Design and early progress of the Comparison of Anticoagulation and anti-Platelet Therapies for Intracranial Vascular Atherostenosis (CAPTIVA) trial

Abstract

Background:

The usual antithrombotic treatment for symptomatic intracranial atherosclerotic stenosis (ICAS) consists of dual treatment with clopidogrel and aspirin for 90 days followed by aspirin alone but the risk of recurrent stroke remains high up to 12 months. The Comparison of Anticoagulation and anti-Platelet Therapies for Intracranial Vascular Atherostenosis (CAPTIVA) trial was designed to determine whether other combinations of dual antithrombotic therapy are superior to clopidogrel and aspirin.

Methods:

CAPTIVA is an ongoing, prospective, double-blinded, three-arm clinical trial at over 100 sites in the United States and Canada that will randomize 1683 high-risk subjects with a symptomatic infarct attributed to 70–99% stenosis of a major intracranial artery to 12 months of treatment with (1) ticagrelor (180 mg loading dose, then 90 mg twice daily), (2) low-dose rivaroxaban (2.5 mg twice daily), or (3) clopidogrel (600 mg loading dose, then 75 mg daily). All subjects receive aspirin (81 mg daily), intensive risk factor management, and will undergo blinded CYP2C19 genotype analysis. The primary goal of the trial is to determine whether rivaroxaban or ticagrelor or both are superior to clopidogrel for lowering the primary endpoint (ischemic stroke, intracerebral hemorrhage (ICH), or vascular death) within 12 months. A prespecified interim safety analysis will be conducted when the first 450 randomized subjects have been followed for 12 months to evaluate the risk of major hemorrhage in the rivaroxaban and ticagrelor arms.

Results:

Enrollment began in August 2022 and, as of 26 June 2024, the 450th subject was randomized into the study.

Conclusion:

CAPTIVA is evaluating two alternative dual antithrombotic therapies to clopidogrel and aspirin to maximize the chance of establishing more effective antithrombotic therapy for symptomatic ICAS, one of the most common and high-risk cerebrovascular diseases worldwide.

Introduction

Intracranial atherosclerotic stenosis (ICAS) is one of the most common causes of stroke worldwide.^1,2 The Stenting and Aggressive Medical Management for Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) trial showed that combining clopidogrel and aspirin with intensive risk factor management was superior to stenting for participants with 70–99% ICAS.^3,4 Nevertheless, the 1-year rate of ischemic stroke (using the American Heart Association (AHA) definition⁵), intracranial hemorrhage (ICH), or vascular death was still 27% in participants in the medical arm of SAMMPRIS who qualified with a symptomatic infarct.^6,7

Several combinations of dual antithrombotic therapy are now used for preventing stroke and myocardial infarction in clinical practice (clopidogrel plus aspirin, ticagrelor plus aspirin, low-dose rivaroxaban plus aspirin), but these dual therapies have not been compared in participants with symptomatic ICAS. This article describes the design and early progress of the Comparison of Anticoagulation and anti-Platelet Therapies for Intracranial Vascular Atherostenosis (CAPTIVA) trial that is evaluating these dual therapies in high-risk participants with symptomatic ICAS.

Methods

Overall design

CAPTIVA is an investigator-initiated, randomized, double-blinded, three-arm, Phase III clinical trial funded by the National Institute of Neurological Disorders and Stroke (NINDS) that is being conducted at more than 100 sites in the United States and Canada (ClinicalTrials.gov, NCT05907629). Participants with a symptomatic infarct within the previous 30 days that is attributed to 70–99% stenosis of an eligible intracranial artery are randomized to 12 months treatment with (1) ticagrelor (180 mg loading dose, then 90 mg twice daily) and aspirin 81 mg per day, or (2) low-dose rivaroxaban (2.5 mg twice daily) and aspirin 81 mg per day, or (3) clopidogrel (600 mg loading dose, then 75 mg daily) and aspirin 81 mg per day. All participants undergo blinded CYP2C19 genotype analysis, receive intensive risk factor management, and are followed for 12 months.

