Central retinal artery occlusion (CRAO) is an eye stroke. A clot or embolus chokes off the main artery feeding the retina, and vision in that eye can go dark in seconds. It is one of the few emergencies where hyperbaric oxygen therapy (HBOT) has a real physiologic rationale and a spot on the recognized-indications list. It is also a place where the marketing runs far ahead of the proof.
This guide separates the two. We walk through why HBOT belongs in the CRAO conversation at all, what the clearest studies actually show, how narrow the time window is, and where the evidence quietly falls apart. No hype. No "miracle" language. Just what the data says and where it stops.
Quick Answer
- CRAO is on the UHMS recognized-indications list; the evidence is real but weak
- The therapeutic window is tiny; benefit fades fast after roughly 6-12 hours
- Best data are retrospective; no large randomized sham-controlled trial exists
- HBOT for CRAO is an ER-grade emergency, not a wellness-clinic booking
What CRAO Is, and Why Oxygen Even Enters the Picture
The retina is metabolically greedy. It burns oxygen faster per gram than almost any other tissue in the body, and it has almost no reserve. When the central retinal artery blocks, the inner retina starts dying within minutes to hours. That is why CRAO presents the way it does: sudden, painless, profound vision loss in one eye, often described as a curtain dropping or the world going gray.
Here is the mechanism that makes HBOT plausible. The retina gets oxygen from two directions. The inner retina relies on the central retinal artery, the vessel that just clotted. The outer retina is fed from behind by the choroid, a separate blood supply that is usually still working. Under normal air pressure, that choroidal oxygen cannot diffuse far enough forward to rescue the starving inner retina. Under hyperbaric pressure, plasma carries dramatically more dissolved oxygen, and the theory is that this supercharged choroidal supply can reach across and keep inner-retinal cells alive until the clot clears or collateral flow opens up.
That is a genuinely reasonable idea. It is also the kind of reasonable idea that has failed clinical testing many times in medicine. So the question is never "does the mechanism make sense" but "does it change what patients can see." Hold that distinction. The rest of this article lives inside it.
Is CRAO an FDA-Approved or UHMS-Approved Use of HBOT?
This is where honest framing matters, because the words "approved" and "FDA-cleared" get thrown around loosely by clinics.
Two different bodies are involved, and they do different things:
- The FDA clears hyperbaric chambers as medical devices and regulates the medical-grade oxygen used inside them. The FDA has repeatedly warned that HBOT is promoted for many conditions it has never cleared it for, and that unproven claims can delay real treatment (FDA, Letter to Health Care Providers).
- The Undersea and Hyperbaric Medical Society (UHMS) is the professional body that maintains the list of indications it considers supported enough for clinical use. This is the list clinicians actually reference, and it is the "14 indications" (now expanded) you see cited across the field (UHMS HBO Indications).
CRAO sits on the UHMS list under the acute arterial-insufficiency / central retinal artery occlusion heading. So a clinic saying "CRAO is a recognized indication" is technically correct. What that phrase does not mean is that HBOT is proven to restore vision, or that a randomized trial has confirmed benefit. UHMS itself grades this indication on limited evidence. Recognized does not equal proven.
| Claim you might hear | Accurate reading |
|---|---|
| "CRAO is FDA-approved for HBOT" | Misleading. FDA clears chambers as devices; it does not "approve" CRAO as an indication the way it approves a drug |
| "CRAO is a UHMS-recognized indication" | True. It is on the list, graded on limited/low-quality evidence |
| "HBOT restores vision in CRAO" | Overstated. Some patients improve; no randomized trial proves the effect |
| "It works if you get there fast enough" | Partly supported. Early treatment correlates with better outcomes, but the studies cannot prove causation |
If you take one thing from this section: CRAO's place on the recognized list reflects biological plausibility and modest observational signals, not the kind of proof that backs, say, HBOT for decompression sickness or carbon monoxide poisoning. For the difference between recognized indications and off-label claims, see our guide to the 14 UHMS-recognized indications and the broader HBOT regulation guide covering the FDA and UHMS.
What Happens to CRAO Vision Without Any Treatment
You cannot judge a treatment until you know the natural course of the disease. This is the single most important context clinics tend to skip, because the natural history of CRAO is grim, and grim natural history makes almost any intervention look like it "did something."
