This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record.
1. Trial at a Glance
CLOSE was a randomized phase 3 trial investigating strategies intended to prevent recurrent stroke in patients with ischemic stroke associated with patent foramen ovale.
| Feature | CLOSE |
|---|---|
| Condition | Ischemic Stroke; Patent Foramen Ovale; Atrial Septal Aneurysm; Migraine |
| Allocation | Randomized |
| Design model | Parallel |
| Masking | None |
| Primary purpose | Prevention |
| Lead sponsor | Assistance Publique - Hôpitaux de Paris |
2. Clinical Question
Population
Patients with ischemic stroke, patent foramen ovale, atrial septal aneurysm, or related registry-defined conditions.
Interventions
The trial evaluated aspirin, anticoagulant therapy options, and devices for patent foramen ovale closure.
Comparator
Antiplatelet therapy comparison was included in the randomized study design.
Primary question
Does one strategy reduce stroke recurrence compared with the other randomized strategies?
3. Trial Design
Randomized
3 arms
Parallel design
Follow-up 2–9 years
| Arm | Intervention |
|---|---|
| Device strategy | Devices for PFO closure |
| Drug strategy | Antivitamins K or rivaroxaban or dabigatran or apixaban |
| Antiplatelet strategy | Aspirin |
4. Endpoints
| Endpoint | Registry definition | Time frame |
|---|---|---|
| Stroke (fatal or not) | Stroke occurrence | During the follow up (between 2 or 9 years) |
5. Planned Analysis
The registry identifies stroke (fatal or not) during follow-up between 2 or 9 years as the primary endpoint. Clinical trial analyses of recurrent stroke outcomes commonly use time-to-event methods because participants may have different follow-up durations and may not all experience an event.
Time-to-event analysis concept
For recurrent stroke prevention studies, methods such as Kaplan–Meier estimation, log-rank comparisons, and hazard models are commonly used to compare event occurrence over follow-up. The ClinicalTrials.gov record does not report posted statistical analyses or outcome estimates for this endpoint.
6. Statistical Methodology
Randomized parallel comparison
Randomization is designed to balance measured and unmeasured factors across treatment groups. The parallel design means participants are assigned to a strategy and followed over time rather than sequentially receiving multiple interventions.
Time-to-event outcomes
The primary endpoint is an event occurring over a follow-up period. Statistical analysis must account for varying observation times and participants who complete follow-up without experiencing stroke.
Clinical interpretation of a hazard ratio
If a hazard ratio analysis is used, it compares estimated event rates over time rather than directly comparing the percentage of participants with an event at a single time point.
7. Statistical Methods Explained
Why is stroke recurrence a time-to-event outcome?
Stroke recurrence can occur at different times after enrollment. Time-to-event methods preserve information from the entire follow-up period.
Why does follow-up duration matter?
The registry reports a follow-up range between 2 or 9 years, meaning participants may contribute different amounts of observation time.
Why is randomization important?
Random allocation supports unbiased comparison between treatment strategies by reducing systematic differences between groups.
Why are multiple study arms important statistically?
With three arms, interpretation requires attention to the comparisons being made and whether the statistical analysis accounts for multiple comparisons.
Why is masking reported as none?
The registry reports no masking. Lack of masking can influence outcomes that depend on clinical decisions or subjective assessment, although objective endpoints may be less affected.
8. Limitations
- The ClinicalTrials.gov record does not report posted statistical analyses or outcome estimates.
- No confidence intervals, p-values, or treatment-effect estimates are reported in the registry record.
- The primary endpoint has a long follow-up window, which requires careful handling of censoring and follow-up differences.
- With three randomized strategies, interpretation depends on the prespecified comparison framework and multiplicity considerations.
9. Why This Trial Matters Statistically
CLOSE is a useful teaching example because it illustrates several important concepts in clinical trial statistics: randomized allocation, multi-arm comparisons, prevention endpoints, and analysis of events occurring over extended follow-up.
| Concept | How it appears in CLOSE |
|---|---|
| Randomization | Randomized allocation |
| Parallel design | Three-arm parallel trial |
| Time-to-event analysis | Stroke occurrence during follow-up |
| Long follow-up | Between 2 or 9 years |
| Prevention endpoint | Stroke recurrence prevention |
10. Sources
- ClinicalTrials.gov record: NCT00562289
- PubMed: 28902593
- PubMed: 32381162
- PubMed: 29910193
- PubMed: 27056964