Background and objective Acute ischemic stroke is a potentially devastating condition and leading cause of morbidity and mortality, affecting an estimated 800 000 people per year in the USA. The natural history of untreated or unrevascularized large vessel occlusions in acute stroke patients results in mortality rates approaching 30%, with only 25% achieving good neurologic outcomes at 90 days. Recently, data have demonstrated that early endovascular recanalization of large vessel occlusions results in better outcomes than medical therapy alone. However, the majority of patients in these studies were treated with a stent retriever based approach. The purpose of COMPASS is to evaluate whether patients treated with a direct aspiration first pass (ADAPT) approach have non-inferior functional outcomes to those treated with a stent retriever as the firstline (SRFL) approach.
Materials and methods All patients who meet the inclusion and exclusion criteria and consent to participate will be enrolled at participating centers. Treatment will be randomly assigned by a central web based system in a 1:1 manner to treatment with either ADAPT or SRFL thrombectomy. Statistical methodology is prespecified with details available in the statistical analysis plan.
Results The trial recently completed enrollment, and data collection/verification is ongoing. The final results will be made available on completion of enrollment and follow-up.
Conclusions This paper details the design of the COMPASS trial, a randomized, blinded adjudicator, concurrent, controlled trial of patients treated with either ADAPT or SRFL approaches in order to evaluate whether ADAPT results in non-inferior functional outcome.
Trial registration number NCT02466893, Results.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Acute ischemic stroke (AIS) remains a potentially devastating condition and is a leading cause of morbidity and mortality, affecting an estimated 800 000 people per year in the USA alone and costing an estimated $41 billion in 2007.1 The natural history of untreated or unrevascularized large vessel occlusions in acute stroke patients results in mortality rates approaching 30%, with only 25–40% of patients achieving good neurologic outcomes at 90 days.1 2
There have been numerous trials reported over the past several years indicating the benefit of intra-arterial therapy to revascularize patients suffering from cerebral large vessel occlusions over best medical therapy (Multicenter Randomized Clinical trial of Endovascular treatment for Acute ischemic stroke in the Netherlands (MR CLEAN), Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke (ESCAPE), Endovascular Revascularization With Solitaire Device Versus Best Medical Therapy in Anterior Circulation Stroke Within 8 Hours (REVASCAT), Extending the Time for Thrombolysis in Emergency Neurological Deficits-Intra-Arterial (EXTEND-IA), The Randomized, Concurrent Controlled Trial to Assess the Penumbra System’s Safety and Effectiveness in the Treatment of Acute Stroke (THERAPY), Solitaire With the Intention For Thrombectomy as Primary Endovascular Treatment (SWIFT PRIME), Trial and Cost Effectiveness Evaluation of Intra-arterial Thrombectomy in Acute Ischemic Stroke (THRACE)).3–5 However, across these trials, stent retrievers were the primary approach to effect thrombectomy.
A direct aspiration first pass technique (ADAPT) is an approach that utilizes the advantages of large bore aspiration catheters that can be tracked into the cerebral circulation to directly remove thrombus with aspiration.6 Clot removal using this technique occurs in one of two ways. Either the clot can be entirely aspirated through the catheter, or the thrombus gets lodged in the catheter tip on application of suction and the clot is removed along with the catheter out of the body. In cases where the application of aspiration alone is not successful in removing the thrombus, with the above mentioned two techniques, then the large catheter provides a conduit for the use of a stent retriever at the location of the thrombus to affect thrombectomy. Initial experience with this approach has shown promising results.6 7 Recent data from the 3D separator trial has provided some support that aspiration is potentially comparable with stent retrievers, as they found no significant difference in revascularization or outcomes between the two groups.8 However, it should be noted that the 3D separator is a novel stent retriever and therefore aspiration was not compared against the established Food and Drug Administration (FDA) approved stent retrievers for stroke treatment. The technical comparability of these two approaches were also suggested, but not proven, with the recent publication of the Contact Aspiration versus Stent Retriever for Successful Revascularization (ASTER) trial, a prospective randomized trial designed to compare the angiographic outcomes between the ADAPT and stent retriever as a firstline (SRFL) approaches.9
ADAPT, while encompassing the advantages of aspiration with the rescue opportunity of stent retrievers in select cases, has not been rigorously tested directly against the approach of SRFL therapy. Despite recent data from the 3D trial, single arm data, and the results of the ASTER trial supporting the potential benefit of ADAPT, there is no high quality randomized trial designed and executed to evaluate whether aspiration based thrombectomy demonstrates equivalent outcomes to an SRFL approach.6 7 10–12 The purpose of this study is to demonstrate, in a prospective, randomized, blinded adjudicator, concurrent, controlled trial that patients treated with the ADAPT approach have non-inferior functional outcomes compared with those treated with an SRFL approach.