The landmark natural-history work comes from Sohan Singh Hayreh and colleagues. In their large cohort of non-arteritic CRAO eyes, visual outcomes were poor and largely fixed by presentation: the great majority of eyes with a complete occlusion ended with counting-fingers vision or worse, and meaningful spontaneous recovery was uncommon (Hayreh & Zimmerman, 2005, PMID 16138997). A later synthesis of ocular vascular occlusive disorders reinforced that final visual acuity in CRAO is heavily determined at onset, with limited late improvement (Hayreh, 2014, PMID 24769221).
But "uncommon" is not "never," and that caveat is the whole problem. A minority of CRAO eyes do improve on their own, often because the embolus fragments and moves, or a cilioretinal artery spares the central vision. When a therapy is given to everyone and a slice of those patients would have improved regardless, the therapy inherits credit it may not have earned. This is exactly why retrospective HBOT-for-CRAO studies must be read carefully rather than at face value.
Two numbers to anchor on:
- CRAO is uncommon, with an incidence in the range of roughly 1 to 2 per 100,000 person-years in population estimates.
- Experimental work on primate retina suggests irreversible inner-retinal damage after only a few hours of complete occlusion, which is the biological basis for the "get treated fast" urgency.
That short tolerance time is the entire drama of CRAO treatment. It also sets up the next section.
One more layer of honesty belongs here. Not all CRAO is the same, and that matters for how the evidence reads. Non-arteritic CRAO, driven by an embolus, is the common form and the one most HBOT studies enroll. Arteritic CRAO, caused by giant cell arteritis, is a different disease with a different treatment path (urgent steroids to protect the other eye), and mixing the two muddies any analysis. When you read that "HBOT helped CRAO," check whether the study specified non-arteritic disease. The better ones do. It is one more reason to treat the headline numbers cautiously.
How HBOT Is Actually Delivered for a CRAO Emergency
If HBOT is used, the delivery looks nothing like a wellness-clinic package of 40 relaxing sessions. It is an acute rescue attempt, and the logistics are unforgiving.
A typical emergency approach involves getting the patient into a chamber as fast as possible, often after quick bedside measures like ocular massage and steps to lower intraocular pressure while transport is arranged. Reported protocols compress oxygen delivery around the first hours, and the number of sessions is far smaller than chronic-wound protocols. In the academic single-center series, patients received a modest handful of treatments, not weeks of them (Kim et al., 2025, PMID 39368618). A 2026 comparison of different HBOT protocols found that how the oxygen was dosed changed outcomes, which tells you the protocol is still being optimized rather than settled (Eye (Lond), 2026, PMID 41667662).
The practical bottleneck is access. Emergency-capable hyperbaric facilities are not on every corner, and the ones equipped to treat an acute CRAO at 2 a.m. are rarer still. A chamber advertised for "wellness" may not be staffed, pressurized, or willing to handle a true ocular emergency, and the delay in finding the right one can consume the entire treatment window. This is a big part of why real-world CRAO outcomes stay poor even where HBOT is available: the clock usually wins.
HBOT is also not risk-free, which the sober framing requires stating plainly. Documented risks include middle-ear and sinus barotrauma from pressure changes, temporary vision changes, oxygen toxicity, and, rarely, seizures. For most healthy patients these risks are manageable under trained supervision, but they are real and are one reason emergency delivery should happen in a properly staffed setting rather than a spa. We cover the broader safety and vision-side-effect picture in our explainer on how HBOT can affect your eyesight.
The Time Window: Where the Evidence Is Least Bad
If HBOT helps CRAO at all, the studies agree the effect concentrates in patients treated early. This is the most consistent finding in the literature, and it is the one clinics get roughly right.
A single-center retrospective series from a major academic center is among the cleaner data points. Among CRAO patients treated with HBOT, those reaching treatment within the earliest hours were markedly more likely to show a clinically meaningful gain in visual acuity than those treated later, with the early-treated subgroup showing the strongest response (Kim et al., 2025, American Journal of Ophthalmology, PMID 39368618). Coverage of that work by the American Academy of Ophthalmology summarized it plainly: early HBOT after CRAO may enhance recovery, with heavy emphasis on the word "may."
Other retrospective cohorts echo the pattern. A retrospective study reported a significant improvement in best-corrected visual acuity from baseline to post-HBOT in treated patients, with better results tied to shorter delay to treatment (Cureus, 2024, PMID 39113814). A 2025 emergency-medicine analysis of non-arteritic CRAO likewise found HBOT associated with visual improvement, again concentrated in earlier presentations (Am J Emerg Med, 2025, PMID 40784185).
