunct to percutaneous coronary intervention for the prevention of cardiac ischemic complications. It also has been used as an adjunct to endovascular procedures or thrombolysis in an animal model4 and for treatment of patients with ischemic cerebrovascular disease, including patients with acute ischemic stroke.5–12
A dose-escalation study in 74 subjects suggested a low risk of symptomatic intracranial hemorrhage (ICH) when the agent was given up to 24 hours after onset of stroke.13 The results of the study suggested that abciximab given as a bolus dose (0.25 mg/kg) followed by a 12-hour infusion at a rate of 0.125 μg/kg per minute was reasonably safe. In order to obtain additional information about safety and potential efficacy, a larger randomized trial was performed.14 In this study of 400 subjects treated with 6 hours of onset of stroke, symptomatic hemorrhages within 5 days were found in 7 of 195 patients (3.6%) given abciximab and 2 of 199 patients (1.0%) in the control group. Treatment with abciximab was associated with a shift of favorable outcomes at 3 months as measured by the modified Rankin Scale (mRS). The trend of potential efficacy was most prominent among subjects treated within 5 hours of onset of stroke.
In order to confirm the efficacy and safety of abciximab in the early treatment of patients with acute ischemic stroke, the Abciximab in Emergency Stroke Treatment Trial–II (AbESTT-II) was performed. The primary objective was to determine whether early administration (within 5 hours since symptom onset) of abciximab would be associated with a significant increase in favorable outcomes at 3 months using the dichotomous scores obtained on mRS as influenced by the baseline severity of the stroke.15 The secondary objective was to compare the proportion of subjects with neurological recovery at 3 months and all-cause mortality at 3 months in the primary cohort. Another objective was to assess the safety and potential efficacy in a group of subjects treated between 5 and 6 hours after symptom onset. In addition, the study enrolled a cohort of patients whose strokes were first detected on awakening but who could be treated within the next 3 hours to address the potential use of emergency therapy to restore perfusion. This aim was to obtain preliminary information about the potential efficacy of early treatment of this group of patients, and it was stimulated by reports that many patients with stroke present on awakening might respond in a manner similar to those patients who are treated within 3 hours of stroke that occur during waking hours.16,17
AbESTT-II was an international randomized, double-blind, placebo-controlled phase 3 trial that tested the efficacy and safety of intravenously administered abciximab given to patients with acute ischemic stroke. The trial was funded through research grants from Eli Lilly & Co and Centocor R&D Inc. Representatives of the sponsors participated in the design and conduct of the trial, including as ex-officio members of the Executive Committee. An Executive Committee with coprincipal investigators (W.H. and H.P.A.) supervised the trial (see Appendix for membership). Clinical coordinating centers were located in Heidelberg, Germany, and Iowa City, Iowa. The members of the Executive Committee and the representatives of the sponsors were unaware of treatment assignment.
The participating sites, with the number of enrolled patients and principal investigators of the research teams, are listed in the Appendix. Assignment of treatments was performed via an interactive voice response system. Brain-imaging studies were collected on patients and sent to a central CT core laboratory, and the studies were interpreted by independent neuroradiologists at the University of Iowa. A physician at Centocor served as the medical monitor for the trial. All causes of death, suspected ICH, stroke progression, and stroke recurrent were reviewed and adjudicated by an independent Clinical Endpoint Committee in a blinded fashion (see Appendix for membership). Their end points were used in the data analyses. All individual safety reports and interim analyses were shared with an independent external safety and efficacy monitoring committee (SEMC); see Appendix). In order to maintain the blinding of the entire Executive Committee, one member of the Committee and one representative of each of the sponsors not otherwise involved in the study interacted with the external SEMC at the time the decision was made to halt the trial prematurely. The other members of the Executive Committee and the sponsors were not aware of treatment allocation or specific information about the analyses at the time the trial stopped.
The Executive committee (Appendix) is responsible for reporting the scientific data and the writing of this article. The investigators had the opportunity to probe the database and to review the results of the analyses. The sponsors had the opportunity to review this publication before its submission for the purpose of ensuring the accuracy of the data and to determine whether documents needed to be submitted to protect the intellectual property interests in the compound or its uses. The trial was registered with the ClinicalTrials. gov with study number NCT 00073372.
