Rivaroxaban for secondary stroke prevention in patients with embolic strokes of undetermined source: Design of the NAVIGATE ESUS randomized trial

Overview and timelines

NAVIGATE ESUS (Clinicaltrials.gov.NCT02313909) is an international, double-blinded, randomized phase III trial comparing rivaroxaban 15 mg once daily (immediate-release, film-coated tablets) with aspirin (enteric-coated) 100 mg once daily, both to be taken with food, in patients with recent ESUS (Figure 1). Matching placebos will be used in this double-blinded trial. The primary hypothesis is that rivaroxaban is superior to aspirin for reducing the risk of recurrent stroke and systemic embolism (primary efficacy outcome). Seven thousand patients will be enrolled to detect a ≥30% reduction in primary efficacy outcome events with 90% power by assignment to rivaroxaban based on an estimated rate of 3.8% per year among aspirin-assigned patients. ¹ Participants will be randomized between seven days and six months following the qualifying ESUS at about 480 sites in 31 countries (Figure 2). Patient recruitment began in December 2014 and is anticipated to finish in 2017. Mean patient follow-up is expected to be about two years, but the study will continue until at least 450 participants have experienced a primary efficacy outcome event. The main results are anticipated to be available in 2018. OPEN IN VIEWER

OPEN IN VIEWER

Figure 2.

Countries participating in NAVIGATE ESUS.

Study population

Screening ischemic stroke patients for participation is based on five key eligibility criteria as summarized in Table 1. In short, patients ≥50 years old with nonlacunar ischemic stroke visualized by neuroimaging and without a clear etiology are included for whom most current guidelines recommend antiplatelet therapy for secondary prevention (Figure 3). After the qualifying stroke, at least 20 h of cardiac rhythm monitoring is required to exclude atrial fibrillation lasting >6 min, although investigators can choose to monitor for longer periods per local clinical practice standards. Intracranial arterial imaging is not required, but if done, the presence of >50% intracranial atherosclerotic stenosis supplying the ischemic area excludes participation. Patients with left ventricular systolic dysfunction of any severity, patent foramen ovale (PFO), and all types of aortic arch plaque are eligible if anticoagulation (or PFO closure) is not planned and potential randomization to aspirin is acceptable to local investigators. Transthoracic echocardiography is mandatory, with transesophageal echocardiography (TEE) optional and an acceptable substitute, with intracardiac thrombus detected by either technique an exclusion criterion. Patients with carotid artery atherosclerotic plaques causing ≤50% stenosis are eligible regardless of ulceration or other features. The outer time limit (i.e. six months) between qualifying stroke and randomization was chosen because the temporal pattern of recurrent stroke is not known for ESUS patients and it was deemed worthwhile to determine the absolute benefits of anticoagulation for ESUS patients who are identified after the acute phase due to diagnostic delays.

OPEN IN VIEWER

Table 1.

Inclusion criteria.

Embolic stroke of undetermined source (ESUS) a between seven days b and six months:

1. Ischemic stroke visualized by CT or MRI that is not lacunar c

2. Absence of extracranial and, if intracranial imaging performed, intracranial atherosclerosis causing >50% luminal stenosis of arteries supplying the area of ischemia d

3. No atrial fibrillation by history, ECG, or after >20 h of cardiac rhythm monitoring e

4. No intracardiac thrombus by echocardiography

5. No other specific cause of ischemic stroke identified (e.g. high-risk cardiac source f usually requiring anticoagulation, cardiac tumor, arteritis, dissection, migraine/vasospasm, cerebral venous thrombosis, drug abuse)

Age ≥50 years g

Ability and willingness to provide written informed consent

OPEN IN VIEWER

Figure 3.

Conceptual scheme for screening to identify ESUS patients. Prior to diagnostic testing, approximately 15% of strokes would be ESUS. With each additional step in screening, the probability increases toward 100%. ⁵ *After excluding less uncommon stroke etiologies (e.g. dissections).

Exclusion criteria include severely disabling stroke (modified Rankin score ≥4 at screening), the presence or plan to insert an implantable ECG loop recorder, specific indication for chronic anticoagulation or for chronic antiplatelet therapy, ongoing regular use of conventional nonsteroidal anti-inflammatory drugs, previous nontraumatic intracranial hemorrhage (an exception is hemorrhagic transformation of ischemic stroke), and required use of strong inhibitors of both cytochrome P450 isoenzyme 3A4 (CYP3A4) and P-glycoprotein (e.g. protease inhibitors and several azole-antimycotic agents) (Table 2).