So what is the window? The honest answer is that it is short and not precisely defined:
| Time to treatment | What the observational data suggest |
|---|---|
| Under ~6 hours | Best reported response rates; strongest rationale before irreversible retinal death |
| ~6-12 hours | Some patients still improve; UHMS considers treatment reasonable within 24 hours |
| 12-24 hours | Diminishing returns; individual case reports of benefit but weak group-level signal |
| Beyond 24 hours | Little expected benefit; retina likely infarcted |
UHMS guidance treats presentation within 24 hours as the outer bound for considering HBOT, but "considerable within 24 hours" and "likely to help at 20 hours" are not the same statement. Every hour matters, and the useful window is realistically measured in the first several hours, not the first day. This is why CRAO is an emergency-department problem: the clock started before the patient even noticed.
Where the Evidence Falls Apart
Now the uncomfortable part. Everything above describes an association, and association is not proof. Here is exactly why the CRAO-HBOT literature cannot yet claim that HBOT restores vision.
No large randomized sham-controlled trial exists. The entire evidence base is retrospective cohorts, case series, and small comparative studies. None randomly assigned patients to real versus sham hyperbaric pressure and then measured vision blind. That is the design that separates a true drug effect from wishful thinking, and for CRAO it has never been done at scale.
Selection bias runs in one direction. Patients who make it to a hyperbaric chamber fast are, by definition, patients near a hyperbaric facility with a fast diagnosis and the mobility to get there. Those patients differ from the general CRAO population in ways that independently predict better outcomes. When early-treated HBOT patients do better, some unknown fraction of that gap is the sorting, not the oxygen.
Comparator studies are mixed and small. A comparative-effectiveness study weighing conservative care, HBOT, and endovascular retinal surgery found differences between approaches but was retrospective and modest in size (Int Ophthalmol, 2024, PMID 39521754). A 2020 analysis explicitly framed the open question as thrombolysis, HBOT, or both, precisely because no approach has clearly won (J Thromb Thrombolysis, 2020, PMID 32166539). A 2026 study comparing HBOT protocols found outcome differences by protocol, which is useful for optimizing dosing but says nothing about whether HBOT beats no HBOT (Eye (Lond), 2026, PMID 41667662).
Meta-analysis inherits the weaknesses. A retrospective study bundled with a systematic review and meta-analysis reported a signal favoring HBOT, but pooling low-quality retrospective data does not upgrade it to high-quality evidence (Korean J Ophthalmol, 2022, PMID 34743490). Garbage in, meta-analyzed garbage out is a real failure mode, and the authors of these reviews routinely call for randomized trials for exactly this reason.
The pattern here mirrors the honest assessment across HBOT's off-label frontier, which we cover in our evidence-first skeptic's guide to whether HBOT actually works. CRAO is one of the better-supported cases and it still lacks a definitive trial. Take that as calibration for everything softer.
HBOT Compared to the Other CRAO Options
CRAO has no proven, guaranteed treatment. That is the blunt reality, and it shapes how HBOT fits. Because there is no clear winner, several interventions get tried, sometimes together, sometimes in sequence.
| Approach | Evidence status | Practical notes |
|---|---|---|
| Conservative measures (ocular massage, lowering eye pressure, breathing exercises) | Weak; largely unproven | Cheap, fast, low-risk; often done while arranging definitive care |
| Hyperbaric oxygen therapy | Recognized indication, low-quality evidence | Time-critical; requires a nearby emergency-capable hyperbaric facility |
| Intra-arterial or intravenous thrombolysis | Investigational; active trials | Bleeding risk; availability and timing constraints |
| Endovascular / surgical approaches | Limited, specialized | Few centers; not standard |
| Treating the underlying cause | Essential regardless | CRAO is a stroke equivalent and demands a full workup |
That last row is the one no clinic should let you skip. CRAO is often the eye's warning shot for systemic vascular disease. A patient with CRAO needs urgent evaluation for carotid disease, cardiac sources of emboli, and stroke risk, because the same clot mechanism can hit the brain next. Whatever you decide about the eye, the stroke workup is not optional. For how HBOT is framed in the broader stroke conversation, see our evidence atlas on HBOT for stroke recovery.
What This Means If It Happens to You or Someone You Love
Sudden painless vision loss in one eye is a 911 emergency. Not a call-your-eye-doctor-tomorrow event. The retina does not wait.
Here is the sober playbook, given everything above:
- Go to an emergency department immediately. Do not drive to a wellness spa advertising oxygen chambers. CRAO needs an ER-grade evaluation, stroke workup, and, if HBOT is on the table, a hospital-affiliated or emergency-capable hyperbaric facility.
- Understand what HBOT can and cannot promise. If offered early, it is a reasonable attempt with genuine physiologic logic and modest supporting data. It is not a guarantee, and outcomes are frequently limited even with fast treatment because the natural history of complete CRAO is poor.