The primary efficacy objective was to compare outcomes, as measured by the proportion of mRS responders at 3 months in the primary population in patients treated with abciximab versus placebo. For patients with baseline National Institutes of Health Stroke Scale (NIHSS) scores of 4 to 7, the goal was a mRS score of 0, for those with a NIHSS score of 8 to 14, the measure of success was a score of 0 or 1, and for patients with a NIHSS score of 15 to 22, the desired mRS score was 0 to 2.15 Assuming the response rate to be 26% to 31% in the placebo group as found in the AbESTT study, the sample size was chosen such that 1200 patients would have a statistical power of 80% to detect an effect size of 10% with a 2-sided significance level of 0.05.14
Secondary efficacy analyses evaluated the proportion of patients with neurological recovery (NIHSS score 0 to 1) or mortality at 3 months; proportions of patients with minimal disability (mRS 0 to 1) at day 5/discharge, 10 days, 6 weeks and 3 months; Barthel Index scores of 95 to 100 at 5 days/discharge, 10 days, 6 weeks and 3 months; Barthel Index scores of 100 at 5 days/discharge, 10 days, 6 weeks, and 3 months. Patients with stroke progression or neurological recovery (NIHSS 0 to 1) at 5 days/discharge and patients with stroke recurrence at 3 months were also analyzed. The distributions of mRS, Barthel Index, and NIHSS scores at 3 months also were assessed. Similar efficacy end points were analyzed among patients treated 5 to 6 hours after stroke and among patients with stroke on awakening.
The primary safety end point was the rate of fatal or nonfatal symptomatic ICH that was diagnosed within 5 days/discharge. Other safety events included other ICH within 3 months, asymptomatic hemorrhages found on CT within 5 days/discharge or at 3 months, true thrombocytopenia within 5 days/discharge and nonintracranial bleeding within 5 days/discharge. The safety analyses were performed independently in the primary and companion populations.
The design of AbESTT-II included recruitment of patients into 2 study cohorts, which were interlinked for baseline assessment, treatment regimen, and follow-up evaluations. However, for the statistical analyses for efficacy and safety, the 2 cohorts were treated as 2 separate studies within the larger trial. This design was reviewed with and approved by regulatory agencies.
The primary cohort included those persons who could be treated within 5 hours of onset of stroke (randomization <4.5 hours). The goal was to recruit 1200 subjects into this cohort. The sample size calculations for the trial were based on the anticipated outcomes in these subjects. The outcomes of the subjects enrolled in the primary cohort were used for the primary safety and efficacy analyses of AbESTT-II, including any decision to prematurely halt the trial.
Because there was interest in testing the safety and efficacy of abciximab among subjects treated at a slightly longer interval and among those with stroke present on awakening, AbESTT-II planned to evaluate safety and efficacy in another 600 subjects who could be enrolled in the companion cohort. The cohort included those patients who could be treated 5 to 6 hours after stroke onset (randomization 4.5 to 5.5 hours) or who could be treated within 3 hours of awakening with signs of stroke (randomization <2.5 hours). Both safety and efficacy outcomes among the subjects in the companion cohort were analyzed independently from the primary cohort. In addition, responses among the subjects in the wake-up group were evaluated independently from the other subjects in the companion cohort.
Eligible patients were assigned to either abciximab or an identically appearing placebo at a 1:1 randomization using minimization procedure with biased-coin assignment.18 The 4 balancing factors used were: study population (primary or companion), site, baseline NIHSS score (4 to 7, 8 to 14, and 15 to 22—three groups) and time since stroke onset.
Patients older than 18 years of age and with an acute ischemic stroke resulting in a baseline NIHSS score ranging from 4 to 22 were recruited at 112 sites (see Appendix). Baseline evaluation included a clinical assessment, non–contrast-enhanced CT of the brain, and laboratory tests. Baseline clinical assessments included a prestroke mRS and Barthel Index. Women of childbearing potential had a pregnancy test and if the test was positive, the woman was excluded. Exclusion criteria for the trial were similar to other acute stroke treatment trials. Those patients that might be treated with intravenous recombinant tissue plasminogen activator, including most seen within 3 hours of onset of stroke, were excluded. Patients with mild (NIHSS score <4) or severe (NIHSS score >22) were excluded. The respective institutions’ human subject research boards had to approve the protocol and consent. Before enrollment, the patient or a surrogate (where permitted) gave signed consent.
Pharmacists at the sites were given identically appearing 5-mL vials of abciximab or placebo. The study medication was administered as an initial bolus dose (0.25 mg/kg) followed by a 12-hour infusion (0.125 μg/kg per minute) up to a maximum infusion of 10 μg/min.
Ancillary Care and Follow-Up Assessments
Patients were not to receive aspirin or other antithrombotic medications until after a follow-up CT study that was performed 36 to 48 hours after randomization. Thereafter, the selection of medications was at the discretion of the treating physician. Mechanical methods for prevention of deep vein thrombosis were allowed at any time during the study. Other interventions, including rehabilitation, were encouraged. The protocol included recommendations for treatment of potential complications including thrombocytopenia or serious bleeding.
