Patent Foramen Ovale and Stroke: A Review

Authors, Journal, Affiliations, Type, DOI

• Authors: David M. Kent (Predictive Analytics and Comparative Effectiveness Center, Tufts Medical Center) and Andy Y. Wang (Department of Neurology, University of California San Francisco). Kent developed the RoPE score and led the SCOPE individual-patient-data meta-analysis that produced the PASCAL classification used throughout this review
• Journal: JAMA — Clinical Review & Education. Published online 28 July 2025; accepted 13 June 2025; CME available
• Type: Narrative review. Search of PubMed from 1 January 2014 to 17 May 2025 yielded 867 articles, of which 79 were included (27 systematic reviews/meta-analyses, 23 cohort studies, 15 RCTs, 8 expert statements, 6 guidelines)
• DOI: https://doi.org/10.1001/jama.2025.10946

Overview

The clearest single account of how PFO-associated stroke is now conceptualised, written by the investigator whose two instruments — RoPE and PASCAL — define it. Its thesis is that the question has changed. It is no longer "does closure work?" but "in this patient, was the PFO the cause?" — a Bayesian question with an answer that ranges from near-zero to about 90% depending on six clinical variables and two anatomical ones.

The numbers make the case better than the argument does. In a patient with a RoPE score under 3, PFO prevalence is 23% — the general population rate — meaning the PFO almost certainly did not cause the stroke. At RoPE 9–10, prevalence is 77% and the attributable fraction is around 90%. Closure reduces recurrent stroke by 90% in the PASCAL "probable" group (HR 0.10) and not at all in "unlikely" (HR 1.14), where it adds atrial fibrillation.

It is also the most useful source in this set on the things that follow from closure: that the stroke risk is still six times the general population's after the device is in, and that the patients who gain least are the ones the procedure hurts most.

Keywords

Patent foramen ovale · cryptogenic stroke · embolic stroke of undetermined source · paradoxical embolism · PFO closure · RoPE score · PASCAL classification

Key Takeaways

Epidemiology and the attributable fraction

• PFO is present in approximately 25% of all adults, with a mean tunnel diameter of 4.9 mm (range 1–19 mm) — most are too small to shunt meaningfully but large enough to pass a venous embolus.
• Paradoxical embolism accounts for about 5% of all ischaemic strokes and 10% of strokes in young and middle-aged adults.
• ~50% of patients aged 60 or younger with ESUS have a PFO, against 25% of the general population.
• Meta-analysis of 23 case-control studies (1,154 cases, 1,852 controls): summary OR 2.9 (2.1–4.0) for PFO in cryptogenic versus known-cause stroke; OR 5.1 (3.3–7.8) below age 55.
• Derived attributable fractions: PFO is the likely cause in 80% (75–84%) of cryptogenic strokes under 55 and 52% (34–66%) of older patients.
• Two mechanisms, not one: paradoxical embolism, and in situ thrombus formation around the interatrial tunnel — the second is why closure might work even without a demonstrable venous source.
• In 2020 an international working group recommended that PFO-associated stroke be removed from the "cryptogenic" category and given its own nomenclature.

RoPE — what the score is actually for

• RoPE was derived from 8 databases and 3,023 patients, and estimates the probability that a PFO is pathogenic rather than incidental — it is not a recurrence-risk score, and the distinction is the source of most of the confusion about it.
• PFO prevalence by score: RoPE <3 → 23% (attributable fraction ≈ 0); RoPE 9–10 → 77% (attributable fraction ≈ 90%). Validated in ASTRAL (455 patients): 23% at RoPE ≤3 to 72% at RoPE 9–10.
• The paradox that limits it: high RoPE scores predict a high chance the PFO caused this stroke, but patients with high scores have low absolute recurrence risk — 2-year stroke/TIA recurrence 2% (0–4%) at RoPE 9–10 in 1,324 patients. The absolute benefit of closure in exactly the patients most likely to have a PFO-related stroke is therefore modest.
• RoPE omits anatomy — large shunt and atrial septal aneurysm are not in it, which is why PASCAL exists.

PASCAL — combining cause and anatomy

• Large shunt is defined as >20 bubbles in the left atrium on TOE; aneurysm as ≥10 mm excursion from midline.

The trials and the heterogeneity

• The first three RCTs (CLOSURE I 909 patients, RESPECT 980, PC 914) reported no significant difference at 2, 2.5 and 4 years. Extended RESPECT at 5.9 years then found 3.6% versus 5.8% (HR 0.55, 0.31–1.00, P=.046).
• Extended RESPECT, REDUCE and CLOSE (all 2017) and DEFENSE-PFO (2018) were positive.
• Meta-analysis of 6 RCTs (3,560 patients): recurrent ischaemic stroke 2% versus 4.6%, RR 0.44 (0.23–0.85), at a median 57 months.
• SCOPE individual-patient-data meta-analysis (6 RCTs, 3,740 patients): annualised stroke 0.47% versus 1.09%, adjusted HR 0.41 (0.28–0.60).

• Large shunt plus aneurysm — both features, disregarding RoPE: 5.9% recurrence on medical therapy versus 0.4% with closure at 2 years, ARR 5.5% (2.7–8.3), against 1.0% for one or neither feature.