- Weigh it against the clock, not the marketing. The value of HBOT in CRAO is almost entirely about how fast it starts. A chamber six hours away that opens tomorrow morning is not the same offer as a chamber down the hall right now.
- Insist on the systemic workup. The eye is the presenting symptom. The vascular disease behind it is the real threat.
None of this is a reason to dismiss HBOT for CRAO. It is a reason to keep it in proportion: a recognized, physiologically sound, time-critical option with weak but non-zero evidence, best delivered in an emergency setting, never oversold.
Frequently Asked Questions
Does HBOT cure central retinal artery occlusion? No. There is no cure for CRAO. HBOT is a recognized treatment attempt with a reasonable mechanism and modest observational support, but no randomized trial proves it restores vision, and many patients see limited recovery even with prompt treatment.
How fast do I need to get HBOT for it to help? As fast as physically possible. The strongest observational signals come from patients treated within the first several hours, and benefit drops sharply after that. UHMS considers treatment within 24 hours, but realistically the useful window is measured in hours, not a full day.
Is HBOT for CRAO FDA-approved? Not in the way people assume. The FDA clears hyperbaric chambers as devices; it does not "approve" CRAO as an indication the way it approves a drug. CRAO is a UHMS-recognized indication, graded on limited-quality evidence. Recognized is not the same as proven.
Should I go to a wellness clinic's oxygen chamber for sudden vision loss? No. Sudden painless vision loss is an emergency. Go to an emergency department. CRAO requires urgent diagnosis, a stroke workup, and, if HBOT is used, an emergency-capable hyperbaric facility, not a spa-style wellness chamber.
Why does CRAO evidence stay weak if it is a recognized indication? Because "recognized" reflects biological plausibility plus small retrospective studies, not a definitive randomized trial. No large sham-controlled study has ever compared real versus fake hyperbaric pressure for CRAO, so selection bias and the disease's variable natural history still cloud the picture.
Medical Disclaimer
This article is for general educational purposes only and is not medical advice. It does not establish a doctor-patient relationship and should not be used to diagnose or treat any condition. Central retinal artery occlusion is a medical emergency; sudden vision loss requires immediate emergency care. HBOT carries risks, including barotrauma to the ears and sinuses, oxygen toxicity, and, rarely, seizures, and it is not appropriate for everyone. Evidence for HBOT in CRAO is limited and largely retrospective. Always consult qualified emergency and ophthalmology clinicians about your specific situation. Never delay emergency care to pursue any elective therapy.
Sources
- Kim BM, Wang KY, Xu TT, et al. Outcomes of Hyperbaric Oxygen Treatment for Central Retinal Artery Occlusion: A Single Center Experience. American Journal of Ophthalmology. 2025. PMID 39368618
- The effects of hyperbaric oxygen treatment for non-arteritic central retinal artery occlusion (HBOT-CRAO). American Journal of Emergency Medicine. 2025. PMID 40784185
- Effects of Hyperbaric Oxygen Therapy in the Treatment of Patients With Central Retinal Artery Occlusion: A Retrospective Study. Cureus. 2024. PMID 39113814
- The Effects of Hyperbaric Oxygen Therapy in Patients with Central Retinal Artery Occlusion: A Retrospective Study, Systematic Review, and Meta-analysis. Korean Journal of Ophthalmology. 2022. PMID 34743490
- Comparative efficacy of conservative, hyperbaric oxygen, and endovascular retinal surgery approaches in central retinal artery occlusion. International Ophthalmology. 2024. PMID 39521754
- Comparative effectiveness of varying hyperbaric oxygen protocols in the treatment of acute central retinal artery occlusion. Eye (London). 2026. PMID 41667662
- Acute phase treatment in central retinal artery occlusion: thrombolysis, hyperbaric oxygen therapy or both? Journal of Thrombosis and Thrombolysis. 2020. PMID 32166539
- Hayreh SS, Zimmerman MB. Central retinal artery occlusion: visual outcome. American Journal of Ophthalmology. 2005. PMID 16138997
- Hayreh SS. Ocular vascular occlusive disorders: natural history of visual outcome. Progress in Retinal and Eye Research. 2014. PMID 24769221
- Undersea and Hyperbaric Medical Society. HBO Indications. uhms.org
- U.S. Food and Drug Administration. Follow Instructions for Safe Use of Hyperbaric Oxygen Therapy Devices: Letter to Health Care Providers. fda.gov
— The HBOT Finder Team