Sudden Vision Loss in One Eye

Vision · §36

Vision goes out in one eye. No pain, no warning, sometimes back in seconds. This is a stroke until proven otherwise — same biology, same time pressure, same emergency department. The eye doctor is the wrong call; 911 is the right one. What follows is why, the window you have, and how to tell apart the four sudden vision losses an adult is most likely to face.

Respond · As-needed Evidence Moderate Chapter Vision

The eye that suddenly stops seeing is sharing blood vessels with the brain, and the clot that just blinded it is sitting one branch upstream from a clot that could disable you next week. Catch it inside roughly four hours and you may save the eye; catch it inside the next two weeks and you almost certainly prevent the stroke that would have followed. The cost is one ambulance ride. The cost of waiting it out is whichever of those two you miss.

The artery that feeds the retina is a direct branch of the artery that feeds the brain. A clot small enough to lodge in the eye is a clot that broke off from a plaque in your neck, or your heart, or your carotid — the same source that throws clots into the brain. The eye is just the first place the clot happened to stop. Cardiologists and stroke neurologists now treat sudden vision loss in one eye as an acute stroke that happens to involve the retina rather than an "eye problem" Mac Grory et al. 2021.

The retina starves fast. In monkey experiments where the central retinal artery was clamped and released at different times, the retina survived complete blockage of less than about 100 minutes with no visible damage. By 105 minutes, damage was permanent. By four hours, the inner retina and optic nerve were destroyed Hayreh et al. 1980. Humans are usually given a working window of about four hours from symptom onset; past six hours, the eye is gone.

The transient version — vision in one eye blackens for seconds to minutes, then comes back — is called amaurosis fugax. People describe it less often as a curtain coming down than as a grey cloud, a fog, or something pulled over the field of view — the classic curtain shows up in only about one patient in four Biousse et al. 2018. Either way it means a small clot transiently lodged in a retinal vessel and washed through. The clot is gone; the source that made it is not. About one in three people who have amaurosis fugax have a carotid artery in their neck that is more than 75% blocked Biousse et al. 2018.

What happens in the next thirty days

Two clocks start the moment the vision goes. The first is the retinal clock, measured in hours: somewhere between one and four hours of clamped blood flow, the back of the eye dies. The second is the brain clock, measured in days: the same source that fired a clot into the eye is most likely to fire another, and the second one usually goes upstream into the brain.

The brain clock is the bigger threat. In a Korean nationwide cohort of patients who had a central retinal artery clot, the rate of cerebral stroke in the first seven days afterward was 44 times the background rate, and stayed elevated through day 90 Park et al. 2015. Roughly three in ten patients who show up with a fresh retinal artery clot already have signs of a fresh brain stroke on MRI when they walk in the door — they just haven't noticed it yet because the part of the brain that was hit happens to be silent Fallico et al. 2020. The reader who sees the vision come back and decides to wait until Monday morning is sitting on those odds.

The transient version carries the same warning. About one in twenty people who have an episode of amaurosis fugax will have a full stroke within the following year, with most of that risk concentrated in the first two weeks Mac Grory et al. 2021. The carotid plaque that fired the first clot is still there. The longer it stays untreated, the more chances it gets.

What to do, right now

The decision rule is small enough to memorize. Sudden painless loss of vision in one eye, in an adult — even if it has already come back — is a 911 call. Not the family doctor. Not the eye doctor. Not "I'll see how it is tomorrow." When you call, the words to use are "I think I'm having a stroke and it's affecting my eye" — that phrasing routes you to a stroke alert. "I can't see out of one eye" can route you to ophthalmology, which is the slow path.

At the hospital, expect a full stroke workup: a brain CT or MRI to see what's happened in the brain, scans of the arteries in your neck and head to find the source of the clot, a heart-rhythm monitor to check for atrial fibrillation, an echocardiogram of the heart, and blood work. If you're over fifty, the blood work will include inflammation markers to screen for the one form of sudden vision loss that is treated with steroids rather than a clot-buster — see the next two sections Mac Grory et al. 2021.

