PC Trial
Percutaneous Closure of Patent Foramen Ovale in Cryptogenic Embolism
Clinical Question
In patients under 60 years with a patent foramen ovale and cryptogenic ischemic stroke, TIA, or peripheral thromboembolism, does percutaneous closure with the Amplatzer PFO Occluder reduce recurrent embolic events or death compared with medical therapy alone?
Bottom Line
In 414 patients <60 years with a PFO and cryptogenic ischemic stroke, TIA, or peripheral thromboembolism, transcatheter PFO closure with the Amplatzer PFO Occluder did not significantly reduce the composite of death, nonfatal stroke, TIA, or peripheral embolism versus medical therapy (3.4% vs 5.2%; HR 0.63, 95% CI 0.24-1.62; P=0.34) over a mean follow-up of ~4 years. Trends favored closure for stroke (HR 0.20) but the trial was substantially underpowered because event rates were less than half of those assumed at design.
Major Points
- Design: investigator-initiated, multicenter, open-label, randomized superiority trial with blinded end-point adjudication in 29 sites across Europe, Canada, Brazil, and Australia (2000-2009).
- Population: 414 patients <60 y with a PFO on TEE and cryptogenic ischemic stroke (79.2%), TIA with a corresponding cerebral ischemic lesion (18.1%), or peripheral thromboembolism (2.7%); mean age 44 y, 49% male, ~24% atrial septal aneurysm, ~22% large right-to-left shunt.
- Intervention: percutaneous PFO closure with the Amplatzer PFO Occluder (n=204) with ASA 100-325 mg/d for >=5-6 months plus ticlopidine or clopidogrel for 1-6 months, versus medical therapy (n=210) with antiplatelet and/or oral anticoagulation at treating-physician discretion.
- Procedure metrics: device implantation attempted in 196/204 and successful in 188 (95.9%); effective closure (no/minimal shunt) achieved in 142/148 (95.9%) at 6-month TEE.
- Primary composite end point (death, nonfatal stroke, TIA, peripheral embolism): 7 (3.4%) closure vs 11 (5.2%) medical over mean follow-up of 4.1 vs 4.0 years; HR 0.63 (95% CI 0.24-1.62), P=0.34; per-protocol HR 0.70 (95% CI 0.27-1.85), P=0.48.
- Stroke: 1 (0.5%) closure vs 5 (2.4%) medical; HR 0.20 (95% CI 0.02-1.72), P=0.14. Exploratory RESPECT-style stroke definition: HR 0.14 (95% CI 0.02-1.17), P=0.07.
- TIA: 5 (2.5%) vs 7 (3.3%); HR 0.71 (95% CI 0.23-2.24), P=0.56. Death: 2 (1.0%) vs 0; RR 5.20 (95% CI 0.25-107.61), P=0.24 (both noncardiovascular — COPD, glioma). No peripheral embolic events in either arm.
- New-onset atrial fibrillation: 6 (2.9%) closure vs 2 (1.0%) medical (HR 3.15, 95% CI 0.64-15.6, P=0.16); no device-associated thrombi; bleeding 3.9% vs 5.7% (HR 0.66, 95% CI 0.27-1.62, P=0.40); PFO-related hospitalization 6.4% vs 6.2%.
- Antithrombotic use diverged: from 12 months onward, medical-therapy patients received significantly more antithrombotic and oral-anticoagulant therapy (P<0.001). 28/210 medical-therapy patients crossed over to closure (median 8.8 months).
- Trial designed to detect a 66% relative reduction (12% to 4% event rate over 4.5 y); observed medical-arm rate was 5.2%, giving <40% power — the null result carries a real risk of type II error and preceded the definitive REDUCE, CLOSE, and long-term RESPECT results that later established closure benefit in selected patients.
Design
Study Type: Multicenter, prospective, open-label, randomized superiority trial with blinded (masked) end-point adjudication
Randomization: 1
Blinding: Open-label treatment; clinical-events committee unaware of study-group assignments adjudicated all potential end points
Enrollment Period: February 24, 2000 - February 19, 2009
Follow-up Duration: Mean 4.1 years in the closure group and 4.0 years in the medical-therapy group (up to 5 years planned); 845.1 and 835.0 patient-years accrued
Centers: 29
Countries: Europe (multiple), Canada, Brazil, Australia
Sample Size: 414
Power Calculation: 205 patients per group provided 80% power to detect a reduction in the primary composite from 3% to 1% per year over a mean 4.5-year follow-up at alpha=0.0492 (allowing for one interim analysis; none was performed, so alpha=0.05 was used). Observed medical-arm event rate (5.2% total; ~1.3%/y) was less than half of the anticipated 12%, reducing actual power to <40%.
Analysis: Intention-to-treat primary analysis; per-protocol sensitivity analysis censoring medical-therapy crossovers at the time of crossover. Cox proportional-hazards models generated hazard ratios, 95% CIs, and P values; for zero-event comparisons a continuity-corrected relative risk with Fisher's exact test was used.