Close observation including frequent measurements of vital signs were done during the first 24 hours after completing the infusion of study agent. Blood pressure parameters and recommendations for treatment of hypertension were provided. Investigators performed daily clinical assessments and scoring of the NIHSS through day 5 or discharge, whichever was sooner. Follow-up assessments of the mRS and Barthel Index were done either by telephone or in person at 10 days and 6 weeks. Investigators performed the NIHSS, mRS, and Barthel Index at ≈3 months. Intercurrent events and interventions were recorded at the time of the follow-up studies.
Hematological studies, including platelet counts, were done at regular intervals until day 5 or discharge. A follow-up CT study was done at 36 to 48 hours after randomization. Additional brain-imaging tests during or after the acute treatment period could be ordered at the discretion of the physician. Other diagnostic studies, including vascular or cardiac imaging, were ordered by the treating physician. Copies of scheduled and unscheduled brain-imaging studies were forwarded for central interpretation.
Adverse Experiences and Outcomes
Serious adverse experiences were reported following regulatory requirements. Breaking of the study blind could be performed in the event of an emergency that would require knowledge of the treatment medication. Because of the effects of abciximab on platelet function, we included a time period of 5 days to monitor for acute treatment period adverse experiences, including symptomatic intracranial hemorrhage.
Any deaths, including the presumed cause, which occurred within 120 days of randomization, were recorded. However, all deaths occurred before 90 days. Intracranial hemorrhages were diagnosed by the results of brain-imaging tests. The definitions of symptomatic or asymptomatic ICH used by this trial were described previously.14 Symptomatic intracranial hemorrhage was diagnosed if neurological worsening was detected and hemorrhage was found on brain imaging. A causal link was required. Both major and minor non-neurological bleeding and thrombocytopenia were recorded using the previously described definitions.14 In addition, other serious or nonserious adverse experiences were recorded through day 5 and 3 months. Investigator-reported bleeding was reported through 3 months. Major and minor bleeding events were reported through 5 days. Adverse events were recorded through 3 months. Serious adverse events were also reviewed through 5 days/discharge.
Outcomes were assessed at 5 days and 3 months (90 to 120 days after randomization) using the mRS score at 3 months as the primary measure of efficacy. Alternative methods to obtain follow-up information, such as house calls or assessment by a nonstudy physician were to be used if the patient could not return to the study site. Any assessment performed >120 days after randomization was not included in the analysis. Other outcomes included neurological improvement, neurological deterioration, stroke progression, or recurrent stroke. The definitions of these outcomes have been described in a prior report.14
The trial examined safety and efficacy results independently in the 2 study cohorts. In addition, results in the companion cohort were evaluated separately for those patients treated 5 to 6 hours after onset of stroke and those with stroke detected on awakening. All efficacy analyses were performed on data from all randomized patients using the intention-to-treat principle. All safety analyses were performed on all patients who received at least some study treatment according to the actual treatment received. For the adjusted dichotomous mRS responder analysis, data were inputed for patients with missing observations on the 3-month mRS by carrying the last available postrandomization value. If no postbaseline values were available for a patient or if a patient died before the 3-month assessment, the patient was considered to be a nonresponder for the 3-month mRS responder end point. Wald test statistic from a logistic regression model with baseline NIHSS groups as a covariate was used to compare the 2 treatment groups.
Planned subgroup analyses in each of the cohorts were performed using adjustments for age, sex, baseline NIHSS scores, time from onset of symptoms, presence of diabetes, use of aspirin within 24 hours or 7 days within randomization, or presence of new stroke on baseline CT. In general, Mantel-Haenszel χ2 statistics were used to compare event rates or proportions between treatment groups for dichotomous secondary end points without a stratification variable. For rare events, Fisher exact test was performed. Mortality was calculated using Kaplan–Meier estimates based on those deaths that occurred within 120 days of randomization. The log-rank test was used to compare mortality between the treatment groups.
Periodic monitoring of safety data were performed by the SEMC. No formal statistical testing was done on this interim safety data. Two interim efficacy analyses were planned. These analyses were to be performed when the first 400 and 700 subjects, respectively, in the primary population had been randomized and 3-month end point data were available. However, because this trial was terminated early, no planned interim analyses were performed. Other unscheduled analyses, primarily for safety, could be performed at the discretion of the SEMC or on alert of the trial’s safety monitor, who was unaware of treatment allocation.
The results of AbESTT-II are reported independently for the primary and companion cohorts. In addition, the results among subjects in the companion cohort are analyzed independently of the cohort of subjects with stroke present on awakening. The trial did not perform analyses aggregating the data from the 3 cohorts.