What the data say about the eye, and what it says about the brain

The framework here changed in 2021. Until then, sudden vision loss was an ophthalmology problem; clot-busting medication was almost never given because no one had run the trial. Now the American Heart Association explicitly recommends managing it the same way the brain version is managed — clot-buster on the table for the 4.5-hour window, secondary stroke prevention as the larger and more reliable payoff Mac Grory et al. 2021.

The vision-saving piece of the evidence is honest but partial. Pooling every case series where someone gave intravenous clot-buster for an acute retinal artery clot, about half of patients treated inside the 4.5-hour window had measurable visual recovery; patients treated after the window had essentially none Mac Grory et al. 2020. Three randomized trials are running or recently completed; results so far are mixed. The case is strong enough that the AHA tells stroke centers to consider it, weak enough that the same statement flags the evidence as preliminary.

The stroke-prevention piece of the evidence is harder to argue with. The cohort numbers cited above — a 44-times-baseline stroke rate in the first week after a retinal artery clot Park et al. 2015, three in ten patients with a fresh brain stroke already on MRI when they arrive Fallico et al. 2020 — are the load-bearing reason to go to the hospital. Once the workup finds the carotid plaque or the heart-rhythm problem that's firing clots, the toolkit that prevents the next stroke is well-established: clearing the carotid surgically when it is more than 70% blocked, starting blood-thinners if the heart is in atrial fibrillation, antiplatelet and statin therapy, blood-pressure control. None of this is novel; what's new is the recognition that someone whose vision blacked out for two minutes last Tuesday qualifies for it.

Three things people get wrong

"It came back, so I'm fine." No. The vision coming back means the clot dissolved. It does not mean the source that made the clot dissolved. Stroke risk is highest in the first 48 to 72 hours after a transient event; that's the window where the prevention workup actually prevents something Mac Grory et al. 2021.

"I should see the eye doctor first." This is the single most common reason people miss the treatment window. The eye doctor — optometrist or ophthalmologist — will diagnose it correctly and send you to the emergency room. The detour costs hours to a day. A cohort at a major academic center found that even in 2017 through 2020, with awareness rising, the median time from vision loss to arrival at a stroke-capable hospital was 48 hours — ten times longer than the treatment window allows Biousse et al. 2018. The fix is to skip the eye doctor and route directly to the emergency room.

"The clot-buster works for the eye like it does for the brain." Not yet proven. The signal is real and the AHA endorses considering it inside the window, but the evidence is observational, not randomized. The more durable reason to go to the hospital is the stroke prevention workup, not the eye-saving medication. Even if the clot-buster doesn't restore vision in your particular case, the workup that the hospital triggers very likely saves you from the brain stroke that was coming next.

When it's something else — and the something-else is still urgent

Most sudden monocular vision loss in older adults is the clot story above. A handful of other things look similar at first and are not clots. The decision rule does not change — go to the hospital — but knowing which of these you might have changes what the hospital looks for, and in two cases changes how fast they need to move.

Sudden vision loss in both eyes at once is a different problem with a different workup — possibly a stroke in the back of the brain rather than the eye, possibly something toxic — and is out of scope for this entry. Same rule: 911, stroke-capable hospital.

Where this falls apart in practice

Three patterns account for almost every missed case.

The vision came back, so the patient stayed home. The most common failure. By the time the second event happens — usually a brain stroke rather than a second amaurosis episode — the prevention window is closed.
The patient called the eye doctor first. The eye doctor's office triages by the patient's words; "I lost vision in one eye" sounds like a same-week appointment, not a 911 call. By the time the patient is seen, examined, and referred, the treatment window for the eye is gone and the highest-risk days for the brain stroke are passed.
The ED triaged it as a vision complaint, not a stroke. Mitigated by naming the symptom as a possible stroke at the front desk. "I think I'm having a stroke" routes to the stroke alert protocol; "I can't see out of my left eye" sometimes routes to a general queue.

What changes if you respond fast

Inside the first hour, the conversation is about saving the eye. A hospital that gives the clot-buster medication inside the 4.5-hour window measures meaningful visual recovery in about half of the patients who get it — vague shapes becoming readable letters, the difference between counting fingers and reading a menu Mac Grory et al. 2020. Outside the window, that number is close to zero Mac Grory et al. 2020. The eye is the time-pressured part of the payoff and the part that fades the most if you wait.