Harm

• AF/flutter 3.7%, vascular complications requiring surgical repair 3.0%, haematoma/haemorrhage 2.7%, tamponade/perforation 0.5%, pneumothorax/haemothorax 0.1%, death 0.3% — from claims data on 1,887 patients.
• Pooled IPD across 6 RCTs: any serious adverse event 28.7% versus 26.4% (risk difference 1.97%, −0.89 to 4.82) — no significant excess overall.
• But AF 5.0% versus 1.1% (RD 3.77%, 2.65–4.89), over half of it periprocedural and often transient. AF beyond 45 days: 2.4% versus 0.8% (RD 1.38%, 0.56–2.19). VTE 1.4% versus 0.5%. Bleeding 1.4% versus 1.7%.
• Harm also concentrates in the patients who do not benefit: post-45-day AF risk difference was 4.41% in PASCAL "unlikely" versus 0.65% in "probable".
• Rare unquantified hazards: nitinol allergy, device dislodgement, fracture, infection.

After the device is in

• Stroke risk remains elevated compared with the general population: 2.5% (1.5–4.0%) at 4 years after closure versus 0.4% (0.3–0.6%), HR 6.3 (3.1–12.6).
• Residual risk is driven by vascular risk factors — age, hypertension, diabetes, dyslipidaemia, smoking — not by the PFO. Vascular risk management does not stop at the device.

Medical therapy

• Anticoagulation versus antiplatelet therapy — meta-analysis of 6 RCTs (1,961 patients): recurrent ischaemic stroke 3.2% versus 5.5%, RR 0.59 (0.35–0.98), P=.04.
• But closure versus anticoagulation remains untested at scale. Of the 6 closure trials, 4 let investigators choose antiplatelet or warfarin, REDUCE mandated antiplatelet therapy, and only CLOSE randomised medical therapy to anticoagulant or antiplatelet. About three-quarters of medically treated patients received antiplatelet therapy.

Over 60

• Only DEFENSE-PFO randomised patients over 60 — 34 patients, 13 closed — which precludes conclusions.
• Two propensity-matched observational studies favour closure (130 matched pairs, HR 0.45; 5,508 Medicare beneficiaries, HR 0.62), both with methodological problems: no common time zero in the first (immortal time bias), and real-world claims data that cannot identify cryptogenic strokes in the second.
• The American Academy of Neurology suggests closure may be reasonable at 60–65 with high-risk features despite RoPE <7.
• CLOSE-2 (NCT05387954) is randomising patients aged 60–80.

Areas of uncertainty

• Whether periprocedural AF should be anticoagulated.
• Whether TIA patients should be closed — they were excluded from most RCTs.
• DVT is found in 7–27% of patients worked up for PFO-associated stroke, is not in any risk score, and early lower-limb Doppler (with pelvic MR venography in selected cases) is recommended by society guidelines despite no outcome data.
• Features outside the scores that may raise or lower the probability of causality: Valsalva at stroke onset, recent prolonged travel, sleep apnoea, DVT or PE at presentation, hypercoagulability, migraine, Chiari network, Eustachian valve.

Limitations of the document

• The authors state that the quality of included studies was not formally evaluated — a narrative review with a systematic search, not a systematic review.
• Relevant studies may have been missed, and the diagnostic workup of stroke varied between included studies, producing substantial heterogeneity that the pooled estimates do not resolve.
• Written by the developer of the two instruments it recommends. Kent derived RoPE and co-led SCOPE, which produced PASCAL. The review is transparent about the provenance but is not an independent appraisal of either tool, and a reader gets no counterweight view.
• PASCAL has never been prospectively validated and was not a stratification variable in any trial — the subgroup effects are derived from the same dataset that generated the classification, which is the definition of overfitting risk.
• The 2-year absolute risk reduction is 2.1% in both "possible" and "probable", with identical NNTs of 47 — which undermines the practical value of separating the two categories, a point the review notes but does not pursue.
• The older-than-60 evidence is presented sympathetically despite the review's own criticism of both studies, which leaves the reader with an impression of benefit the data do not support.
• Only one paragraph on device choice, and none on device-specific outcomes, operator volume or institutional requirements — the procedural questions an interventional reader would ask.
• No cost or equity analysis, in a review recommending a device procedure for a 2% absolute benefit.

Key Concepts Mentioned

concepts/Patent-Foramen-Ovale · concepts/PFO-Closure-Selection · concepts/PFO-Detection-and-Anatomy · concepts/Poststroke-AF-Monitoring · concepts/PFO-Non-Stroke-Indications

Key Entities Mentioned

entities/Ischemic-Stroke · entities/Atrial-Fibrillation · entities/Pulmonary-Embolism · concepts/Atrial-Septal-Defect

Named but unlinked: RoPE · PASCAL · SCOPE · RESPECT · REDUCE · CLOSE · CLOSURE I · PC Trial · DEFENSE-PFO · NAVIGATE ESUS · Amplatzer PFO Occluder · Gore Cardioform · NobleStitch · ESUS

Wiki Pages Updated

• sources/pfo-jama-2025 — created
• concepts/Patent-Foramen-Ovale — created
• concepts/PFO-Closure-Selection — created
• concepts/PFO-Detection-and-Anatomy — created
• entities/Ischemic-Stroke — ESUS and the cryptogenic category
• concepts/Poststroke-AF-Monitoring — AF detection before and after closure
• wiki/wikiindex · wiki/sourceindex