Between December 2003 and September 2005, the trial randomized a total of 808 patients in 112 sites. At an unscheduled interim analysis in the wake-up stroke patients due to a higher than anticipated overall intracranial hemorrhage rate, the SEMC recommended that recruitment of patients into this part of the study be halted effective May 20, 2005, because the rate of symptomatic ICH among patients receiving abciximab exceeded the anticipated safety margins. Subsequently, in September 2005, the SEMC recommended suspending enrollment in the entire trial. After a review of the efficacy data through 6 weeks of follow-up in >90% of enrolled patients and through 3 months of 80% of all patients enrolled, the SEMC recommended that the trial be halted because of safety concerns and a less than acceptable benefit-risk ratio. This recommendation was accepted unanimously by the Executive Committee and the sponsors. Although the trial halted recruitment, other study procedures, including scheduled follow-up assessments were completed on all previously enrolled subjects. The database was not locked and the final analyses were not performed until these data were received and verified.
Of 808 patients randomized, 7 patients were not included in this analysis (5 withdrew consent, 2 have no informed consent), for a total of 801 analyzable patients. During the period of enrollment, 9011 patients were screened. The primary reasons for not being enrolled in the trial are included in Table 1. Many patients had multiple reasons for exclusion.
Four hundred thirty-nine, 319, and 43 patients were enrolled in the primary, companion, and wake-up cohorts, respectively (Figure 1). Ten randomized patients did not receive study agent, most commonly because of an elevated blood pressure that met an exclusion criterion that developed after entry (Figure 1). One patient assigned to receive abciximab was administered placebo as a study agent. Three patients in the companion cohort were not randomized but received study drug and hence are included in the safety analyses. The study blind was broken in 8 instances (abciximab–6/placebo–2). The leading reason was a symptomatic ICH that required a surgical intervention. Seventy-three patients (abciximab–38/placebo–35) were treated with antithrombotic medication, most commonly aspirin, after randomization and before the follow-up CT scan.
The baseline characteristics of the study cohorts are described in Tables 2 and 3⇓. No differences were found in regard to risk factors, prior use of antithrombotic medications, interval from stroke, laboratory findings, results of baseline CT scan, or severity of stroke. The stratification assured that numbers of patients in the 3 subgroups by NIHSS score (4 to 7, 8 to 14, and 15 to 22) were similar. The mean and median NIHSS scores in the abciximab and placebo-treated groups are included in the tables. Approximately 20% of the patients in the primary population were enrolled within 3 hours of onset of stroke.
Figure 2 shows the distribution of outcomes (mRS scores) and the rates of favorable outcomes as defined in the responder analysis in the primary population at 3 months. In addition, the rates of favorable outcomes in subgroups divided by NIHSS scores of 4 to 7, 8 to 14, and 15 to 22 are shown. No treatment effect was noted in the primary cohort between the abciximab and placebo-treated patients at 3 months. Adjustments for baseline NIHSS score, age, diabetes mellitus, prior aspirin use, or baseline CT findings did not affect the results of the study. Other efficacy outcomes are included in Table 4. Stroke progression was diagnosed within the first 5 days in 17 abciximab-treated and 19 placebo-treated patients. Recurrent strokes were diagnosed during the follow-up period in 7 abciximab-treated and 3 placebo-treated patients. No benefit from treatment could be detected in these secondary analyses.
Symptomatic (Fatal or Nonfatal) ICH
In the primary cohort, symptomatic ICH was diagnosed within 5 days in 13 patients (abciximab–12 [5.5%]/placebo–1 [0.5%]; P=0.002; Tables 5 and 6⇓). These hemorrhages were fatal in 8 patients (abciximab–8 [3.7%]/placebo–0 [0%]; P=0.004). Within 3 months, symptomatic or fatal ICH was diagnosed in 2 additional patients in the abciximab group, both of which were fatal (total abciximab–14; P=<0.001; Table 4). Patients randomized within 3 hours of onset of stroke did not have any difference in rates of bleeding or efficacy components compared with those patients treated at slightly longer periods. There was not a relationship of the severity of neurological impairments, the baseline CT findings, or the prior use of medications and the development of symptomatic hemorrhage in the entire primary cohort or those enrolled within 3 hours.
Overall, 60 patients (abciximab–35, placebo–25) died during the follow-up period (Table 7). Figure 3 shows the Kaplan–Meier plot of the mortality through 3 months after randomization of the primary population. The primary causes of death were ICH in 11 patients (abciximab–11, placebo–0), initial stroke in 15 patients (abciximab–6, placebo–9), and recurrent stroke in 3 (abciximab–3, placebo–0). Other causes of death were evenly distributed between the abciximab and placebo patients.
Other Safety Outcomes
The rates of asymptomatic ICH, other serious bleeding and thrombocytopenia are listed in Table 7. Other serious adverse experiences were diagnosed within 5 days in 30 abciximab-treated and 33 placebo-treated patients. Through the 3-month time period, serious adverse experiences were diagnosed in 123 patients (abciximab–62, placebo–61). Most of these other serious adverse events were neurological or neurovascular and related to the initial stroke as well as subsequent respiratory disorders.