By the next morning, the conversation has moved on from the eye and is about the brain. The MRI shows whether you've already had a small brain stroke you didn't notice. The carotid scan finds the plaque that fired the clot, and a surgical or stent procedure within two weeks of the event prevents most of the strokes that would otherwise have followed. The heart-rhythm monitor catches the atrial fibrillation that you didn't know you had, and a blood thinner cuts that stroke risk by roughly two-thirds. A statin starts. Blood pressure gets tightened up. This is the larger payoff and the one that holds up no matter what happens with the eye.

A year out, the eye is whatever the eye is — sometimes fully recovered, often not, in older patients more often not than yes. But the stroke that would have happened in the first month after the warning shot didn't happen. Your partner is not the one bathing a stroke survivor; you are the one mowing the lawn. The reader who treated a transient vision loss as a stroke and showed up at the hospital that night is, statistically, the reader who is still walking around with a working left hand a decade later.

Related things worth knowing

This entry covers sudden vision loss in one eye. Adjacent topics a reader looking at this may also want to know about:

Sudden vision loss in both eyes at the same time. A different differential — most often a stroke in the back of the brain, sometimes a vasculitis taking both eyes, sometimes toxic. Same hospital, different workup.
Gradual loss of vision over weeks to months. Cataract, glaucoma, macular degeneration, a compressing mass behind the eye. Not an emergency, but worth following up on its own track.
Headache with vision changes. Could be migraine with aura, could be giant cell arteritis in an older adult, could be a mass effect. The two latter are urgent.
Carotid screening before any symptoms. Different question. Catalogue entry on its own; the routine recommendation for asymptomatic adults is currently against screening.
Sleep apnea and the optic nerve. Untreated apnea increases risk of the ischemic-optic-nerve variant of sudden vision loss; treating apnea reduces that risk.

Related in the handbook

Worsens / aggravates

— A clot thrown from atrial fibrillation can lodge in the eye's artery and black out your vision — the same mechanism as a brain stroke.

— The clot that blinded one eye usually broke off upstream, and high pressure is a top reason it formed. Get the pressure controlled.
— Flashes, floaters, or a curtain point to the retina, not a stroke — knowing which sudden vision change you have decides whether you call 911 or an eye doctor.
— Gradual bent or missing lines on the grid is a slow macular problem — very different from the sudden total blackout that means call an ambulance.
— The clot that blacked out your eye broke off a plaque. ApoB and Lp(a) are the lipid numbers driving that plaque.
— Like sudden vision loss, sudden hearing loss in one ear is a same-day emergency — the window is short.

Substance + claimed effects

Sudden monocular vision loss in an adult — painless dimming, greying, or blackout in one eye, transient (amaurosis fugax, seconds to minutes) or persistent (minutes to permanent) — is the ophthalmic presentation of acute retinal arterial ischemia and, in older adults, a cluster of vascular and inflammatory emergencies that share final common pathways with cerebral stroke Mac Grory et al. 2021. The dominant differential is central or branch retinal artery occlusion (CRAO / BRAO), now formally classified as a stroke equivalent by the American Heart Association Mac Grory et al. 2021. Other time-critical causes that this entry covers as the framing differential: giant cell arteritis (GCA) with arteritic anterior ischemic optic neuropathy in patients over 50; non-arteritic anterior ischemic optic neuropathy (NAION); retinal detachment and vitreous hemorrhage; optic neuritis in younger adults. The entry's substance is the symptom and the response — recognition, time-to-ED, secondary stroke workup. Claimed consequences: (1) vision preservation when reperfusion is achieved inside the retinal tolerance window, (2) prevention of subsequent ischemic stroke and myocardial infarction via the secondary-prevention workup the ED triggers, (3) a longevity benefit that depends entirely on actually getting the workup done.

