Enter your search text in the box above

Select a result to preview

#patent-foramen-ovale #cryptogenic-stroke #embolic-stroke-of-undetermined-source #pfo-closure #paradoxical-embolism #rope-score #pascal-classification #atrial-fibrillation

Patent Foramen Ovale

Contents. #Introduction · #Definition · #Key Concepts · #Detection and Anatomy · #Indication for Closure · #Indications Beyond Stroke · #Rationale and Pathophysiology · #Contradictions / Open Questions · #Connections · #Sources

Detailed pages. concepts/PFO-Detection-and-Anatomy · concepts/PFO-Closure-Selection · concepts/PFO-Non-Stroke-Indications

Introduction

A patent foramen ovale is a flap of fetal atrial septal tissue that never fused — a tunnel, not a hole — present in about a quarter of all adults and harmless in almost all of them. Its clinical importance is entirely conditional: it matters only when something crosses it. The consequence is a diagnostic problem unlike most in cardiology, because the structure is common, the disease is rare, and demonstrating that this PFO caused this stroke is a probabilistic argument rather than a measurement.

Abbreviations: ASA, atrial septal aneurysm · APT, antiplatelet therapy · c-TCD, contrast transcranial Doppler · c-TOE, contrast transoesophageal echocardiography · c-TTE, contrast transthoracic echocardiography · DCI, decompression illness · ESUS, embolic stroke of undetermined source · ILR, implantable loop recorder · OAC, oral anticoagulation · PASCAL, PFO-Associated Stroke Causal Likelihood · PFO, patent foramen ovale · POS, platypnea-orthodeoxia syndrome · RLS, right-to-left shunt · RoPE, Risk of Paradoxical Embolism

Definition

A patent foramen ovale is a persistent tunnel between the septum primum and septum secundum, allowing a right-to-left shunt when right atrial pressure transiently exceeds left. It is not a deficiency of atrial septal tissue — that is an atrial septal defect, and the distinction governs everything downstream, because a PFO needs a pressure gradient to shunt and an ASD shunts continuously.

PFO-associated stroke — the term that replaced "cryptogenic stroke with a PFO" in a 2020 nomenclature proposal — is an ischaemic stroke with a cortical, large white-matter or retinal infarct in a patient with a PFO and no other identified likely cause. The redefinition matters because it moves the PFO from a finding adjacent to the diagnosis into the diagnosis itself, and thereby makes the causal question explicit.

Key Concepts

Detection and Anatomy

Full detail: concepts/PFO-Detection-and-Anatomy.

No test is a gold standard, and the conventional reference is imperfect. Contrast transcranial Doppler is the most sensitive screen (0.96) but cannot see the septum; contrast transthoracic echocardiography is specific (0.99) but insensitive (0.71) — a fair screen and a bad rule-out; contrast transoesophageal echocardiography allows anatomical measurement but has a sensitivity of only 0.89 against autopsy, surgery or catheterisation, partly because Valsalva is difficult under sedation. ESO 2024 consequently issues no evidence-based recommendation on diagnosis at all, only consensus: local algorithms, TCD when echo is inconclusive, and TOE after a positive screen to define anatomy.

Anatomy is not decoration — it is the second axis of the treatment decision. Atrial septal aneurysm (≥10 mm excursion), a large shunt (>20 bubbles within 3 cardiac cycles at TOE, or >10 high-intensity transient signals at TCD), septal hypermobility and PFO size each predict recurrence and each raise the treatment effect.

Indication for Closure

Full detail: concepts/PFO-Closure-Selection.

Close the PFO in a patient aged 18–60 with a PFO-associated stroke, when the PFO is more likely than not to have been the cause. The decision takes two steps, and it is the second one that changes management.

Step 1 — RoPE asks whether this PFO caused this stroke. Add a point for each favourable feature, then the age score (sources/pfo-jama-2025, rating: high).

Factor Points
No hypertension 1
No diabetes 1
No prior stroke or TIA 1
Non-smoker 1
Cortical infarct on imaging 1
Age 18–29 / 30–39 / 40–49 / 50–59 / 60–69 / ≥70 5 / 4 / 3 / 2 / 1 / 0

The score converts into a probability that the PFO is pathogenic rather than incidental:

RoPE total PFO prevalence in that stratum PFO-attributable fraction
<3 23% ≈0
9–10 77% ≈90%

RoPE is a causal-probability instrument, not a recurrence-risk score, and using it as one is the commonest error made with it. A high score means the stroke was probably caused by the PFO; it does not mean another stroke is likely.

