PICSS: Patent Foramen Ovale in Cryptogenic Stroke Study: Low-Intensity Warfarin vs Aspirin in the WARSS TEE Substudy (Homma et al., 2002)
In patients with cryptogenic stroke and a patent foramen ovale, does low-intensity warfarin reduce recurrent stroke or death compared with aspirin? Read this as a subgroup of a substudy, not as a trial result.
Homma S et al. (Circulation 2002;105(22):2625-2631) · doi:10.1161/01.CIR.0000017498.88393.44 · 98 patients
Population
Included
- Enrolled in WARSS: non-cardioembolic ischemic stroke, excluding atrial fibrillation and extracranial carotid stenosis
- Underwent transesophageal echocardiography (referral was clinician-driven, not protocolised)
- For the published subgroup: cryptogenic stroke subtype AND a patent foramen ovale on TEE
Verifier-confirmed inclusion
- Enrolled in WARSS: non-cardioembolic ischemic stroke
- Underwent transesophageal echocardiography as part of the PICSS substudy (630 of the WARSS cohort, 42 US centres, 1993 to 2000)
- For the subgroup NeuroWiki publishes: cryptogenic stroke subtype AND a patent foramen ovale on TEE (98 patients, warfarin 42 and aspirin 56)
Excluded
- Atrial fibrillation (excluded by the parent trial)
- Extracranial carotid stenosis (excluded by the parent trial)
- Cardioembolic stroke mechanism
Verifier-confirmed exclusion (parent trial)
- Atrial fibrillation
- Extracranial carotid stenosis
Remaining exclusions (not recoverable)
- The full published exclusion list was not retrieved: ahajournals.org returned 403 to the evidence-verifier on 2026-07-31 and the PICSS and WARSS full texts were not read. 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
SUBGROUP, NOT A PRIMARY
NOT A TRIAL RESULT. This is a non-prespecified subgroup of the PICSS transesophageal-echo substudy of WARSS: 98 patients and 14 events in total. The anticoagulant was LOW-INTENSITY warfarin, INR 1.4 to 2.8, so nothing here transfers to modern anticoagulation or to any direct oral anticoagulant.
Low-intensity warfarin
4 of 42
9.5% recurrent stroke or death at 2 years
Aspirin 325 mg
10 of 56
17.9% recurrent stroke or death at 2 years
Hazard ratio
0.52
95% CI 0.16 to 1.67 · p = 0.28
The interval includes 1, so no difference was demonstrated.
Per-arm counts are shown rather than a 100-dot grid because one patient is 2.4 percentage points in the warfarin arm: a dot grid would imply a precision these 98 patients cannot support. Two published extractions of this subgroup disagree on the numerators while carrying the same hazard ratio and interval, which cannot both be right; the composite figures shown here match the endpoint the parent trial actually set. No NNT is computed.
Study Arms
- Agent
- Warfarin
- Dose
- Titrated to a target INR of 1.4 to 2.8
- Route
- Oral
- Frequency
- Daily, dose-adjusted
- Duration
- 2 years
THE governing caveat on this trial. This is LOW-INTENSITY anticoagulation. It is not the INR 2 to 3 used in CLOSE and it is not a direct oral anticoagulant, so PICSS supports no inference about standard-intensity vitamin K antagonism or about any DOAC.
- Agent
- Aspirin
- Dose
- 325 mg daily
- Route
- Oral
- Frequency
- Daily
- Duration
- 2 years
Double-blind against warfarin in the parent WARSS design.
Trial Design
Type
- Transesophageal-echocardiography SUBSTUDY of the WARSS randomized trial, not a standalone trial
- Parent design: double-blind randomized comparison of low-intensity warfarin (INR 1.4 to 2.8) against aspirin 325 mg daily
- Parent population: non-cardioembolic ischemic stroke, excluding atrial fibrillation and extracranial carotid stenosis
- TEE referral was clinician-driven rather than protocolised, so PICSS is a SELECTED subset of WARSS
- The cryptogenic-stroke-with-PFO comparison is a NON-PRESPECIFIED SUBGROUP of that substudy
- Parent primary endpoint: recurrent ischemic stroke or death from any cause within two years
Timeline
42 US centres, enrolled 1993 to 2000; 2-year follow-up. Published Circulation 2002;105(22):2625-2631
N
98
Enrollment
630 WARSS patients underwent transesophageal echocardiography and a PFO was found in 203 (33.8%). The cell displayed here is the 98 patients who had both cryptogenic stroke and a PFO. WARSS randomized warfarin titrated to INR 1.4 to 2.8 against aspirin 325 mg daily, double-blind. Published Circulation 2002; parent trial NEJM 2001.
Bedside Pearl
PICSS (Circulation 2002) cannot answer whether anticoagulation beats aspirin in PFO-associated cryptogenic stroke. The quoted comparison, 4 of 42 (9.5%) vs 10 of 56 (17.9%) for recurrent ischemic stroke or death at 2 years, HR 0.52 (95% CI 0.16 to 1.67), P=0.28, is a NON-PRESPECIFIED SUBGROUP OF A SUBSTUDY resting on 14 events, and the anticoagulant was LOW-INTENSITY warfarin at INR 1.4 to 2.8, not modern anticoagulation. Hypothesis-generating only. The AAN 2020 advisory (Statement 3a, Level C) treats the two agents as equally acceptable. No NNT.