Evidence by addressing question

Mechanism

The central retinal artery is a branch of the ophthalmic artery, itself the first intracranial branch of the internal carotid. Embolic or thrombotic occlusion of the CRA — most often by carotid plaque debris, cardioembolic source, or in situ thrombosis — interrupts inner-retinal perfusion. The retinal ganglion-cell layer is among the most metabolically demanding tissues in the body and tolerates ischemia poorly. Hayreh's seminal rhesus-monkey experiments established the operational tolerance window: complete CRAO of under ~97 minutes produced no permanent retinal damage; about 105 minutes produced irreversible damage; 240 minutes caused massive irreversible neuroretinal and optic-disc damage Hayreh et al. 1980. The human tolerance window is conventionally extrapolated to ~4 hours, with vanishing recovery probability beyond ~6 hours. The mechanism is identical to cerebral large-vessel occlusion stroke — embolic or thrombotic interruption of arterial flow to neural tissue — which is why the AHA scientific statement reclassified CRAO from "ophthalmic emergency" to "acute ischemic stroke of the eye" Mac Grory et al. 2021.

Amaurosis fugax — the transient form — is the ophthalmic analog of a cerebral TIA: a microembolus from carotid plaque transiently lodges in a retinal arteriole, then washes through. The "curtain descending" description is classical but uncommon; only about a quarter of patients describe a curtain, with most reporting blur, cloud, or fog instead Biousse et al. 2018. About one-third of amaurosis-fugax patients have ≥75% ipsilateral carotid stenosis on imaging Biousse et al. 2018.

NAION operates through a different mechanism: hypoperfusion of the short posterior ciliary arteries supplying the anterior optic-nerve head, typically in a structurally crowded ("disc-at-risk") optic nerve, often precipitated by nocturnal hypotension. Vision loss is classically present on awakening, painless, with an altitudinal field defect (most often inferior). NAION is not embolic and not a stroke equivalent; thrombolysis is not indicated.

Arteritic AION from GCA is a vasculitic occlusion of the same posterior ciliary arteries by an inflammatory infiltrate, occurring almost exclusively in patients over 50. Untreated, fellow-eye involvement follows the first eye in roughly half of cases within days to weeks. Steroids are vision-preserving only when started early.

Evidence

Subsequent stroke risk after CRAO is the single best-replicated finding in this literature. A Korean nationwide cohort (Park et al.) found the incidence rate ratio for stroke peaked at 44.5 in the first 1–7 days after CRAO (95% CI 27.07–73.20), with elevated rates extending out to 90 days Park et al. 2015. A 2020 systematic review and meta-analysis of 15 studies found pooled prevalence of acute cerebral ischemia on MRI of ~30% in patients presenting with acute CRAO — the equivalent of a high-risk TIA Fallico et al. 2020. CRAO and minor stroke / high-risk TIA carry comparable subsequent stroke risk Mac Grory et al. 2021.

For amaurosis fugax specifically, the post-event stroke rate runs 5–10% in the subsequent year, with much of that risk concentrated in the first weeks; trial-derived estimates put 90-day stroke risk at ~6% when grouped with other TIA presentations Mac Grory et al. 2021. Among amaurosis-fugax patients found to have ≥70% ipsilateral carotid stenosis, urgent carotid endarterectomy within two weeks of the event reduces five-year stroke risk substantially (NASCET-era data, generalized in current AHA secondary-prevention guidance) Mac Grory et al. 2021.

The vision-preservation evidence is weaker and more contested. No randomized trial has shown a treatment that meaningfully improves visual acuity in CRAO. Mac Grory's patient-level meta-analysis of intravenous fibrinolysis found that treatment within 4.5 hours was associated with visual recovery in ~50% of patients, versus essentially no recovery beyond the window; thrombolysis had no measurable benefit when given after 4.5 hours Mac Grory et al. 2020. The natural-history baseline is poor: in Hayreh's 244-patient case series, fewer than 20% of non-arteritic CRAO patients regained functional vision (≥20/200) spontaneously, with the recovery that did occur happening mostly in the first 7 days Hayreh & Zimmerman 2005.

For GCA, the dominant evidence base is Hayreh's prospective cohort of 170 biopsy-positive patients: amaurosis fugax preceded permanent vision loss in 31% of patients overall, 39% of arteritic AION cases, and 15% of GCA-related CRAO cases, with permanent loss following the transient symptom by a median of ~8.5 days. Patients started on systemic steroids within 24 hours of visual symptom onset experienced improvement in 58% of cases; with delayed treatment, only 6% improved Hayreh et al. 1998.