Step 2 — PASCAL adds anatomy, and this is where the decision is made. The anatomy is a large shunt (>20 bubbles in the left atrium) and/or an atrial septal aneurysm (≥10 mm excursion from midline) (sources/pfo-jama-2021, rating: very high).

High RoPE (≥7) High-risk anatomy Category HR (95% CI) 2-year ARR NNT Decision
No No Unlikely (~15%) 1.14 (0.53–2.46) −0.7% (−4.0 to 2.6) — Do not close
No Yes Possible 0.38 (0.22–0.65) 2.1% (0.6–3.6) 47 Close
Yes No Possible 0.38 (0.22–0.65) 2.1% (0.6–3.6) 47 Close
Yes Yes Probable 0.10 (0.03–0.35) 2.1% (0.9–3.4) 47 Close

What the decision buys. Annualised stroke risk on medical therapy is 1.09% against 0.47% with closure, an adjusted HR of 0.41 (0.28–0.60) and a 2-year ARR of 1.72% — an NNT of about 47, comparable to antihypertensives, antiplatelets and statins in secondary prevention. When a large shunt and an aneurysm coexist the ARR rises to 5.5% (2.7–8.3) and the NNT falls to 14–18; morphological selection alone produced no strokes in the closure arm of DEFENSE-PFO (sources/pfo-jama-2021; sources/pfo-jacc-2018).

The edges of the evidence. Both guidelines agree on closure plus antiplatelet therapy within 18–60, and both concede there is nothing behind the boundaries: 34 over-60 patients randomised in one trial, one patient under 18 in another, leaving the group with the highest stroke incidence — and the highest competing risk — managed by consensus. No trial has compared closure with a direct oral anticoagulant, and every trial enrolled within months of the index stroke, so timing rests on consensus rather than data (sources/pfo-eso-2024, rating: very high; sources/pfo-jscai-2022).

Indications Beyond Stroke

Full detail: concepts/PFO-Non-Stroke-Indications.

Eight scenarios have no trial evidence behind them. One is for closure — platypnea-orthodeoxia, on a 14-point rise in oxygen saturation; one by analogy — systemic embolism, on three patients; six against — migraine, decompression illness, thrombophilia, isolated atrial septal aneurysm, TIA and DVT. The migraine verdict is the counter-intuitive one: three randomised trials failed their primary endpoints, a post-hoc pooled analysis found a nine-fold rise in complete cessation, and SCAI recommends against routine closure while conceding that refractory patients may reasonably choose it. Every recommendation is conditional, and all but migraine rest on very low certainty (sources/pfo-jscai-2022, rating: very high).

Rationale and Pathophysiology

Why a small hole causes a large stroke. A PFO is a filter bypass. The pulmonary capillary bed normally removes small venous thrombi; a shunt of a few millimetres is wide enough to pass an embolus large enough to occlude a cerebral branch. The brain is the organ that notices, both because the anatomical path reaches it first — emboli must pass the cerebral circulation to reach the renal, splenic or peripheral arteries — and because neural tissue tolerates brief ischaemia worse than any other.

Why it is intermittent. The tunnel is held closed by left atrial pressure exceeding right. Anything that reverses that gradient — Valsalva, cough, defecation, lifting, a breath-hold dive — opens it. This explains the clinical features that sit outside the scores: Valsalva at stroke onset, recent prolonged immobility or air travel, obstructive sleep apnoea, and the striking association with decompression illness in divers, where the shunt allows nitrogen bubbles to bypass the lung.

Why the effect size varies so much between patients. The probability that a given PFO caused a given stroke is a product of two things: how likely a paradoxical embolus was in that patient (young, no vascular risk factors, cortical infarct, no atrial fibrillation) and how permissive the anatomy is (shunt size, aneurysm, tunnel length). RoPE measures the first, PASCAL adds the second. When both are favourable the treatment effect is large; when the first is unfavourable the "PFO-associated" stroke was probably due to occult AF or small-vessel disease, and closing the PFO removes nothing while adding a device that provokes AF.

Why the harm concentrates where the benefit does not. Older patients with vascular risk factors have both a lower attributable fraction and a higher risk of procedure- and device-related AF. The same variables that make a PFO unlikely to be causal make the atrium more likely to fibrillate after an occluder is sewn into it. This is not a paradox but a consequence: the two outcomes share a substrate. That is the strongest argument in the literature for using a causal-likelihood instrument before, rather than after, offering the procedure.

Contradictions / Open Questions

Connections

Sources

Guidelines

The pooled primary evidence

Reviews and trials