Methods and design
This is a prospective, randomized, blinded adjudicator, concurrent, controlled trial comparing mechanical thrombectomy with the ADAPT approach and an SRFL approach in patients presenting with AIS from large vessel occlusion within 6 hours of symptom onset. Any FDA cleared mechanical stent retriever (Solitaire, Medtronic, and Trevo, Styker, at the time of the trial) or cleared aspiration catheter device (Penumbra System, at the time of the trial) is approved for use. As a quality initiative, prior to releasing any sites to enroll patients, all centers’ most recent 20 AIS cases will be reviewed to ensure practitioner quality and adequacy of follow-up. The center must also demonstrate at least five patients treated with both ADAPT and SRFL approaches.
Patients who meet the inclusion and exclusion criteria, consent to participate, and who are randomized will be considered enrolled. The treatment arm will be randomly assigned by a central web based system in a 1:1 manner to treatment with either ADAPT or SRFL thrombectomy. Patients randomized to either arm will be treated with aspiration alone or a stent retriever, as randomly assigned to the ADAPT or SRFL cohort, respectively, for up to three passes and thereafter, in cases of persistent occlusion, allowed to employ any therapy according to physician preference. Patients randomized to the SRFL approach can use adjunctive technology during the initial stent retriever passes according to the technique the treating physician uses as their standard of care (regarding the use of a balloon guide catheter, concurrent distal aspiration, both adjunctive tools, or no adjunctive tools). Data on each patient will be collected at the time of enrollment and treatment, and at subsequent follow-up visits.
All sites will keep a screen failure log of all acute stroke patients treated with intra-arterial therapy but who are not randomized into the study. Reason(s) for exclusion will be recorded. Logs will be data entered by the clinical sites on a monthly basis. Recruitment rates will be tracked over time for each hospital. The actual recruitment rates, as well as potential recruitment rates, will be useful for planning further clinical trials and determining the widespread impact of the therapy.
Age 18 years or older (ie, candidates must have had their 18th birthday).
National Institutes of Health Stroke Scale (NIHSS) score ≥5 at the time of neuroimaging.
Presenting or persistent symptoms within 6 hours of when groin puncture can be obtained.
Neuroimaging demonstrates large vessel proximal occlusion (distal internal carotid artery through middle cerebral artery (MCA) bifurcation).
The operator feels that the stroke can be appropriately treated with traditional endovascular approaches (the ADAPT approach or conventional firstline stent retriever approach).
Pre-event nodified Rankin Scale score 0–1.
Non-contrast CT/CT angiography (CTA) or MRI/MR angiography (MRA) for trial eligibility performed or repeated at treating stroke center.
Consenting requirements met according to local institutional review board.
Patient is more than 6 hours from symptom onset.
Rapidly improving neurologic examination.
Absence of large vessel occlusion on non-invasive imaging.
Presence of an existing or pre-existing large territory infarction.
Known or suspected pre-existing (chronic) large vessel occlusion in the symptomatic territory.
Absent femoral pulses.
Excessive vascular access tortuosity that will likely result in unstable access platform.
Pregnancy; if a woman is of childbearing potential a urine or serum beta HCG test is positive.
Severe contrast allergy or absolute contraindication to iodinated contrast.
Clinical history, past, imaging or clinical judgment suggests that the intracranial occlusion is chronic.
Patient has severe or fatal comorbidities that will likely prevent improvement or follow-up or that will render the procedure unlikely to benefit the patient.
Head CT or MRI scan exclusion criteria
Presence of blood on imaging (subarachnoid hemorrhage, intracerebral hemorrhage, etc).
High density lesion consistent with hemorrhage of any degree.
Significant mass effect with midline shift.
Core infarct lesion volume >50 mL.
Angiographic evidence of carotid dissection or tandem cervical occlusion, or stenosis requiring treatment.