Inclusion Criteria
- Age less than 60 years
- Patent foramen ovale documented on transesophageal echocardiography
- One of: (a) clinically and neuroradiologically verified ischemic stroke, (b) TIA with a corresponding cerebral ischemic lesion on neuroimaging, or (c) clinically and radiologically verified extracranial peripheral thromboembolic event
- No other identifiable cause of the qualifying stroke, TIA, or peripheral thromboembolism
- Written informed consent
Exclusion Criteria
- Identifiable non-PFO cause of the cryptogenic event
- Age 60 years or older
- Additional eligibility exclusions detailed in the Supplementary Appendix
Arms
| Field | PFO Closure | Control |
|---|---|---|
| Intervention | Percutaneous transcatheter closure with the Amplatzer PFO Occluder (St. Jude Medical), typically under local anesthesia with fluoroscopic guidance +/- TEE/ICE. Post-procedure: ASA 100-325 mg/d for at least 5-6 months plus ticlopidine 250-500 mg/d or clopidogrel 75-150 mg/d for 1-6 months (ticlopidine or clopidogrel alone if ASA intolerant). Endocarditis prophylaxis for 2-6 months. | Antithrombotic therapy at treating-physician discretion — antiplatelet therapy and/or oral anticoagulation, with at least one antithrombotic drug required. |
| N | 204 | 210 |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| Composite of death, nonfatal stroke, TIA, or peripheral embolism during follow-up (mean ~4 y), analyzed by intention-to-treat with Cox proportional hazards | Primary | 11/210 (5.2%) | 7/204 (3.4%) | 0.63 | 0.34 |
| Nonfatal stroke | Secondary | 5/210 (2.4%) | 1/204 (0.5%) | 0.2 | 0.14 |
| TIA | Secondary | 7/210 (3.3%) | 5/204 (2.5%) | 0.71 | 0.56 |
| Death (any cause) | Secondary | 0/210 | 2/204 (1.0%) | 5.2 | 0.24 |
| Cardiovascular death | Secondary | 0/210 | 0/204 | NA | |
| Noncardiovascular death | Secondary | 0/210 | 2/204 (1.0%) (COPD, glioma) | 5.2 | 0.24 |
| Peripheral embolism | Secondary | 0/210 | 0/204 | NA | |
| Secondary composite (stroke, TIA, peripheral embolism) | Secondary | 11/210 (5.2%) | 5/204 (2.5%) | 0.45 | 0.14 |
| Primary composite - per-protocol analysis | Secondary | — | — | 0.7 | 0.48 |
| Stroke - exploratory RESPECT definition (post hoc) | Secondary | 7 | 1 | 0.14 | 0.07 |
| New-onset atrial fibrillation | Safety | 2/210 (1.0%) | 6/204 (2.9%) | 3.15 | 0.16 |
| Bleeding (any) | Safety | 12/210 (5.7%) | 8/204 (3.9%) | 0.66 | 0.40 |
| Myocardial infarction | Safety | 1/210 (0.5%) | 2/204 (1.0%) | 2.04 | 0.62 |
| PFO-related hospital admission | Safety | 13/210 (6.2%) | 13/204 (6.4%) | 1.02 | 0.95 |
| Procedural complication (all minor) | Safety | 0/210 | 3/204 (1.5%) | 0.12 | |
| Any adverse event | Safety | 62/210 (29.5%) | 71/204 (34.8%) | 0.25 | |
| Any serious adverse event | Safety | 37/210 (17.6%) | 43/204 (21.1%) | 0.37 | |
| Any minor adverse event | Safety | 42/210 (20.0%) | 40/204 (19.6%) | 0.92 | |
| Device-associated thrombus | Safety | — | 0/204 | NA | |
| Successful device implantation | Safety | — | 188/196 attempted (95.9%) | NA | |
| Effective closure at 6 mo (TEE, n=148) | Safety | — | 142/148 (95.9%) | NA | |
| Medical-therapy crossovers to PFO closure | Safety | 28/210 (median 8.8 mo, IQR 1.2-26.4) | — | NA |
Subgroup Analysis
Prespecified subgroups analyzed for the primary composite showed statistical trends toward interaction for age (P=0.10) and for the presence versus absence of an atrial septal aneurysm (P=0.09), but no formally significant heterogeneity in effect.
Criticisms
- Substantially underpowered: observed medical-arm event rate (5.2%) was less than half of the anticipated 12%, giving <40% power to detect the pre-specified 66% relative reduction and a real risk of type II error.
- Very long recruitment window (2000-2009) reflecting slow enrollment; the resulting selected, low-risk cohort limits generalizability.
- Open-label treatment: outcome adjudication was blinded, but investigators and patients were not — with more potential events discounted in the medical-therapy arm, this could bias event ascertainment.
- Composite primary end point mixed hard outcomes (death, stroke) with softer ones (TIA), diluting the effect size seen for stroke alone (HR 0.20) into a null overall result.
- High attrition (7 withdrawals + 24 lost in closure; 11 + 31 in medical) and a 13% crossover rate from medical to closure (median 8.8 months) undermined the ITT contrast.
- TEE-based shunt grading and PFO characterization were not fully centralized (data available for 185/184 patients only).
- Antithrombotic regimens differed by design (mandated antiplatelet in the closure arm; treating-physician discretion in the medical arm), and from 12 months onward medical-arm patients received significantly more antithrombotics and oral anticoagulation, potentially blunting between-group differences.
- Amplatzer PFO Occluder is a specific device; findings do not necessarily extrapolate to other septal occluders.
Funding
Supported by St. Jude Medical (manufacturer of the Amplatzer PFO Occluder). The funder provided organizational support for event adjudication and DSMB meetings but was not involved in trial design, conduct, analysis, manuscript writing, or the decision to submit.
Based on: PC Trial (N Engl J Med, 2013)
Authors: Meier B, Kalesan B, Mattle HP, et al; PC Trial Investigators
Citation: N Engl J Med 2013;368:1083-1091. DOI: 10.1056/nejmoa1211716
Content summarized and formatted by NeuroTrials.ai.