For optic neuritis in young adults, the Optic Neuritis Treatment Trial established the natural history: median time to maximum vision loss ~4–7 days, spontaneous recovery to ≥20/40 in ~95% by 6 months, IV methylprednisolone speeding recovery but not changing final visual acuity Beck et al. 1992.

Protocol

The AHA 2021 statement is unambiguous: any patient with acute monocular vision loss within the prior 4.5 hours should be triaged as a stroke, not as an ophthalmic outpatient case. Specifically: bypass community optometry / ophthalmology and go directly to a stroke-capable emergency department; expect a CODE STROKE workup including non-contrast head CT, CT or MR angiography of the head and neck, MRI brain with DWI, telemetry / 12-lead ECG, transthoracic echocardiogram, lipid panel, HbA1c, ESR and CRP (to screen GCA in patients over 50) Mac Grory et al. 2021. Intravenous tissue plasminogen activator may be considered within the 4.5-hour window for non-arteritic CRAO at centers with stroke / neuro-ophthalmology expertise, with explicit acknowledgement that the evidence is observational and not RCT-grade Mac Grory et al. 2020 Mac Grory et al. 2021. For amaurosis fugax with resolved symptoms, the same stroke-protocol workup applies because the embolic source — and the stroke risk — has not gone away.

For suspected GCA in any patient over 50 with sudden vision loss, headache, jaw claudication, or polymyalgia symptoms, high-dose corticosteroids (oral prednisone 1 mg/kg/day, or IV methylprednisolone 1 g/day × 3 days if vision-threatening) should be started immediately on suspicion — biopsy can wait for several days without losing diagnostic yield, but a delayed steroid start loses fellow-eye vision Hayreh et al. 1998.

Contraindications

Not all sudden monocular vision loss is CRAO. The key red-flag features that change the differential:

Headache, scalp tenderness, jaw claudication, age >50 → GCA. Different mechanism, different drug (steroids, not tPA), same urgency. ESR/CRP must be drawn before steroids.
Sudden burst of new floaters, flashes of light, "curtain" descending → retinal detachment or vitreous hemorrhage. Different mechanism (mechanical separation, not ischemic), different treatment (surgical), still urgent but not stroke-protocol urgent. The "curtain" description here is the genuine peripheral-to-central darkening of a progressing detachment, distinct from the embolic curtain of amaurosis fugax.
Pain on eye movement, age <45, female → optic neuritis. Different mechanism (demyelinating inflammation), different treatment (high-dose IV steroids reduce time to recovery but not final acuity per ONTT), much better prognosis Beck et al. 1992. Workup orients toward MS rather than stroke.
Bilateral simultaneous loss → not in scope. Implies cortical (occipital) stroke, GCA with bilateral involvement, toxic / metabolic insult, or hysterical / functional vision loss. Different workup, different differential.

The protocol contraindications for CRAO thrombolysis are the same as for cerebral stroke tPA: recent surgery, active bleeding, intracranial hemorrhage on imaging, recent stroke, uncontrolled hypertension, anticoagulation outside therapeutic range.

Misconceptions

Three dominant misconceptions in lay readers and in clinical practice both:

"It came back, so I'm fine." Amaurosis fugax that resolves spontaneously is treated as a completed ophthalmic TIA. Stroke risk is highest in the first 48–72 hours after a transient embolic event Mac Grory et al. 2021. Resolution of the symptom does not resolve the embolic source.
"Go to the eye doctor first." This is the single largest source of preventable delay in CRAO management. Biousse's cohort documented median time-to-presentation of 144 hours in 2010–2013, improving to 48 hours in 2017–2020 as awareness grew — both numbers far outside the 4.5-hour treatment window Biousse et al. 2018. Routing through community optometry adds an average half-day to a one-day delay because optometrists must refer to ophthalmology, who then refer to the stroke service.
"Thrombolysis works for eye stroke like it does for brain stroke." Observational data support a benefit signal in the 4.5-hour window Mac Grory et al. 2020, but no RCT has confirmed it. The AHA statement is explicit that the evidence remains preliminary; the stronger reason to go to the ED is the secondary stroke-prevention workup, not the eye-saving tPA Mac Grory et al. 2021.