Large (more than one-third of the MCA) regions of clear hypodensity on the baseline CT scan or Alberta Stroke Program Early CT score (ASPECTS) <7; sulcal effacement and/or loss of gray–white differentiation alone are not contraindications for treatment.
AIS patients will undergo mechanical thrombectomy within 6 hours of symptom onset or time last seen normal. Appropriate patients will be randomized to treatment by the ADAPT approach or the SRFL approach.
A stratified randomization will take place centrally within REDCap (Research Electronic Data Capture), the database system that will be used for this study. Randomization occurs in a 1:1 ratio to either the ADAPT or SRFL approach. Once a patient is determined to meet all study eligibility criteria, immediately prior to the procedure, the Investigator (or an authorized team member) will log on to REDCap to randomize the subject. The investigator will not be blinded to treatment assignment. Once a patient is randomized, the patient is considered enrolled in the study and must be followed through to the end of the study.
The subject should be clinically evaluated in the same manner as any routine AIS patient. Clinical assessment documenting NIHSS and significant past medical history should be obtained. Imaging with CT or MR per the institution standard of care is required to exclude acute intracranial hemorrhage. Additional anatomic and/or physiologic imaging with CT or MR perfusion imaging per the institution standard of care should then be performed on patients that do not have evidence of significant ischemia on initial scans. Anatomic imaging can use CTA with contrast bolus imaging to visualize the vessels of the head as well as the presence of collateral circulation. Similar anatomic cerebral imaging can be performed with MRA. The studies must demonstrate an acute major vessel intracranial anterior circulation occlusion (internal carotid artery or MCA).
Patient selection will be based on the institutional standard of care. There must be at least 50% volume tissue at risk or an identified eloquent region at risk on CT perfusion for those centers using perfusion selection. The presence of more than one-third territory MCA infarction, ASPECTS <7, or acute infarction >50 mL in volume is considered a boundary for exclusion. Acute volumes will be estimated at the site using A×B×C/2 calculations.
A schedule of events is depicted in table 1.
Primary efficacy endpoint
The primary objective is to evaluate whether AIS patients, with appropriate image selection, treated with the ADAPT mechanical thrombectomy approach within 6 hours of symptom onset do not have inferior clinical outcomes to those treated with SRFL, as evaluated by 90 day global disability, assessed using the modified Rankin Scale (mRS) score, analyzed using success criteria as mRS 0–2. The primary efficacy endpoint, and all subsequent analyses, and the statistical methodology is presented in greater detail in the statistical analysis plan (see online supplementary appendix).
Supplementary file 1
Secondary efficacy endpoints
The secondary objective is to demonstrate that the ADAPT approach is technically superior to the SRFL approach in the treatment of AIS through the evaluation of: time from groin puncture to Thrombolysis in Cerebral Infarction (TICI) 2b or better revascularization, rate of TICI 2c or greater revascularization within 45 min of access, rate of TICI 3 revascularization within 45 min of access, and 90 day disability, assessed across the overall distribution of mRS.
Secondary efficacy outcomes
Secondary efficacy outcomes will evaluate:
TICI 2b or greater revascularization within 45 min of access.
Occurrence of emboli to a new territory.
Presence of vasospasm involving the accessed vascular tree.
90 day global disability assessed via the overall distribution of the utility weighted mRS.
Reduction in stroke severity (NIHSS) at 24 hours post treatment.
Reduction in stroke severity (NIHSS) at 7 days post treatment or discharge (whichever occurs first).
Stroke Impact Score.13
First pass TICI 2b or greater efficacy.
Safety outcomes will evaluate:
Symptomatic intracranial hemorrhage at 24 hours post-randomization.
Asymptomatic intracranial hemorrhage at 24 hours post-randomization.
Intracranial hemorrhage within 90 days of randomization.
All intracranial hemorrhage with distinction of PH2 hemorrhage (neurological deterioration
(NIHSS worsening, score of 4 or more)) within 36 hours of randomization.
Clinically significant complications (pneumonia, sepsis, urinary tract infection, etc) at the time of discharge or 7 days post-randomization (whichever comes first).
Mortality rates at 30 days post-randomization.
Mortality rates at 90 days post-randomization.
Treatment related serious adverse events up to 48 hours post-randomization.
Procedure related serious adverse events.
Cost outcomes will evaluate:
Device related costs for the procedure.
