Cryptogenic Stroke: Hidden Causes, PFO, and AFib

Cryptogenic stroke is an ischemic stroke whose cause remains unknown after a standard medical workup. It accounts for roughly one in four ischemic strokes, making it one of the most common and frustrating diagnoses in stroke medicine. The label sounds definitive, but it is really an admission that available tests came up short. In many cases the cause is there, just hiding below the reach of routine screening, and a growing body of research is reshaping how clinicians search for it and what they do when the search stalls.

What Makes a Stroke “Cryptogenic”

Ischemic strokes have traditionally been sorted into buckets: large-artery disease, small-vessel disease, cardioembolism (a clot thrown from the heart), or a known unusual cause like a blood disorder. The widely used TOAST classification system places any stroke that does not fit neatly into one of those categories, or that fits into more than one, into the cryptogenic bin. The definition served its purpose for clinical recordkeeping, but it was too vague to build treatment trials around. In 2014, researchers introduced a narrower concept called embolic stroke of undetermined source, or ESUS, which targets patients whose brain imaging pattern looks embolic (a clot that traveled from somewhere else) after large-artery disease, small-vessel disease, and major cardioembolic sources have been ruled out.¹ ESUS was designed to create a workable patient population for randomized trials of blood thinners, and it has since become one of the most studied categories in stroke research.

The distinction matters because “cryptogenic” includes everything from strokes that were almost certainly embolic to strokes with incomplete testing and strokes where two competing causes muddied the picture. ESUS carves out the subset most likely caused by a traveling clot, which focuses the search on potential embolic sources: irregular heart rhythms, holes in the heart, arterial plaques, or abnormal blood clotting.

Atrial Fibrillation That Stays Hidden

The single most hunted culprit behind cryptogenic stroke is atrial fibrillation (AF), an irregular heart rhythm that lets blood pool and form clots in the heart’s upper chambers. AF can be brief and intermittent, producing no symptoms and escaping detection during a standard hospital stay. When patients with cryptogenic stroke are fitted with insertable cardiac monitors (small devices placed under the skin of the chest), AF is detected in about one in five within two years, with a median time to detection of roughly four months.² That number alone shows how much AF is missed by brief monitoring.

Finding AF matters because it changes treatment. Patients with known AF benefit from anticoagulant drugs rather than simple antiplatelet therapy like aspirin. But a large study following 840 cryptogenic stroke patients with implanted monitors found that among the 112 who went on to have another stroke, about 70% had never had AF detected at any point.³ That statistic is sobering: even with prolonged monitoring, most recurrent strokes in this population occur in patients who never show AF. It points to a reality that many neurologists and cardiologists have come to accept: AF is not the only cardiac explanation for these strokes.

Atrial Cardiopathy Without Fibrillation

A concept gaining traction is atrial cardiopathy, a diseased state of the heart’s upper chambers that can promote clot formation even when the heart rhythm looks normal on a monitor. Structural changes like atrial enlargement, scarring (fibrosis), and chronic low-grade inflammation can create a sluggish, clot-friendly environment independent of AF episodes.⁴ This helps explain why some patients with cryptogenic stroke have elevated heart-stress biomarkers and enlarged atria on imaging but never develop detectable AF. The atrial tissue itself is the problem, and AF, when it eventually appears, may be a symptom of the same underlying disease rather than the sole mechanism.

Patent Foramen Ovale

Nearly half of patients with cryptogenic stroke turn out to have a patent foramen ovale (PFO), a small flap-like opening between the two upper chambers of the heart that normally closes shortly after birth but persists in roughly a quarter of the general population.⁵ In most people it causes no trouble, but in certain circumstances a venous blood clot can slip through the opening and travel to the brain, a process called paradoxical embolism.