OPEN IN VIEWER

Table 2.

Key exclusion criteria.

1.	Severely disabling stroke (modified Rankin score ≥4 at screening)
2.	Patent foramen ovale with plans for closure
3.	Known serious infection or inflammatory disease that may be the cause of stroke
4.	Patient has or is intended to receive an implantable ECG loop recorder
5.	Indication for chronic anticoagulation
6.	Indication for chronic antiplatelet therapy
7.	Active bleeding/major bleeding within last six months/ previous nontraumatic intracranial hemorrhage (any type, ever)/high risk for serious bleeding
8.	Hepatic disease associated with coagulopathy
9.	Renal disease with estimated GFR < 30 ml/min/1.73 m2
10.	Life expectancy less than six months
11.	Use of strong inhibitors of both cytochrome P450 isoenzyme 3A4 (CYP3A4) and P-glycoprotein (e.g. protease inhibitors and several azole-antimycotic agents)
12.	Female of childbearing potential who is not surgically sterile or who is sexually active and not using reliable contraception
13.	Chronic, regular use of a conventional nonsteroidal anti-inflammatory drug

Patients are randomly allocated by an interactive voice/web response system to either rivaroxaban or aspirin in a 1:1 ratio, with the block size sequestered until the end of recruitment. Randomization is stratified by country and by age <60 and ≥60 years.

Follow-up

Participants return for outpatient office visits at one, six, and 12 months and then every six months until a common end-study date. At three months, participants are contacted by telephone, and a telephone contact is done one month after the end-of-treatment visit. Participants are assessed for the occurrence of safety and efficacy events, adherence, adverse events and vital signs, and quality of life, functional status and cognitive function are recorded. Adherence to assigned therapy is assessed by interview and pill counts at each visit. Low-dose (≤100 mg daily) aspirin in addition to study-assigned therapy is permitted if a new indication develops during follow-up (e.g. stable coronary artery disease), but dual antiplatelet therapy mandates cessation of study drug for the duration of use.

Outcome events

The primary efficacy outcome is time to recurrent stroke (ischemic, hemorrhagic, and undefined stroke, including TIAs with positive neuroimaging) or systemic embolism. This outcome was chosen in part to align with trials demonstrating a substantial reduction in the risk of embolic events by DOACs in atrial fibrillation patients. Secondary efficacy outcomes are outlined in Table 3. The primary safety outcome is major bleeding according to the criteria of the International Society of Thrombosis and Haemostasis. ¹² We anticipate that the annual rate of intracerebral hemorrhage will be relatively low (<0.4%/yr) due to prohibition of concomitant dual antiplatelet therapy, exclusion of lacunar strokes as qualifying events, and delayed initiation of anticoagulation for a least seven days after the qualifying stroke, and be equal in both treatment arms. Extracranial major hemorrhage is anticipated to be increased among those assigned rivaroxaban versus aspirin (2%/year versus 1%/year, respectively).^13,14

OPEN IN VIEWER

Table 3.

Trial outcome events.

Primary efficacy outcome:

• Stroke (ischemic, hemorrhagic, and undefined stroke, TIA with positive neuroimaging) and systemic embolism

Secondary efficacy outcomes: a

• Cardiovascular death (including death due to all types of hemorrhage), recurrent stroke, systemic embolism, and myocardial infarction

• All-cause mortality

• Individual components of the primary and secondary efficacy outcomes (stroke, cardiovascular death, and myocardial infarction) as well as

• Recurrent ischemic stroke (including TIA with positive neuroimaging)

• Disabling/fatal stroke (modified Rankin score ≥4) at 90 days after recurrence

Primary safety outcome:

• Major bleeding meeting ISTH criteria 12

Secondary safety outcomes:

• Life-threatening bleeding

• Clinically relevant nonmajor bleeding

• Intracranial hemorrhage

Statistical analysis plan

The primary efficacy analyses will be based on the intent-to-treat population. Rivaroxaban-assigned patients will be compared with the aspirin control group using a log–rank test. Kaplan–Meier estimates will be used to plot the cumulative incidence risk over time. Risk reduction will be estimated with the Cox proportional hazards model. Secondary efficacy outcomes will be analyzed using similar methods as for the primary efficacy analysis, with testing performed in hierarchical order to control the global type 1 error level (Table 3). The statistical analysis plan includes several prespecified subgroup analyses based on age, sex, global region, and time from index stroke to randomization of <30 days, 31 days to 3 months, and >3 months.