Stakes

The downside of ignoring or delaying the response is large in two distinct directions. Direction one: a missed treatment window for vision. With Hayreh's retinal tolerance window at ~100 minutes for full preservation and ~240 minutes for any preservation, the patient who waits to see if vision returns has typically blown the window by the time they call Hayreh et al. 1980. Spontaneous recovery to functional vision occurs in <20% of untreated CRAO Hayreh & Zimmerman 2005. Direction two: a missed stroke workup. With IRR for cerebral stroke peaking at ~44 in the first week after CRAO Park et al. 2015 and ~30% prevalence of silent or symptomatic cerebral infarction already present on MRI at presentation Fallico et al. 2020, the patient who treats CRAO as "just" an eye problem is sitting on a high-risk embolic source for the days when most of the strokes happen. In GCA, the cost of delay is the fellow eye — fellow-eye involvement in untreated GCA approaches 50% within days Hayreh et al. 1998.

Practicalities

Practical reality of the response: the symptom typically presents without pain, often without warning, and with no other systemic clue. The reader must trust the symptom alone. The decision rule for an adult with sudden painless monocular vision loss is: ED now, not the eye doctor, not "wait and see." If the symptom is transient (resolves in seconds to minutes), the action is the same — the embolic source is still there. Stroke-capable EDs are most US Level 1 / 2 trauma centers and most academic medical centers; rural EDs may transfer. The 911 call frames the presentation as "I think I'm having a stroke that's affecting my eye" — this routes care correctly, where "I can't see out of one eye" may route to ophthalmology.

Failure modes

The system fails in characteristic ways:

Patient waits, vision partially returns, patient does not come in. Embolic source remains. Subsequent stroke days later.
Patient routes to optometrist / ophthalmologist first. Diagnosis is made correctly. Window has closed by the time the ED sees them. Stroke workup still gets done; vision is lost.
ED triages as "vision complaint," patient sits in waiting room. Mitigated by 911-trained stroke triage when the patient names "stroke" rather than "vision loss" in the initial complaint.
GCA missed because ESR / CRP were not drawn or because steroids were withheld pending biopsy. Fellow eye lost.
NAION diagnosed correctly (vascular but not embolic), no further workup done — missed underlying sleep apnea, severe hypertension, or other treatable vascular precipitants.

Audience

The substance applies to all adults but the risk distribution is highly age-skewed. Most CRAO is in patients over 60, peak in the seventh and eighth decades, with vascular-risk-factor prevalence (hypertension, diabetes, smoking, atrial fibrillation, hyperlipidemia, carotid disease) Mac Grory et al. 2021. GCA is essentially exclusive to over-50, with peak incidence in the seventies and a female predominance. Optic neuritis is the inverse: young adults, female-predominant, peak 20–40 Beck et al. 1992. Retinal detachment skews to high myopes and patients with prior cataract surgery. NAION occurs in 50+, often with sleep apnea and small "crowded" optic discs.

Payoff

For patients who present inside the 4.5-hour window and receive thrombolysis at a stroke-capable center, observational data suggest visual recovery of one or more lines of acuity in roughly half — versus essentially zero recovery beyond the window Mac Grory et al. 2020. The larger and more reliably-banked payoff is the secondary stroke prevention: identification and treatment of carotid stenosis (carotid endarterectomy where indicated), initiation of antiplatelet therapy, anticoagulation if atrial fibrillation is found on telemetry, statin therapy, blood pressure and glycemic optimization — this is the same toolkit that drops post-TIA five-year stroke risk substantially in NASCET-era and modern data Mac Grory et al. 2021. For GCA, early steroid initiation preserves the fellow eye in roughly half of cases that would otherwise have lost it within days Hayreh et al. 1998.

Out of scope

Bilateral simultaneous vision loss (cortical stroke, toxic optic neuropathy, functional vision loss) is a different differential and warrants a separate entry. Gradual vision loss over weeks to months (cataract, glaucoma, macular degeneration, compressive optic neuropathy from intracranial mass) is not part of this entry. Pediatric sudden vision loss is a different differential (papilledema from idiopathic intracranial hypertension, optic glioma, accommodation spasm) and is out of scope. Routine carotid screening in asymptomatic patients is also out of scope — the trigger here is the symptomatic event.