Efficacy and Safety in the Companion Cohort Enrolled 4.5 to 5.5 Hours After Onset of Stroke
Figure 4 shows the distribution of outcomes (mRS scores) and the rates of favorable outcomes as defined in the responder analysis in the companion cohort treated 5 to 6 hours after stroke at 3 months. In addition, the rates of favorable outcomes in subgroups divided by NIHSS scores of 4 to 7, 8 to 14, and 15 to 22 are shown. No treatment effect was noted. Adjustment for NIHSS score, age, diabetes mellitus, prior aspirin use, or baseline CT findings did not affect the results of the study. Other efficacy outcomes are included in Table 4. Stroke progression within 5 days was diagnosed in 37 patients (abciximab–16/placebo–21). Recurrent strokes at 3 months occurred in 7 patients (abciximab–4/placebo–3). No benefit from treatment could be detected in these secondary analyses.
Symptomatic (Fatal or Nonfatal) ICH
In this cohort, no significant difference in hemorrhages was noted between the 2 treatment groups, symptomatic or fatal ICH was detected within 5 days in 4 patients (abciximab–3/placebo—1; Tables 5 and 6⇑). These hemorrhages were fatal in 3 patients (abciximab–2/placebo–1). Another 3 hemorrhages were diagnosed during the follow-up period (abciximab–1/placebo–2), both being fatal in placebo.
Overall, 35 patients in the companion cohort died (abciximab–18, placebo–17; Table 7). The primary cause of death was initial stroke occurring in 11 of the 35 patients (abciximab–4/placebo–7). Other causes of death were evenly distributed between the abciximab and placebo patients.
Other Safety Outcomes
The rates of asymptomatic ICH, other serious bleeding, and thrombocytopenia are described in Table 5. Other serious adverse experiences were diagnosed within 5 days in 23 abciximab-treated and 33 placebo-treated patients. Through 3 months, serious adverse experiences were diagnosed in 105 patients (abciximab–48/placebo–57). Most of these other serious adverse events were neurological or neurovascular and related to the initial stroke as well as subsequent respiratory disorders and heart failure.
Efficacy and Safety in the Companion Cohort With Wake-Up Stroke
Figure 5 shows the distribution of outcomes (mRS scores) and the rates of favorable outcomes as defined in the responder analysis in the companion cohort with wake-up stroke at 3 months. In addition, the rates of favorable outcomes in subgroups divided by NIHSS scores of 4 to 7, 8 to 14, and 15 to 22 are shown. No significant difference in treatment effect was noted. The numbers of patients in the 2 treatment groups are relatively small; adjustment for NIHSS score, age, diabetes mellitus, prior aspirin use, or baseline CT findings did not affect the results of the study. Other efficacy outcomes are included in Table 4. Stroke progression within 5 days was diagnosed in 7 patients (abciximab–2/placebo–5). Recurrent strokes at 3 months occurred in 2 patients (abciximab–1/placebo–1). No benefit from treatment could be detected.
Symptomatic (Fatal or Nonfatal) ICH
Symptomatic ICH was diagnosed within 5 days in 4 patients (abciximab–3/placebo–1; Tables 4 and 5⇑). These hemorrhages were fatal in 3 patients (abciximab–2/placebo–1). A subsequent fatal hemorrhage was diagnosed in the follow-up period in one patient treated with abciximab.
Besides having a high rate of bleeding complications, overall this group of patients had a high rate of death, especially among those treated with abciximab. Overall, 9 patients (abciximab–6/placebo–3) died during the follow-up period (Table 7). The primary causes of death in this group were ICH in 4 patients (abciximab–3/placebo–1) and cerebral infarction in 1 patient (abciximab–1/placebo–0).
Other Safety Outcomes
The rates of asymptomatic ICH, other serious bleeding and thrombocytopenia are reported in Table 5. Other serious adverse experiences were diagnosed within 5 days in 3 abciximab-treated and 5 placebo-treated patients. Through 3 months, serious adverse experiences were diagnosed in 13 patients (abciximab–6/placebo–7). Most of these other serious adverse events were due to atrial arrhythmias.
This large international trial planned to enroll 1800 patients (1200 in the primary cohort) to test the hypothesis that abciximab would improve outcomes at 3 months after stroke using a dichotomous end point on the mRS as influenced by the baseline NIHSS score. The current trial, which was a phase 3 trial to test efficacy, was built on the success of the dose escalation pilot study and the confirmatory phase IIb trial that were previously reported.13,14 The trial did not show benefit from treatment with abciximab in any cohort examined although the primary cohort had a significant increase in symptomatic or fatal ICH. Based on the lack of a favorable benefit:risk profile for the study at an interim analysis, and that it was highly unlikely to demonstrate a favorable profile by enrolling more patients, the SEMC concluded and the investigators accepted to discontinue the study.