Study development and committees

The trial coprincipal investigators from the Population Health Research Institute (RGH, SJC) initially proposed the trial to Bayer Pharma AG (SDB). The Steering Committee is responsible for the scientific leadership and consists of the national leaders (n = 32) of each participating country, the coprincipal investigators, sponsor representatives, and additional experts in thrombosis and clinical trials. The Publications Committee approves all manuscripts prior to submission and consists of four national leaders (rotating annually), the coprincipal investigators, and a sponsor representative. Sponsor representatives (who comprise <15% of these two committees) have voting rights. Outcome event verification is overseen by a central Adjudication Committee that will assess whether reported outcome events meet study criteria.

An independent Data Monitoring Committee monitors participant safety on an ongoing basis and may recommend modification of the study protocol in case unexpected safety concerns arise or may terminate the study for safety concerns that are not offset by benefits related to stroke reduction. Two formal interim analyses will occur when approximately 50 and 67% of primary efficacy outcome events have accrued, and the trial will be stopped for overwhelming efficacy by one treatment relative to the other if not mitigated by safety issues.

Two substudies

The primary objective of the MIND MRI Substudy is to determine the effect of rivaroxaban versus aspirin on MRI-defined clinically silent (or covert) brain infarcts and clinical ischemic strokes. Covert infarcts are neither asymptomatic nor benign and result in considerable morbidity including cognitive decline, loss of independence, gait impairment, and falls. While stroke recurrence increases the likelihood of dementia, poststroke cognitive decline often occurs in the absence of clinical recurrence. The mechanism is thought to be either a manifestation of the increasing vascular burden or an interaction between subclinical ischemia and coexisting neurodegenerative pathology, principally Alzheimer’s disease. Therapies aimed at interrupting the accumulating vascular pathology and its consequences may improve the cognitive and functional trajectory of ESUS patients.

The incidence of covert brain infarcts is expected to exceed recurrent clinical ischemic strokes, but owing to the possible heterogeneous pathogenesis of these lesions, covert infarcts may be less sensitive to anticoagulants regarding prevention. About 1000 participants will undergo brain MRI using a standardized acquisition protocol near the time of randomization that will be repeated near study end. MRIs will be interpreted at a core MRI facility by experts unaware of treatment assignment.

The Biomarker, Genetics, Gene Expression Substudy will collect plasma, RNA, and DNA samples from up to 3000 participants for exploratory analyses aimed at establishing whether ESUS is a distinct clinical entity or the result of multiple unrelated causes that have in common a propensity for thrombosis. The substudy will integrate biomarker information, comprehensive genomics characterization, and gene expression (including miRNA) data to identify pathways linked to ESUS and stroke recurrence. Integration of these approaches has the potential to be more revealing than the use of any one alone. Anticipated analyses include the predictive value of D-dimer for recurrent stroke, NT-proBNP levels as predicting atrial fibrillation during follow-up, and several as yet undefined, exploratory biomarkers for predicting clinical events.

Choice of aspirin as the control antiplatelet therapy

Aspirin is the best characterized and most widely used antiplatelet therapy for secondary stroke prevention. Most experts believe that there is no compelling evidence that other antiplatelet agents offer important benefits over aspirin, and hence many major guidelines include aspirin as acceptable chronic antiplatelet therapy for secondary stroke prevention, including for patients with cryptogenic ischemic strokes.^15–17 However, in some countries, aspirin monotherapy is not the preferred antiplatelet therapy for secondary stroke prevention. ¹⁸ If investigators were allowed to choose the antiplatelet comparator and dose based on local preference, a double-blind comparison with rivaroxaban would not be feasible, and it was elected to choose a single antiplatelet comparator. For ESUS patients, there are no existing data about relative efficacy of different antiplatelet therapies. Hence, aspirin is the most reasonable choice for the antiplatelet control arm of NAVIGATE ESUS and in-line with current global practice.^5,18