Data safety monitoring board
A data safety monitoring board (DSMB) will comprise three members not participating in the trial and will include a neuroradiologist, neurologist, and neurosurgeon. Additionally, a statistician will be available to facilitate DSMB review. The DSMB will exercise review of the overall safety of the trial, periodically review all adverse events occurring in the trial, and make recommendations for adjustments in the study protocol, should any be considered necessary for safety or other related reasons.
For this study, the sample size was determined by assuming that the true proportions of subjects with mRS outcomes of 0–2 at the 90 day follow-up visit (referred to as a success) are similar to the rate of 32.6% (76/233) from recent data from the MR CLEAN trial. The sample size calculations assumed that 33% of ADAPT patients experience success (mRS 0–2) and 33% of SRFL patients experience success. Based on a one-sided, normal approximation test for non-inferiority with a non-inferiority margin of 15% and alpha=0.05, 122 patients per treatment group will have 80% power. The sample size was adjusted to 135 patients per treatment group to account for up to 10% attrition. Assuming that the observed success rate for both the ADAPT arm and the SRFL arm is 33% (40/122), the 90% normal approximation CI for the true difference in percentages between treatments is (−9.9%, 9.9%). Sample size was computed using SAS V.9.4.
COMPASS is a multicenter, randomized, phase III investigator initiated and run clinical trial evaluating the clinical outcomes of AIS patients treated with the ADAPT mechanical thrombectomy approach within 6 hours of symptom onset compared with those treated with an SRFL approach, as measured by 90 day functional outcomes using the mRS. The primary hypothesis to be tested is that treatment with the ADAPT thrombectomy approach will not have inferior clinical outcomes at 90 days compared with patients treated with the SRFL approach. Each eligible subject will be randomized in a 1:1 ratio to either the ADAPT approach or the SRFL approach. Randomization will take place centrally through the web based REDCap database system. The centrally controlled randomization will help ensure the treatment balance at any interim analysis as well as in the final analysis.
This is a multi-device clinical study where various FDA approved devices may be used to treat vessel occlusion (see table 2). The study is conducted with standardization of subject enrollment, data entry, and adverse event reporting. All investigational sites will follow the requirements of a common protocol, data collection procedures, and forms, and will use cleared mechanical thrombectomy devices in accordance with product labeling.
The primary effectiveness endpoint is mRS success, defined as an mRS score of 0–2 at the 90 day follow-up visit. The number and proportion of patients in each treatment group who are a success will be calculated and the 90% CI for the difference in proportions between treatment groups will be presented. The null and alternative hypotheses for this endpoint are as follows:
H0: pc – pt ≥0.15
H1: pc – pt <0.15
where pt and pc represent the true success proportions for treatment (ADAPT) and control (SRFL), respectively. This endpoint will be analyzed using a logistic regression model with the following terms in the model: treatment, ASPECTS score at baseline, patient age, sidedness, and any other baseline characteristic for which there is a statistically significant difference between treatments. The P value for the test of the null hypothesis will be obtained based on a statistic for the difference (control−treatment) in least squares means (actually proportions rather than means) for each treatment minus the non-inferiority margin of 0.15, all divided by the standard errors of the difference in the least squares means and assuming an approximate normal distribution.
Subjects deceased during study follow-up will be scored as mRS 6, as per the standard scoring for mRS. For subjects missing data for the primary efficacy endpoint, missing values will be imputed using the last observation carried forward (LOCF) method—that is, using the mRS value as of the last available follow-up visit or discharge (whichever is later). A sensitivity analysis will be performed for the intention to treat population in which subjects with missing data for the primary efficacy endpoint will be scored as mRS 6 (ie, failure). A second sensitivity analysis will be performed in which subjects with missing data for the primary efficacy endpoint will be excluded. Under the intention to treat principle, all patients who are randomized are included in the analysis. Additional details can be found in the statistical analysis plan.