The previous study suggested that those patients treated within 5 hours of onset of stroke likely would have the greatest benefit from treatment with abciximab.14 This group of patients became the primary cohort, for which the overall trial’s primary safety and efficacy analyses were judged. The concept that early restoration of perfusion is associated with the highest likelihood of improved outcomes is supported by the meta-analyses of the trials of intravenous thrombolysis.19 This trial did not replicate the previous trial. No improvement in the rates of favorable outcomes with early treatment with abciximab was found. Approximately 20% of the patients in this group were treated within 3 hours of onset of stroke. Even including this group of patients treated soon after onset of stroke, an improvement in the rate of favorable outcomes with treatment with abciximab could not be demonstrated. Analysis of the influence on results by baseline severity of stroke showed heterogeneity in the results among the different populations and did not correlate with any evidence of benefit. Although the previous study found abciximab to be most effective in moderate-to-severe stroke (NIHSS scores <15), severity of stroke at baseline did not affect response to treatment in this trial. Both the primary and secondary outcome measures of efficacy did not demonstrate a treatment-effect. Subgroups based on clinical and demographic characteristics also demonstrated a lack of efficacy in the primary outcomes. In summary, this trial did not find that early administration of abciximab was effective in improving outcomes at 3 months.
In order to determine whether patients could be successfully treated at longer intervals (up to 6 hours after stroke), the companion population was included in this study. This group of patients was assigned treatment through an independent randomization process. Because the interval from stroke until treatment would be longer, it was anticipated that the rates of favorable outcomes would be less. Indeed, favorable outcomes dropped from 32.1%–33.0% to 23.3%–25.6%. Although the rate of favorable outcomes was slightly higher with treatment with abciximab, these results were not significantly different. There is no obvious reason why intravenous administration of abciximab would be effective at slightly longer treatment period. This finding probably reflects the relatively small number of patients in this treatment group and did not justify continued recruitment of patients in this group when the results among patients in the primary population were negative.
To date, patients who awaken after several hours of sleep have not been enrolled in clinical trials because the time of onset of stroke has not been known. Serena et al16 observed that some patients in this group might respond favorably to early reperfusion therapy. Therefore, some patients were enrolled as a subgroup of the companion population. Based on CT findings, Todo et al20 concluded that stroke at awakening in many patients may be secondary to events that happen shortly before arousal. They concluded that these patients might be good candidates for acute stroke interventions. Therefore, this trial included an exploratory arm to evaluate the potential efficacy of treatment among patients with wake-up stroke. Although the mean NIHSS score was slightly higher among the patients included in the wake-up stroke group, their outcomes in both treatment groups were worse than those patients whose strokes occurred while awake. The patients in the group receiving abciximab had more complications and overall outcomes were poorer. Because of a concern about the potential safety of abciximab among the patients in the wake-up stroke group, enrollment was suspended. Although the high rate of poor outcomes may be unique to our study, our experience should cause other investigators to approach this group of patients with caution.
Evaluation of the components of the design of the trial did not identify any variable to explain the difference between the results of this trial and our previous study. The treatment regimen, entry criteria, ancillary care, and follow-up assessments were the same. Although the primary outcome measure of efficacy in this trial was changed to the adjusted dichotomous responder analysis for this trial, no benefit from treatment was seen in any other end point chosen. It is unlikely that the lack of benefit from abciximab shown in this trial is the result of trial design but more likely the failure of treatment. Other measures of favorable outcome, including a mRS score of 0 to 1 at 3 months regardless of baseline severity of stroke, neurological recovery, score of 95 to 100 on the Barthel Index also showed no treatment effect. Other statistical methods, such as the shift in distribution analysis, have been used in clinical trials.21 Such an analysis would not have provided different results for AbESTT-II.
Symptomatic ICH is a potential complication of any intervention used to restore or improve perfusion to the brain after acute ischemic stroke. In AbESTT, the risk of symptomatic ICH was ≈4% and none of the cases were fatal.14 In contrast, whereas the overall risk of serious intracranial bleeding in this trial was approximately the same, many of the complications were fatal. In addition, the risk of symptomatic intracranial hemorrhage with treatment with abciximab in AbESTT-II was less than that reported with intravenous administration of thrombolytic agents.22,23 Still, comparisons about the hemorrhagic risks found in AbESTT-II and the trials of thrombolysis should be made with caution. The agents have different effects on thrombosis. Differences among the trials include the severity of baseline strokes, interval from stroke onset until treatment, and the period of observation for symptomatic hemorrhage (24 to 36 hours versus 5 days). In addition, many patients screened for AbESTT-II were treated with thrombolysis and were excluded.