Choice of 15 mg once daily dose of rivaroxaban

The selection of a dose of anticoagulant for a new indication often involves a phase 2 dose-ranging study. However, such a study was not considered feasible due to the relatively low rate of recurrent stroke in ESUS patients and the absence of an established surrogate biomarker. Consequently, the dosage was extrapolated from clinical evidence on the use of rivaroxaban in other patient populations. Rivaroxaban 20 mg daily (15 mg daily for those with an estimated creatinine clearance between 30 and 49 ml/min) was efficacious and safe in patients with atrial fibrillation, in whom embolism is the dominant cause of stroke. ¹⁹ Rivaroxaban 15 mg daily dose appears to be efficacious for stroke prevention in Japanese patients with atrial fibrillation. ²⁰ Due to concern about a potentially higher rate of intracerebral hemorrhage in patients with recent ischemic stroke, a dosage of 15 mg once daily in ESUS patients, without dose reduction for moderate renal impairment, was deemed appropriate. Modeling data show an overlap of exposure (i.e. of rivaroxaban plasma concentrations over time) for the 15 and 20 mg doses.²¹ For this patient population, rivaroxaban 15 mg daily was chosen to optimally balance efficacy with safety.

Screening participants for covert paroxysmal atrial fibrillation

When the trial was designed in 2013–2014, the approach to detection of atrial fibrillation at many stroke centers was an ECG at hospital admission and 24 h of inpatient telemetry or Holter monitoring.^5,22 In mid-2014, two influential studies reported that brief episodes (several minutes) of previously unrecognized atrial fibrillation could be detected in 10–20% of patients with cryptogenic ischemic stroke if the duration of cardiac monitoring was prolonged to 30 days.^23,24 However, it is unknown whether brief (≤6 min) episodes of atrial fibrillation detected weeks or months after stroke identify patients who benefit from anticoagulation. Further, short episodes of atrial fibrillation are not necessarily temporally associated with stroke risk, raising further doubt about their relevance to the pathogenesis of stroke. ²⁵

For NAVIGATE ESUS eligibility, screening for covert paroxysmal atrial fibrillation with ≥20 h of cardiac rhythm monitoring is required, with episodes >6 min duration mandating exclusion based on the ASSERT trial. ²⁶ Sites are permitted to monitor for longer periods (e.g. 7–30 days) based on local clinical practice, but monitoring must be completed prior to randomization. Randomization of patients undergoing prolonged (e.g. months to years) cardiac rhythm monitoring using implanted recorders is an exclusion criterion because of the uncertain implications of subgroup analysis in those with implanted monitoring devices and the likelihood of cross-over based on current enthusiasm in some countries for the use of anticoagulants if even brief episodes of atrial fibrillation are detected.

If NAVIGATE ESUS results demonstrate a reduction in recurrent ischemic stroke with anticoagulation, could this overall result be driven by a large treatment effect among a subgroup of subjects with covert paroxysmal atrial fibrillation? While this is unlikely based on predicted frequencies, this possibility will be assessed by routine clinical screening to determine if atrial fibrillation is present at the time of primary efficacy events and comparing the frequency of atrial fibrillation between treatment arms. Further, if rivaroxaban shows superior efficacy versus aspirin and is relatively safe for ESUS patients, it will challenge the need for prolonged cardiac rhythm monitoring for most patients with cryptogenic stroke.

Relative effects of rivaroxaban versus aspirin on different embolic sources

ESUS includes multiple sources (cardiac, arterial, paradoxical). Are different embolic sources likely to respond similarly to rivaroxaban versus aspirin? Clinical trials have not shown a benefit of anticoagulation over antiplatelet therapy for secondary prevention of stroke due to arterial sources (including intracranial stenosis and aortic arch atheroma), but subgroup analyses support the biologic plausibility if a consistent anticoagulant effect can be maintained.^27–29 Further, often two or more potential embolic sources are detected in individual ESUS patients, and the specific culprit cannot be identified, so that it is pragmatic to assess the relative effect of antithrombotic therapies among all ESUS patients. A series of exploratory subgroup analyses will examine recurrent stroke rates and response to anticoagulation associated with individual potential embolic sources as well as the presence of prothrombotic disorders (e.g. active cancer).^30,31

Should TEE be required for eligibility?

Minor-risk embolic sources are more frequently detected using TEE than with precordial echocardiography, ³² and TEE is sometimes recommended routinely for young patients with cryptogenic stroke. However, global surveys indicate that TEE is infrequently done in routine clinical practice,^5,22 and with uncertain evidence-based management implications. ³³ While it would be ideal to more accurately characterize minor-risk embolic sources, mandating TEE (mildly invasive and costly) for trial eligibility would severely restrict the number of participating sites and bias recruitment to patients willing to undergo this procedure.