Formal hypothesis tests will be performed for the secondary efficacy endpoints only if the null hypothesis is rejected for the primary efficacy endpoint. Holm’s stepwise testing procedure will be used to adjust the significance levels of the tests in the analysis of the secondary efficacy endpoints in order to take multiplicity into account. Holm’s procedure is conducted as follows. The ordered P values p(1), p(2), p(3), and p(4) will be obtained, where the ordering is from least to greatest. The testing procedure begins with the null hypothesis associated with the most significant P value (ie, with H(1)). This hypothesis is rejected if p(1) ≤0.05/4. If H(1) is not rejected, stop the hypothesis testing. Otherwise, proceed to test H(2), which will be rejected if p(2) ≤0.05/3. If H(2) is not rejected, stop the hypothesis testing. Otherwise, proceed to test H(3), which will be rejected if p(3) ≤0.05/2. If H(3) is not rejected, stop the hypothesis testing. Otherwise, proceed to test H(4), which will be rejected if p(4) ≤0.05/1. If the null hypothesis is rejected for a given secondary efficacy endpoint, P values for subsequent endpoints will be presented for descriptive purposes only. Additional details for subgroup and secondary analyses are available in the statistical analysis plan.
Study organization and funding
The trial was funded in May 2015 with sponsorship from Penumbra. Enrollment began in June 2015 and patient enrollment was completed in July 2017. The last patient follow-up was completed in October 2017. Core laboratory review, site verification of data, and data coordinating center data integrity confirmation is ongoing. The clinical coordination for the trial is being performed at the Medical University of South Carolina, Charleston, South Carolina, USA. The statistical and data coordination for the trial is being conducted at Mount Sinai Hospital New York, New York, USA.
Neurothrombectomy has been shown to be safe and effective for the revascularization of occluded cerebral vessels in multiple randomized trials, generating level 1 evidence for the performance of neurothrombectomy. However, the majority of patients randomized to thrombectomy in recent trials were treated with an SRFL approach. COMPASS represents the first prospective, randomized, blinded adjudicator, concurrent, controlled trial designed to evaluate a clinical outcome assessment comparing the ADAPT approach with the SRFL approach in AIS patients treated within 6 hours of symptom onset.
The authors wish to thank COMPASS Invesitgators: Dave Fiorella, Ricardo Hanel, Keith Woodward, DonFrei, Josser Delgado, Adam Arthur, Ansaar Rai, Joey English, Mike Kelly, ItaloLinfante, Jovin Tudor, Raul Nogueira, Bobby Starke.
Contributors Each author listed should receive authorship credit based on the material contribution to this article, their revision of this article, and their final approval of this article for submission to this journal. Alyssa Pierce assisted with the editing and revision of this manuscript.
Funding This research is supported by Penumbra Inc.
Competing interests AST: Codman, consulting, research grants;Penumbra, consulting, research grants; Microvention, consulting, researchgrants; Blockade, stock, consulting; Pulsar Vascular, stock, consulting, research grants; Medtronic, consulting, research grants; and Siemens, consulting.Investor: Cerebrotech, Endostream, Apama, The Stroke Project, Serenity, 3RiversMedical, and Synchron. Consultant: Cerebrotech, Endostream, The Stroke Project, Serenity, 3Rivers Medical, and Vastrax. JM: Investor: Rebound Medical, Viseon, Cerebrotech, Endostream, Apama, The Stroke Project, Comet Medical, Serenity,3Rivers Medical, and Synchron. Consultant: Rebound Medical, Viseon, Cerebrotech, Endostream, The Stroke Project, Serenity, 3Rivers Medical, and Synchron.AHS: Financial interests: StimSox, Valor Medical, Neuro Technology Investors, Cardinal, Medina Medical Systems, Buffalo Technology Partners Inc, and InternationalMedical Distribution Partners. Consultant: Codman, Medtronic, GuidePoint GlobalConsulting, Penumbra, Stryker, MicroVention, WL Gore and Associates, 3RiversMedical, Corindus, Amnis Therapeutics, CereVasc, Pulsar Vascular, The StrokeProject, Cerebrotech Medical Systems, Rapid Medical, Lazarus (acquired byMedtronic), Medina Medical (acquired by Medtronic), Reverse Medical (acquiredby Medtronic), Covidien (acquired by Medtronic), Neuravi, Silk Road Medical, andRebound Medical. PI/National Steering Committees: Penumbra, 3D Separator Trial, Covidien (now Medtronic), SWIFT PRIME and SWIFT DIRECT trials, MicroVention, FRED Trial, MicroVention, CONFIDENCE Study, LARGE Trial, POSITIVE Trial, Penumbra, COMPASS Trial Penumbra, and INVEST Trial. Board member: IntersocietalAccreditation Committee.
Ethics approval The study was approved by the institutional review board of the Medical University of South Carolina.
Provenance and peer review Not commissioned; externally peer reviewed.