Kevin O’Connor, MD
About a third of ischemic strokes are categorized as being cryptogenic. Embolic strokes of unknown source (ESUS) represent a subset of these cryptogenic strokes. Clinicians are often faced with choosing between an anticoagulant and a platelet antiaggregant as one component of secondary prevention in patients having an ESUS. The NAVIGATE-ESUS trial provides data on the frequency and predictors of major bleeding in ESUS patients based on antithrombotic therapy choice, which can help inform treatment decisions.
NAVIGATE-ESUS was an international, double-blind, phase III trial that included 7213 participants from 31 countries who were randomized to receive rivaroxaban 15 mg once daily or aspirin 100 mg once daily. Sites were in Europe, East Asia, North America, and Latin America. The primary outcome of this analysis was major bleeding. Clinically-relevant nonmajor bleeding (CRNMB) was excluded for several reasons: CRNMB determinations were not determined centrally, whereas there was centralized adjudication of major bleeding events; intracranial bleeding was not considered to be a CRNMB; and analysis of independent predictors of major bleeding and CRNMBs such as epistaxis requiring medical attention would have been potentially spurious. NAVIGATE-ESUS was terminated before targeted enrollment was completed after an interim analysis found increased bleeding among rivaroxaban-treated participants, but no reduction in recurrent stroke.
During a median follow-up of 11 months, 85 first major bleeds occurred with 62 among patients on rivaroxaban and 23 among patients on aspirin. Univariable analyses identified older age, East Asia region, Asian race, systolic blood pressure (SBP) ≥ 160 mm Hg, reduced estimated glomerular filtration rate (eGFR, < 50 mL/min per 1.73 m2), and assignment to rivaroxaban as being associated with major bleeding. Independently associated characteristics, as determined by multivariable analysis, included East Asia region, Asian race, SBP ≥ 160 mm Hg, reduced eGFR, and assignment to rivaroxaban.
A possible explanation of the increased bleeding risk associated with reduced eGFR is higher levels of rivaroxaban stemming from its impaired renal clearance. Although it is primarily hepatically metabolized, rivaroxaban is partially renally excreted. The underlying reason for the association between reduced eGFR and increased bleeding with aspirin use is less clear.
Participants from East Asia had a six-fold higher annualized rate of ICH compared to other regions (0.67%/year versus 0.11%/year with a wide confidence interval (HR, 6.3 [95% CI, 2.2-18]). Although East Asians comprised 19% of the trial population, they accounted for 60% of the ICHs (9 of 15). Asia region as an independent risk factor for bleeding has previously been reported in RE-SPECT ESUS, as well as studies of non-stroke patients receiving antithrombotic therapy.
The majority of major bleeding in NAVIGATE-ESUS was gastrointestinal (38%), 29% was intracranial, and the remainder occurred elsewhere. Univariable analyses revealed region, diastolic blood pressure, prior stroke or transient ischemic attack, and hemorrhagic transformation of the qualifying stroke (inclusion criteria) were associated with nontraumatic intracerebral hemorrhage (ICH). A majority (7 of 10) of ICHs were ipsilateral to the qualifying infarct.
This study identified several predictors of major bleeding associated with antithrombotic therapy in ESUS patients. In particular, impaired renal function is an independent predictor of major bleeding with antithrombotic therapy, both with rivaroxaban and with aspirin. Although the underlying cause of this association remains unclear, it is important to remember when risk stratifying patients with ESUS. Although the East Asian region and Asian race were also independent predictors, they could not be differentiated as only 67 Asian participants were outside the East Asian region, whereas 3 non-Asians were within the region. As such, it may be difficult to rely on either predictor independent of the other to stratify risk. Data for the size of infarcts was not available; however, knowing the relationship between infarct size and ICH would be useful when considering antithrombotic treatments for patients with an ESUS. Oral anticoagulants have been separately evaluated in the setting of ESUS, but they have not been compared in the same trial. Such a study may be helpful to further guide antithrombotic decisions when managing ESUS patients.