We have not identified factors, including ancillary care or previous use of aspirin, that predicted an increased risk of intracranial bleeding in this study. Although AbESTT-II used the same treatment regimen as the previous studies, there always is a possibility that a lower dose or medication or another treatment regimen could have lowered the risk of intracranial bleeding. Because considerable interest remains about the possible role of abciximab in treating patients with acute cerebral ischemia in a variety of situations, future researchers may wish to look at a lower dose or different treatment regimen. At present, there seems little reason to test intravenous administration of abciximab in the dose and regimen that was evaluated in this study.
Although the number of patients enrolled in the wake-up group is small, randomization into this arm of the study was halted earlier than the remainder of the trial because of the high rate (13.6% as compared with 5% for placebo) of symptomatic ICH among patients treated with abciximab. Given this experience, other investigators should use caution when administering interventions to restore perfusion in patients with wake-up strokes. Possibly, new criteria for eligibility based on clinical or imaging variables are needed before enrolling such patients in future trials.
Although the rates of symptomatic intracranial hemorrhage among the patients enrolled in the primary population and those treated within 5 to 6 hours of stroke are important, these rates of bleeding might have been acceptable if there were improved outcomes from treatment. Thus, the lack of evidence of efficacy combined with the increased potential risks from treatment resulted in this trial’s premature discontinuation of enrollment. The trial suggests that intravenous abciximab does not have a role in the management of patients with acute ischemic stroke. The results of this trial should not be construed to reflect the lack of efficacy or safety of abciximab in other forms of acute vascular disease, including acute myocardial ischemia, or in other clinical settings, such as the use of the agent in conjunction with endovascular interventions.
Executive Committee Members
H.P. Adams, Jr, W. Hacke, A. Davalos, J. Torner, J. Frayne, P. Teal, J. Leclerc, B. Oemar, M.B. Effron.
Steering Committee Members
A. Demchuck, A. Massaro, R. Atkinson, P. Trouillas, M.-H. Mahagne, M. Kaste, A. Grau, G. Ford, J. Frayne.
Iowa: H.P. Adams, E. Leira, P. Davis, K. Grimsman; Heidelberg: W. Hacke, P. Ringleb, P. Schnitzer.
D. Sherman, J. Easton, M. Eliasziw, G. Winslow, Dr Wolf-Dieter Heiss, B. Thornton.
Investigators Who Recruited at Least on Patient (number of patients enrolled in each country included in parenthesis)
Argentina (45): M. Pardal, A. Firstenfeld, B. Maskin, J. Furst. Australia (12): R. Markus, D. Crimmins. Austria (13): F. Fazekas, F. Aichner. Belgium (25): P. Cras, G. Vanhooren, V. Thijs, C. Willems. Brazil (22): F. Yamamoto, R. Gagliardi, A. Massaro. Canada (80): N. Bayer, A. Penn, P. Teal, V. Hachinski, L. Berger, A. Shuaib, A. Demchuk, D. Selchen. Spain (139): E. Diez-Tejedor, R. Fernandes-Bolanos, F. Lopez, A. Chamorro, J. Roquer, J. Castillo, J. Alvarez-Sabin, A. Davalos, Y. Silva. Filand (19): J. Sivenius, M. Kaste. France (62): C. Magne, P. Trouillas, F. Woimant, M. Mahagne, P. Amarenco. Great Britain (2): G. Ford. Germany (108): R. Schneider, D. Nabavi, G. Seidel, R. Toepper, C. Kessler, P. Vogel, M. Jauss, G. Gahn, A. Grau, A. Mueller-Jensen, P. Berlit, W. Boessenecker, M. Schabet, A. Hetzel, H. Audebert, J. Glahn, R. Huber, P. Ringleb. Italy (17): P. Nichelli, I. Santilli, G. Comi. Netherlands (2): Dr van der Worp. Poland (79): J. Pniewski, A. Szczudlik, H. Hertmanowska, W. Nyka, Z. Stelmasiak, W. Drozdowski, W. Fryze, A. Kuczynska-Zardzewialy, A. Czlonkowska. Portugal (29): L. Cunha, G. Lopes, T. Pinho e Melo, A. Salgado. South Africa (31): C. Guldenpfennig, J. Roos. Switzerland (11): J. Bogousslavsky, R. Baumgartner. United States (112) D. Kleindorfer, S. Starkman, J. Yasen, S. Braheny, J. Broderick, E. Jauch, S. Cruz-Flores, M. Schneck, E.M. Hekler, P.L. Kwen, C. Lewandowski, S. Rasmus, E. Leira, S. Kasner, R. Bernstein, L. Anderson, B. Demaerschalk, W. Felton, III, E.C. Haley, Jr, J. Hilburn, K. Ng, K. Levin, T. Devlin, M. Hammer.