The primary goal of CAPTIVA is to determine if rivaroxaban plus aspirin or ticagrelor plus aspirin or both are superior to the clopidogrel plus aspirin for lowering the 12-month rate of the primary efficacy endpoint (ischemic stroke, ICH, or vascular death). A prespecified safety analysis will be conducted after the first 450 subjects have been followed for 12 months to evaluate whether there is an excess risk of intracerebral hemorrhage (ICH) and non-ICH major hemorrhage in the ticagrelor and rivaroxaban arms. Excess risk will trigger an early futility analysis to evaluate whether the trial should continue to randomize participants to those arms (see Figure 1).

Figure 1.

Trial flow diagram. When the first 450 randomized subjects have completed 12 months of follow-up, the safety analysis will determine whether either experimental arm (ticagrelor + aspirin or low-dose rivaroxaban + aspirin) demonstrates an excess risk of ICH and non-ICH major hemorrhage, using the historical rates from the SAMMPRIS trial as the comparator. If an experimental arm crosses the safety boundary, this will initiate an early futility analysis to weigh the risk: benefit ratio of the corresponding experimental arm. Otherwise, the futility analysis will be conducted after the first 842 subjects have been randomized, by comparing each experimental arm to the standard of care arm. If an experimental arm crosses the futility boundary, the corresponding experimental arm will be terminated. If neither arm crosses the futility boundary, the trial will continue with both experimental arms.

An exploratory aim is to estimate the impact of CYP2C19 loss-of-function (LoF) carrier status on any benefit that the ticagrelor or low-dose rivaroxaban arms may have in lowering the primary endpoint compared with the clopidogrel arm.

The rationale for key design decisions in CAPTIVA (choice of treatment arms, duration of dual antithrombotic therapy and follow-up, inclusion of CYP2C19 LOF carriers) is provided in Supplementary Materials.

Patient population, consent, and randomization

Subjects enrolled in this trial are inpatients or outpatients with a stroke (using the AHA / American Stroke Association (ASA) definition⁵) that occurred within the previous 30 days and is attributed to 70–99% stenosis (or flow gap on magnetic resonance angiography (MRA)) of an eligible intracranial artery. The full inclusion and exclusion criteria for CAPTIVA are provided in Supplementary Materials. The stenosis can be identified by computed tomography angiography (CTA), MRA, or catheter angiography. Enrollment is based on the site investigators’ measurements of 70–99% stenosis using the Warfarin–Aspirin Symptomatic Intracranial Disease (WASID) criteria,^8,9 but all cerebral vascular and brain imaging is subsequently reviewed by a central neuroradiologist (K.R.P) to enable monitoring of the consistency of site and central readings. The central readings are considered the final readings for all planned analyses. All participants are required to give personal written informed consent to participate. Randomization is conducted centrally within the WebDCU^™ system and is designed to balance the treatment distribution within each clinical site and overall in the trial.

Enrollment procedures

Study medications.

For blinding purposes, the clopidogrel, ticagrelor, and rivaroxaban tablets are encapsulated (see details of encapsulation in Supplementary Materials). Participants who have been on clopidogrel or ticagrelor for at least five consecutive days prior to randomization or received a ≽ 300 mg loading dose of clopidogrel or a ≽ 180 mg loading dose of ticagrelor within 5 days before randomization followed by daily clopidogrel or ticagrelor do not receive the loading dose of study antithrombotic medications. The loading dose consists of eight encapsulated tablets: those randomized to clopidogrel take eight active tablets, (ticagrelor: two active tablets and six placebo tablets), [rivaroxaban: one active tablet and seven placebo tablets]. During the entire 12-month follow-up period, subjects take one encapsulated medication in the morning and one in the evening (participants randomized to clopidogrel are taking a blinded placebo in the evening).

Intensive risk factor management.