For years, randomized trials of PFO closure failed to show a clear benefit over medical therapy alone, and the debate raged. Longer follow-up from the RESPECT trial and two newer trials, CLOSE and REDUCE, changed the picture. In the CLOSE trial, ischemic stroke occurred in about 1.4% of patients who had their PFO closed versus roughly 5.4% in the group receiving antiplatelet drugs alone over a median follow-up of just over three years.⁶ Rates of serious adverse events were similar between the groups, though new-onset AF after the closure procedure occurred in about 7% of patients, usually transient. The evidence now supports PFO closure in carefully selected patients, particularly younger individuals with larger shunts and features like an atrial septal aneurysm that raise the likelihood the PFO was actually involved in the stroke.⁷

Non-Stenotic Arterial Plaques

Conventional stroke workups look for arteries with severe narrowing, typically 50% or more. Plaques that do not reach that threshold tend to be dismissed. Yet emerging evidence shows that non-stenotic carotid plaques, fatty buildups on the carotid artery walls that cause less than 50% narrowing, may account for a sizable share of cryptogenic strokes. One estimate puts the figure at 20% to 30% of unilateral embolic strokes of undetermined source.⁸

The evidence for causation comes from a consistent pattern: high-risk plaque features are far more common on the side feeding the stroke territory than on the opposite side. In one imaging study, plaques five millimeters thick or larger were present on the same side as the stroke in 11% of patients and on the other side in only 1%.⁹ That lopsided distribution is hard to explain by chance. Other research has found that patients with these ipsilateral plaques are less likely to have AF detected on follow-up, suggesting the plaque itself, not a hidden heart rhythm, was the embolic source.¹⁰

The same principle extends inside the skull. High-resolution vessel-wall MRI can reveal intracranial plaques that standard imaging misses entirely because they do not narrow the artery enough to change blood flow. Research supports an etiologic role for these non-stenotic intracranial plaques in ESUS, though this type of imaging is not yet part of routine clinical practice.¹¹

Aortic Arch Plaques

The aorta’s arch, the curved section where the aorta leaves the heart and directs blood toward the brain, can also harbor large plaques. In patients with cryptogenic stroke specifically, large aortic arch plaques (four millimeters thick or more) have been linked to a dramatically higher risk of recurrent events. One study found an adjusted hazard ratio above 6 for large plaques and above 9 for complex plaques with ulceration or mobile components among cryptogenic stroke patients.¹² Microembolic signals, detected by transcranial Doppler, have been recorded downstream from thick aortic plaques, providing direct evidence of their embolic potential.¹³

A more recent study, however, found that the link between large arch plaques and ischemic stroke specifically faded after adjusting for other cardiovascular risk factors, suggesting these plaques may function more as a marker of widespread atherosclerosis than as a direct embolic launcher.¹⁴ The clinical implication: finding a big aortic arch plaque does not necessarily solve the mystery of a cryptogenic stroke, but it does flag a patient at high cardiovascular risk who needs aggressive risk-factor management.

Cancer and Blood-Clotting Disorders

Sometimes a cryptogenic stroke is the first sign of an occult cancer. Tumors, especially mucin-producing cancers of the gastrointestinal tract, pancreas, and lung, can trigger a hypercoagulable state sometimes called Trousseau’s syndrome. The resulting widespread clot formation can send emboli to the brain well before the cancer itself produces obvious symptoms.¹⁵

A panel of blood-clotting and hemostatic markers, sometimes grouped under the acronym MOCHA, can help identify these patients early. In one study of ESUS patients, those with abnormal MOCHA levels were significantly more likely during follow-up to be diagnosed with cancer (21% versus 0% among those with normal levels), venous blood clots, or other hypercoagulable states.¹⁶ A cryptogenic stroke with markedly elevated D-dimer or other clotting markers, especially in an older patient, should prompt a careful search for underlying malignancy.

Blood Biomarkers for Predicting Hidden AF

Not every cryptogenic stroke patient can or wants to have a cardiac monitor implanted. Blood biomarkers offer a way to triage who should get prolonged monitoring. NT-proBNP, a protein released when the heart is under stress, stands out as the most consistent predictor. A meta-analysis of nine studies found that elevated NT-proBNP was associated with roughly a threefold increase in the odds of AF being detected, and low levels had a pooled negative predictive value above 90%, meaning they were good at ruling AF out.¹⁷ Other cardiac biomarkers like BNP and cardiac troponins also tend to be higher in patients who later prove to have AF, though NT-proBNP has the most robust data behind it.¹⁸

In practice, a low NT-proBNP after cryptogenic stroke might steer the diagnostic search away from the heart and toward vascular or hematologic explanations, while a high level would strengthen the case for immediate prolonged cardiac monitoring.¹⁹