Credibility range

The optimist case

Treating sudden monocular vision loss as a stroke is unambiguously the right framework. The mechanism is identical to cerebral ischemic stroke. The downstream stroke risk after CRAO is biologically equivalent to high-risk TIA and Cohort-level data are robust and consistent across continents (Korea, US, Europe). The AHA statement, written by stroke neurologists and neuro-ophthalmologists together, formalizes what was already standard practice at academic stroke centers. The vision-preservation payoff from in-window thrombolysis is observational but the signal is large, biologically plausible (reperfusion of ischemic tissue within tolerance window), and consistent with the cerebral-stroke evidence base from which it is extrapolated. The bigger payoff — secondary stroke prevention — does not depend on the contested thrombolysis evidence at all; it depends only on doing the workup. Public-awareness intervention targeting this symptom would meaningfully reduce stroke incidence in high-risk older adults.

The skeptic case

Two skeptic threads, both real. First, the visual-recovery evidence for thrombolysis in CRAO is observational, retrospective, and subject to selection bias: patients who present quickly are also patients whose CRAO may be a different physiological entity (transient embolism rather than fixed thrombosis), making the apparent treatment benefit confounded by who shows up early. Three randomized trials of IV tPA in CRAO have been initiated; results to date have been mixed or underpowered. The AHA statement explicitly notes the evidence is preliminary. Second, the recommendation to route all sudden monocular vision loss through the stroke pathway carries opportunity costs — patients with non-arteritic AION (vascular but not embolic), retinal detachment (mechanical), optic neuritis (demyelinating), and functional vision loss will all flow through the stroke ED, consuming neurology and imaging resources without benefiting from stroke-specific workup. The signal-to-noise ratio of this protocol is unclear and may worsen as awareness improves.

The author's call

Land firmly on the optimist side, with a qualifier. The stroke-equivalent classification is settled and the secondary-prevention payoff is robust regardless of where the thrombolysis evidence eventually lands; for the reader, "go to a stroke ED" is the correct rule. The vision-preservation tPA story is correctly framed as "consider at expert centers within 4.5 hours" rather than "this works the way brain-stroke tPA works." The opportunity-cost critique is real but the alternative — routing to community ophthalmology first — has documented harm in delay Biousse et al. 2018. For the catalogue-typical reader (older adult, vascular risk factors), the rule "treat any sudden painless monocular vision loss as a stroke and call 911" yields the right answer in almost all cases. meta.controversy = 2 reflects the residual thrombolysis-efficacy debate; meta.evidence = 4 reflects the strong cohort base for the broader claim.

Stakeholder + incentive map

Stroke neurology + neuro-ophthalmology subspecialty: drove the reclassification of CRAO as stroke equivalent. Aligned incentive (correctly identify a previously-missed stroke population) and clean evidence base. Authoring body of the AHA statement.
Community ophthalmology / optometry: historical default route for "eye problems," now being asked to triage out of their pathway for time-critical cases. Some resistance, mostly resolved as the AHA statement filtered into CME.
ED triage protocols: "stroke alert" pathway expansion to include monocular vision loss is incomplete; many EDs still triage these as ophthalmology consults rather than stroke alerts. Variable by institution.
Patient advocacy: minimal — CRAO has no analog to the "FAST" stroke campaign. The substance is under-publicized relative to its incidence and severity.

Population variability

The substance applies to all adults, but the population-level value of the response is heavily age-weighted. Most CRAO and most amaurosis fugax occurs in adults over 60 with established or occult atherosclerotic disease; the stroke-prevention payoff in this group is large. In adults under 40 with sudden monocular vision loss, the differential shifts toward optic neuritis, migraine variants, retinal detachment in high myopes, and rare cardioembolic events (patent foramen ovale, vasculitis, vasospasm, sickle cell). The stroke-equivalent framework still applies — embolic events in young adults are the lower-base-rate but higher-stakes scenario, often signaling an undiagnosed structural cardiac or coagulation issue. GCA is essentially exclusive to over-50. Optic neuritis is essentially exclusive to under-50. The protocol — go to a stroke-capable ED — works across the age range because the workup branches correctly there even when the underlying mechanism is not embolic.