CAPTIVA uses the same risk factor management protocols as used in SAMMPRIS.¹⁰ Intensive risk factor management will primarily target systolic blood pressure (SBP) < 140 mmHg (the AHA recommended target for patients with symptomatic ICAS¹¹) and low-density lipoprotein (LDL) < 70 mg/dL.¹² In addition, a lifestyle health coaching program (INTERVENT) helps manage other risk factors: diabetes (targeting a hemoglobin A1c of < 7%), smoking, sedentary lifestyle, and high body mass index.^13–16 A statin and one medication from each class of antihypertensive agents are available at no cost to study participants through a national retail pharmacy chain.

CYP2C19 testing.

After enrollment, a mouthwash or buccal swab sample is collected and shipped to the University of Florida for CYP2C19 genotype testing. Genomic DNA is isolated and genotyped for the LoF *2 (c.681G > A; rs4244285) and *3 (c.636G > A; rs4986893) alleles and the increased function (gain-of-function) *17 (c.−806 C > T; rs12248560) allele.¹⁷ The results of this testing are kept blinded from site investigators and participants until the end of the trial.

Subject follow-up

All participants are seen by the site neurologist and the coordinator at 1, 4, 8, and 12 months; when their blood pressure is checked, risk factors are optimized, pill counts of the study antithrombotic medications are obtained, and adverse events are documented. For participants whose SBP is ≽ 140/90 mmHg, an adjustment in the blood pressure medications is made and the subject’s blood pressure level is checked again 30 days later.

Primary and secondary endpoints

The primary safety endpoints of CAPTIVA are (1) parenchymal intracerebral brain hemorrhage (ICH) and (2) major non-ICH hemorrhage according to the International Society on Thrombosis and Hemostasis (ISTH) criteria.¹⁸

The primary efficacy endpoint is a composite of ischemic stroke, ICH, and vascular death at 12 months. The secondary efficacy endpoints are (1) Composite of ischemic stroke, ICH, myocardial infarction (fourth universal definition¹⁹) or vascular death; (2) All stroke (ischemic and ICH); (3) Ischemic stroke; (4) Ischemic stroke in the territory of the qualifying stenotic artery; and (5) All death. An exploratory cognitive outcome at 12 months is measured by the Montreal Cognitive Assessment (MoCA).²⁰ Definitions of all study endpoints are provided in the Supplementary Materials.

Adjudication of potential endpoints

All reported neurological events, bleeding episodes, and cardiac events are adjudicated centrally by two experienced blinded adjudicators (neurologists for neurological and bleeding events, cardiologists for cardiac events), and a third adjudicator if there is disagreement among the first two adjudicators.

External oversight

The study protocol was approved by a central institutional review board (Advarra^®), and the Data and Safety Monitoring Board (DSMB) was appointed by the National Institute of Neurological Disorders and Stroke (NINDS). The US Food and Drug Administration (FDA) issued an investigational new drug exemption (IND) to carry out the study.

Further details of study coordination and oversight and site monitoring are provided in Supplementary Materials.

Statistical considerations

Evaluation of primary safety outcome.

The safety analysis is prespecified to occur when the 450th randomized subject completes 12 months of follow-up. In SAMMPRIS, the rates per 100 patient-years were 0.5 (exact 80% CI 0.05–1.9) for ICH and 2.4 (exact 80% CI 1.2–4.5) for non-ICH major hemorrhage on clopidogrel plus aspirin. The analysis will evaluate whether there is an excess risk of ICH and non-ICH major hemorrhage in the rivaroxaban and ticagrelor arms, compared with the upper 80% confidence limits derived from SAMMPRIS. If an experimental arm (either ticagrelor + aspirin or low-dose rivaroxaban + aspirin) crosses this safety boundary, this will initiate a futility analysis to weigh the risk: benefit of the corresponding arm (see Figure 1).

Sample size determination and statistical analysis for the primary efficacy outcome.

