There is no definitive, gold-standard trial that settles HBOT for brain injury. That absence is real. But it's worth asking honestly why the trial was never run — because the answer turns out to be about the structure of how medicine gets funded, not about the treatment failing a test. No one here is a villain. The system produces this outcome, and the system is what we're describing.
A note on voice. Nothing below blames a person. Researchers, clinicians and regulators are all acting reasonably inside the incentives they're given. That's exactly the point: a structure can produce a bad outcome even when everyone in it means well.
A definitive multi-centre blinded RCT costs millions. In medicine that money almost always comes from a company protecting a molecule it owns. Oxygen isn't ownable — no company profits if HBOT is proven, so there's no one to write the cheque. The trials that get funded are for patented drugs. Not because those drugs matter more — because someone stands to make the money back.
To blind the trial you need a sham chamber that pressurises enough to pop the patient's ears — so they believe they're being treated — while delivering little real oxygen. That's a whole second rig, extra staff, extra cost. Most centres can't or won't build it, so even well-meaning trials come out unblinded. (This is exactly why the Weaver 2025 double-blind sham trial is such a big deal — someone finally built it, and real oxygen still won.)
In some countries HBOT went mainstream decades ago. Once a treatment is standard care, you've lost "equipoise" — the honest uncertainty a placebo trial requires. You can't randomise a patient to fake oxygen when your own guidelines call the real thing indicated. So the studies there are pragmatic "HBOT-plus-usual-care vs usual-care" — huge in number, unblinded by necessity. Vast literature, soft method, at the same time.
It's a closed loop. The word "unproven" never has to face a result, because nobody funded the result that would test it. The label sustains itself.
Put those together and you get a treatment stuck in permanent limbo — not because it was tested and failed, but because the structure of medical funding means the test that would settle it is in nobody's financial interest to pay for. The absence of proof is a fact about money and patents, not about the oxygen.
Mainstream medicine does have one recognised active neuroprotective treatment for a brain starved of oxygen: therapeutic hypothermia — cooling. It's worth looking honestly at how well it works, and what we call it.
This is the quietly important part.
The recognised treatment for this injury succeeds on the order of ~25% relative benefit and fails for nearly half the babies who receive it — yet it is never called "unproven." It is simply standard care.
That's not an argument against cooling. Cooling saves real children and should be used. It's an argument about the double standard in the label: a treatment at that level of benefit is "standard," while a treatment with three RCTs, a double-blind sham trial, imaging endpoints and an FDA IND is "experimental." The difference isn't the evidence. It's who paid for it, and when it entered practice.
Cooling figures: relative risk ≈ 0.75, NNT ≈ 7 for neonatal hypoxic-ischaemic encephalopathy. For adult post-cardiac-arrest the picture is weaker still — later TTM trials found no advantage of 33 °C over simply controlling fever. Figures as summarised in this project's research file.
When a brain is already starved of oxygen, mainstream medicine's answer is to keep the patient stable and wait — ventilation, blood pressure, seizure control, managing swelling — plus rehabilitation later, and cooling only inside a narrow window of hours. Past that window, cooling isn't on the table.
There is no competing drug or procedure that actively repairs an established hypoxic brain. No pharmaceutical neuro-repair therapy has proven effective beyond that acute cooling window. So the honest "alternative" to trying HBOT isn't a rival treatment — it's supportive care, rehab, and waiting, or comfort care.
The choice, told straight, is HBOT versus nothing that acts on the injury itself.
The placebo effect is one of the most replicated findings in all of science — belief, demonstrably, heals. And then, having proved it, medicine declares that healing a nuisance: noise, a variable to subtract and throw in the bin. It calls any treatment that works "only through placebo" a failure — even though the patient actually got better.
There's a genuine tension worth naming honestly: oxygen either reduces the swelling in a brain or it doesn't, and a child's optimism has no vote in that. Where the injury is physical, we do want to know the oxygen is doing the work — that's fair, and it's why the sham-controlled trial matters. But there's something strange in demanding a belief-cancelling test even for patients who cannot believe, while treating our own most-replicated healing effect as an embarrassment.
The certainty is what blocks the proof. — the equipoise problem, in one line
HBOT for the brain sits in limbo not because it was tested and failed, but because the test that would settle it is in nobody's financial interest to fund, is expensive to blind, and is ethically fraught where the treatment is already standard. That's a story about a funding structure. It says nothing about whether the oxygen works — the studies on this site speak to that, and they're worth reading for yourself.