Knowledge gaps

What hasn't been settled: the efficacy of IV tPA for CRAO at randomized-trial grade — three trials in progress or recently completed (REVISION, THEIA, TenCRAOS) have produced mixed or pending results; the AHA statement will likely be updated as those data mature. The optimal duration and intensity of secondary stroke prevention specifically after CRAO (versus general TIA protocols) is not well-defined. Whether OCT and modern retinal imaging can identify the very narrow subset of CRAO patients who will recover with reperfusion is unknown. Whether targeted public-awareness campaigns ("if your vision goes out, call 911") would meaningfully shift presentation timing — and downstream stroke incidence — has not been tested. The role of mechanical thrombectomy in CRAO (analogous to large-vessel-occlusion stroke) is theoretical and not yet supported by trial data.

Scope relative to the brief. Brief called for sudden monocular vision loss in adults, including amaurosis fugax, with stroke-equivalent framing and downstream evaluation. Entry covers all of those plus the three differentials a reader will plausibly mistake for it (giant cell arteritis, retinal detachment / vitreous hemorrhage, optic neuritis) because the reader's decision is "is this an emergency and which one." Those alternates are present as warning callouts inside contraindications rather than as full sections — they share the same response ("go to ED, name as stroke") but split on what the ED then does. Full entries for each alternate are good follow-up candidates.

Hard call on action and cadence. Chose action: respond and cadence: as-needed. The substance is fundamentally a symptom recognition + response rule rather than an ongoing habit. know was considered and rejected because the entry's value depends on the reader actually responding, not just being aware.

Hard call on health_short_term score. Scored 2. The vision-preservation evidence (Mac Grory et al. 2020) is observational and the window is narrow, so 3 felt overconfident. The score captures real but conditional benefit — meaningful when the response happens inside 4.5 hours, near-zero outside.

Hard call on longevity score. Scored 3 rather than 4. The stroke-prevention payoff is large in the affected population but conditional on the event occurring. A reader who never experiences sudden vision loss derives no longevity benefit. Scoring 4 felt like it would over-promise; scoring 2 would under-rate the magnitude of the prevention payoff for those affected.

Controversy = 2. Stroke-equivalent classification and secondary-prevention workup are not contested. Thrombolysis efficacy for visual recovery is — three trials in progress (REVISION, THEIA, TenCRAOS), AHA statement itself flags the evidence as preliminary. Reflected this in the article evidence section without overweighting it; the controlling reason to go to the ED is the brain-stroke prevention workup, which does not depend on the thrombolysis evidence.

Excluded. Detailed retinal anatomy, OCT findings, ERG patterns — kept the mechanism section to what shapes the response decision. Pediatric sudden vision loss has a different differential (papilledema from idiopathic intracranial hypertension, optic glioma) and was excluded. Bilateral simultaneous loss has a different differential (cortical stroke, toxic optic neuropathy) and was pointed to in out-of-scope as a candidate for its own entry. Mechanical thrombectomy for CRAO is theoretical and trial-stage; excluded.

Separate-entry candidates flagged. Bilateral sudden vision loss (different differential, including occipital stroke). Giant cell arteritis as its own entry (it sprawls beyond vision into systemic vasculitis with polymyalgia rheumatica overlap, and warrants more depth than the warning callout here). NAION specifically and its link to sleep apnea. Carotid screening in asymptomatic adults (currently against per USPSTF — worth its own entry for that reason).

Future links. When entries exist for the above plus an "Atrial fibrillation screening" entry, a "Statin therapy after vascular event" entry, and a "Recognising stroke (FAST)" entry, this entry should cross-link to all of them — the workup branches into each.

Voice tradeoffs. Resisted the urge to put a single dramatic patient anecdote at the top — the felt-experience anchor here is "vision goes out in one eye," which is already vivid without dramatization. Resisted the urge to soften the "do not go to the eye doctor first" line; the literature on time-to-presentation (Biousse et al. 2018) makes the routing choice load-bearing.