The sample size was calculated based on a two-sample survival analysis comparing the time-to-event of the rivaroxaban and ticagrelor arms to the clopidogrel arm in EAST v6 (EAST Trial Design Software, Cytel, Waltham, MA). The assumed primary event rate in the clopidogrel arm is based on the participants in the medical arm of SAMMPRIS who presented with a symptomatic infarct. Their 12-month rate of ischemic stroke, ICH, or vascular death was 27%.^6,7 Using clopidogrel for 12 (CAPTIVA) rather than 3 months (SAMMPRIS) is likely to lower the rate so we have assumed a 24% primary event rate in the clopidogrel arm. We are seeking a hazard ratio (HR) of 0.66 as the reduction required to overcome the anticipated higher risk of bleeding in the rivaroxaban and ticagrelor arms and the higher cost of these medications (see Supplementary Material for Justification for HR used in the Sample Size Analysis).

To detect an HR of 0.66 with 80% power, using a two-sided 0.05 level of significance, the required sample size per comparison is 948 (474 per arm). This calculation includes inflation for an interim futility analysis and to account for approximately 5% dropout. An additional inflation factor of ~1.18 is applied to account for approximately 8% non-compliance with assigned antithrombotic therapy, yielding a sample size per comparison of 1122 participants (561 per arm). The inflation factor is derived via 1/(1−R²), where R represents the non-compliance rate.²¹ Because the same clopidogrel arm is used as the control in both comparisons, an additional 561 participants are required. This results in a total sample size of 1683 participants. Since the comparison of the ticagrelor and rivaroxaban arms to the clopidogrel arm can be interpreted independently of the other, a multiplicity adjustment is not required.^22–24

An interim futility analysis conducted according to the beta-spending approach using O’Brien-Fleming-like Type II error stopping boundaries will be performed after 50% of subjects have been randomized. An interim efficacy analysis is not planned because stopping a trial early for benefit results in an overestimate of the treatment effect for the primary outcome²⁵ and limits a more comprehensive evaluation of safety and important secondary endpoints.²⁶

The primary hypothesis will be tested using two-sided log-rank tests to compare the time to primary endpoint between the rivaroxaban and ticagrelor arms versus the clopidogrel arm. Patients lost to follow-up will be censored at the last contact date. An HR will be calculated using Cox proportional hazards regression. All analyses will be intention-to-treat unless specified otherwise.

Statistical considerations for the exploratory aim related to CYP2C19 status are provided in Supplementary Materials.

Progress of CAPTIVA trial

The first participant in CAPTIVA was enrolled on 2 August 2022 and, as of 17 December 2024, 575 subjects (34% of the planned sample size) have been enrolled. The 450th subject was randomized on 26 June 2024, so the prespecified safety analysis will be conducted at the end of June 2025, once this subject has been followed for 12 months.

For the subjects enrolled thus far, the median interval between qualifying infarct and randomization is 7.0 days (interquartile range 3.0–17.0).

Discussion

CAPTIVA is essentially two trials: (1) a comparison of low-dose rivaroxaban and aspirin versus clopidogrel and aspirin and (2) a comparison of ticagrelor and aspirin versus clopidogrel and aspirin. The same clopidogrel and aspirin arm is used as the control in both comparisons for efficiency. As such, CAPTIVA will establish whether low-dose rivaroxaban or ticagrelor or both, when added to aspirin, are more effective than the current usual care of clopidogrel plus aspirin for preventing recurrent stroke or vascular death in high-risk patients with ICAS.

Since CAPTIVA is not designed or powered to compare the ticagrelor versus the rivaroxaban arms, there may not be definitive data indicating which of these two treatments is superior if they are both more effective than clopidogrel. Nevertheless, by combining the efficacy, safety, cost, and patient acceptance of these two therapies, a clear choice may emerge for these patients and the physicians taking care of them. This is similar to the scenario with atrial fibrillation in which several novel anticoagulants have been shown to be more effective than warfarin, but none have been directly compared with each other in a randomized trial of subjects with atrial fibrillation.

In conclusion, CAPTIVA will help clarify optimal dual antithrombotic treatment in high-risk patients with ICAS and will provide North American data on the effect of CYP2C19 LOF carrier status on the outcomes of ICAS subjects treated with these dual therapies.