NAVIGATE-ESUS (PFO subgroup): Rivaroxaban vs Aspirin in the Prespecified Patent Foramen Ovale Subgroup of NAVIGATE-ESUS (Kasner et al., 2018)
In embolic stroke of undetermined source with a detected patent foramen ovale, does rivaroxaban 15 mg reduce recurrent ischaemic stroke compared with aspirin? This is the prespecified PFO subgroup.
Kasner SE et al. (Lancet Neurology 2018;17(12):1053-1060); parent Hart RG et al. (NEJM 2018;378(23):2191-2201) · doi:10.1016/S1474-4422(18)30319-3 · 534 patients
Population
Included
- Embolic stroke of undetermined source (ESUS)
- For the subgroup: a patent foramen ovale detected during standard workup
Verifier-confirmed inclusion
- Embolic stroke of undetermined source, enrolled at 459 centres in 31 countries
- For the PFO subgroup: a patent foramen ovale detected during workup (534 of 7,213 patients, 7.4%; rivaroxaban 259 and aspirin 275)
Excluded
- Atrial fibrillation on the monitoring performed before randomization
- An identified cardioembolic, large-artery or small-vessel stroke mechanism (the ESUS construct excludes these by definition)
Ascertainment note that functions as an eligibility caveat
- The trial did NOT mandate transesophageal echocardiography or bubble contrast, so entry into the PFO subgroup depended on whichever imaging the enrolling site happened to perform. A 7.4% detected prevalence, against the 33.8% PICSS detected in a cohort that did undergo transesophageal echocardiography, means the subgroup is incompletely and non-randomly ascertained. This is not a PFO-selected population.
Remaining exclusions (not recoverable)
- The full published exclusion list was not retrieved by the evidence packet. The ClinicalTrials.gov record (NCT02313909) was not mirrored, following the precedent set for the PFO-for-migraine cluster. Read this as a gap in what NeuroWiki can source, not as an absence of exclusion criteria in the trial.
Result in the PFO group
PRESPECIFIED SUBGROUP, 45% POWER
PRESPECIFIED SUBGROUP AT 45% POWER. The parent trial was stopped early at interim for futility PLUS excess bleeding, at a median follow-up of 11 months, and the authors record that substantial imprecision remains. The interaction test was null (p=0.18): PFO status did not modify the treatment effect.
Rivaroxaban 15 mg
2.6
events per 100 patient-years (7 events)
Aspirin 100 mg
4.8
events per 100 patient-years (13 events)
Hazard ratio
0.54
95% CI 0.22 to 1.36
The interval includes 1, so no difference was demonstrated.
These two figures are RATES PER 100 PATIENT-YEARS, not percentages, and they are not drawn as a dot grid for that reason. Major bleeding in this subgroup ran the other way, hazard ratio 2.05 with a 95% interval of 0.51 to 8.18, which spans a halving and an eight-fold increase and cannot be interpreted. In the parent trial major bleeding was significantly higher with rivaroxaban. No NNT is computed.
Study Arms
- Agent
- Rivaroxaban
- Dose
- 15 mg once daily
- Route
- Oral
- Frequency
- Once daily
Note the dose. 15 mg once daily is NOT the 20 mg dose used for stroke prevention in atrial fibrillation, so this trial does not test the AF-strength regimen.
- Agent
- Aspirin
- Dose
- 100 mg daily
- Route
- Oral
- Frequency
- Once daily
Trial Design
Type
- Prespecified PFO subgroup of NAVIGATE-ESUS, a randomized superiority trial of rivaroxaban 15 mg once daily against aspirin 100 mg daily
- Parent population: 7,213 patients with embolic stroke of undetermined source at 459 centres in 31 countries
- Parent trial TERMINATED EARLY at interim analysis for FUTILITY PLUS EXCESS BLEEDING; median follow-up 11 months
- The subgroup analysis was prespecified, described by the authors as planned before completion of the trial
- Subgroup power 45%; the authors state that substantial imprecision remains
- The trial did NOT mandate transesophageal echocardiography or bubble contrast, so PFO ascertainment is incomplete and non-random
Timeline
Median follow-up 11 months after early termination. Subgroup published Lancet Neurology 2018;17(12):1053-1060; parent NEJM 2018;378(23):2191-2201
N
534
Enrollment
7,213 patients with embolic stroke of undetermined source at 459 centres in 31 countries. A PFO was detected in 534 (7.4%), rivaroxaban 259 and aspirin 275. Note the dose: rivaroxaban 15 mg once daily, not the 20 mg used for atrial fibrillation. Neither this trial nor RE-SPECT ESUS mandated transesophageal echocardiography or bubble contrast, so the subgroup is incompletely and non-randomly ascertained. Published Lancet Neurology 2018; parent trial NEJM 2018.
ClinicalTrials.gov
NCT02313909Bedside Pearl
NAVIGATE-ESUS PFO subgroup (Lancet Neurol 2018): rivaroxaban 15 mg vs aspirin 100 mg in the 534 of 7,213 patients (7.4%) with a detected PFO. HR 0.54 (95% CI 0.22 to 1.36) on 20 events, p-interaction 0.18. The subgroup WAS prespecified but had 45% POWER, in a parent trial stopped early for FUTILITY PLUS EXCESS BLEEDING (parent major bleeding 1.8% vs 0.7% per year). Subgroup major bleeding HR 2.05 (95% CI 0.51 to 8.18), too imprecise to act on. The pooled OR 0.48 (95% CI 0.24 to 0.96) reported in this paper became OR 0.70 (95% CI 0.43 to 1.14) once RE-SPECT ESUS was added to the same pool. No NNT.