J. Maley, M. White, J. Weigele.
K. Sila, N. Pasca.
S. Bhat, M. White, J. Maley, J. Weigele.
B. Oemar, G. Holtkamp, V. Valbert, M. Etzi, J. Kearl, T. Henrick, J.D. Lu, L. Damaraju, E. Barnathan.
J. Croaning, R. Kovacs, J. Leclerc, V. Helm, S. Winings, R. Lockhart, M.B. Effron.
Sources of Funding
Supported by Eli Lilly and Company and Centocor Research and Development (see text describing interactions between the sponsors and the trial).
H.P.A. has a financially significant research grant that is sponsored by Eli Lilly and Company and Centocor Research and Development. M.E. has a significant ownership interest in Eli Lilly and Company, and has been employed by Eli Lilly and Company. A.D., J.F., and P.T. have a financially significant research grant that is sponsored by Eli Lilly and Company and Centocor Research and Development. J.L., L.P., E.B., and W.H. have significant ownership interests and have been employed by Centocor Research and Development. B.O. and W.H. have ownership interests and have been employed by Eli Lilly and Company. J.C.T. has nothing to disclose.
Presented in part at the European Stroke Conference, Brussels, Belgium, May 19, 2006.
- Received October 26, 2006.
- Revision received May 24, 2007.
- Accepted June 14, 2007.
Zhang ZG, Chopp M, Goussev A, Lu D, Morris D, Tsang W, Powers C, Ho KL. Cerebral microvascular obstruction by fibrin is associated with upregulation of PAI-1 acutely after onset of focal embolic ischemia in rats. J Neurosci. 1999; 19: 10898–10907.
Zhang ZG, Zhang L, Tsang W, Goussev A, Powers C, Ho KL, Morris D, Smyth SS, Coller BS, Chopp M. Dynamic platelet accumulation at the site of the occluded middle cerebral artery and in downstream microvessels is associated with loss of microvascular integrity after embolic middle cerebral artery occlusion. Brain Res. 2001; 912: 181–194.
Zhang L, Zhang Z, Zhang C, Zhang RL, Chopp M. Intravenous administration of a GPIIb/IIIa receptor antagonist extends teh therapeutic window of intra-arterial tenecteplase-tissue-tissue plasminogen activator in a rat stroke model. Stroke. 2004; 35: 2890–2895.
Zhang L, Zhang ZG, Zhang R, Morris D, Lu M, Coller BS, Chopp M. Adjuvant treatment with a glycoprotein IIb/IIIa receptor inhibitor increases the therapeutic window for low-dose tissue plasminogen administration in a rat model of embolic stroke. Circulation. 2003; 107: 2837–2843.
Morris DC, Zhang L, Zhang ZG, Lu M, Berens KL, Brown PM, Chopp M. Extension of the therapeutic window for recombinant tissue plasminogen activator with argatroban in a rat model of embolic stroke. Stroke. 2001; 32: 2635–2640.
lee KY, Heo JH, Lee SI, Yoon WB. Rescue treatment with abciximab in acute ischemic stroke. Neurology. 2001; 56: 1585–1587.
Mitsias PD, Lu M, Morris D, Silver B, Lewandowski C, Papamitsakis NI, Khaligi MM, Ebadian HB, Zhao Q, Soltanian-Zadeh H, Daley S, Hearshen D, Ewing JR, Patel SC, Chopp M. Treatment of acute supratentorial ischemic stroke with abciximab is safe and may result in early neurological improvement: A Preliminary Report. Cerebrovasc Dis. 2004; 18: 249–250.
The Abciximab in Ischemic Stroke Investigators. Abciximab in acute ischemic stroke: a randomized, double-blind, placebo-controlled, dose-escalation study. Stroke. 2000; 31: 601–609.
Abciximab Emergent Stroke Treatment Trial (AbESTT) Investigators. Emergency administration of abciximab for treatment of patients with acute ischemic stroke. Results of a randomized phase 2 trial. Stroke. 2005; 36: 880–890.
Fink JN, Kumar S, Horkan C, Linfante I, Selim MH, Caplan LR, Schlaug G. The stroke patient who woke up: clinical and radiological features, including diffusion and perfusion MRI. Stroke. 2002; 33: 988–993.
Hacke W, Donnan G, Fieschi C, Kaste M, von Kummer R, Broderick JP, Brott T, Frankel M, Grotta JC, Haley EC Jr, Kwiatowski T, Levine SR, Lewandowski C, Lu M, Lyden P, Marler JR, Patel S, Tilley BC, Albers G, Bluhmki E, Wilhelm M, Hamilton S; ATLANTIS Trials Investigators; ECASS Trials Investigators; NINDS rt-PA Study Group Investigators. Association of outcome with early stroke treatment: pooled analysis of ATLANTIS, ECASS, and NINDS rt-PA stroke trials. Lancet. 2004; 363: 768–774.
The NINDS t-PA Stroke Study Group. Intracerebral hemorrhage after intravenous t-PA therapy for ischemic stroke. Stroke. 1997; 28: 2109–2118.