Why Blood Thinners Have Not Replaced Aspirin for ESUS

If most ESUS strokes are embolic, it seems logical that anticoagulant drugs, which are more potent clot preventers than aspirin, would be better at preventing recurrence. Several large trials tested this idea head-to-head, comparing direct oral anticoagulants (DOACs) like rivaroxaban and dabigatran to aspirin. The results were disappointing. A comprehensive meta-analysis found no statistically significant reduction in recurrent stroke with DOACs compared to aspirin, with a trend toward benefit that did not reach significance.²⁰ An updated meta-analysis confirmed those findings and added that DOACs were associated with a significantly higher risk of clinically relevant non-major bleeding.²¹

A smaller trial of apixaban versus aspirin was stopped early for futility, finding no meaningful difference in new brain lesions between the two groups.²² The likely explanation for these flat results is that ESUS lumps together patients with very different underlying problems. Anticoagulants would logically help someone whose stroke was caused by hidden AF or a cardiac clot, but they would do little for a stroke caused by an arterial plaque rupture or a paradoxical embolism through a PFO. When you average together those different biologies, the benefit in one subgroup gets diluted by the lack of benefit, or even harm, in others. The consensus now is that the right treatment depends on identifying the individual mechanism, not on treating all ESUS patients the same way.

Cryptogenic Stroke in Younger Adults

Stroke in people under 50 is less common but disproportionately cryptogenic. Up to half of ischemic strokes in younger adults end up in the cryptogenic or poorly explained category, often linked to PFO, clotting disorders, or conditions with uncertain causal relationships.²³ The recurrence risk in this group tends to be lower than in older patients: one long-term study found an average annual recurrence rate of around 1% to 1.6%, regardless of PFO status, and more than 60% of young patients achieved a good functional outcome.²⁴

When researchers tried to match young cryptogenic stroke patients to known risk-factor profiles for either vascular or cardiac stroke, about half fit a vascular pattern, a tiny fraction fit a cardiac pattern, and roughly 45% did not match either profile at all.²⁵ That large unclassifiable group underscores how much remains unknown about stroke mechanisms in younger people, and why ongoing research into triggers like migraine with aura, hormonal contraception, and genetic clotting tendencies remains important.

Long-Term Outlook

A population-based study tracking cryptogenic stroke patients over a decade found that their outcomes were broadly similar to those of patients with known non-cardioembolic strokes. About 23% were dead or dependent at six months, and the 10-year risk of recurrence was around 32%, not statistically different from the rates seen in large-artery or small-vessel stroke.²⁶ The cryptogenic label does not mean a worse prognosis than other stroke subtypes. It does mean a nagging uncertainty about what to target for prevention, which is why the diagnostic workup matters so much.

The Economics of Prolonged Heart Monitoring

Insertable cardiac monitors are not cheap, and they require a minor implant procedure. Economic analyses have consistently found them cost-effective for cryptogenic stroke patients nonetheless. One UK-based model estimated that monitoring was associated with fewer recurrent strokes and more quality-adjusted life years compared to standard care, at an incremental cost well below established willingness-to-pay thresholds.²⁷ A US analysis found that implanting the monitor immediately after stroke was cost-effective compared to standard care and actually cost-saving compared to a strategy of waiting and implanting only after a Holter monitor failed to detect AF, because the delayed approach missed AF episodes during the waiting period and led to preventable recurrent strokes.²⁸ The upfront cost of the device tends to be offset by the downstream savings from preventing even one recurrent stroke, which carries enormous treatment and disability costs.

Psychological Impact and Quality of Life After PFO Closure

Living with the knowledge that your stroke has no clear cause can be psychologically taxing. Depression and anxiety are common after any stroke, but the uncertainty of a cryptogenic diagnosis can amplify those feelings. Research comparing cryptogenic stroke patients who underwent PFO closure with those who did not found that the closure group reported better functioning and substantially lower rates of depression and anxiety.²⁹ The strongest predictor of psychological well-being was overall functioning level, not the procedure itself, but it is plausible that closing a known structural abnormality and reducing recurrence risk gives patients a sense of resolution that medical therapy alone does not. Female sex, older age, lower education, and the presence of other health problems were the main drivers of higher depression and anxiety scores, a reminder that psychological recovery after cryptogenic stroke depends on many factors beyond the stroke itself.