The FDA has approved hyperbaric oxygen therapy for fourteen distinct medical conditions. The list includes carbon monoxide poisoning, crush injuries, radiation damage, diabetic foot wounds, necrotizing soft tissue infections, bone infections that have not responded to antibiotics, sudden unexplained hearing loss, and several others. In each of these approvals, the same basic biology is at work. The therapy has a documented effect on inflammation, cell death, oxygen delivery to damaged tissue, the growth of new blood vessels, and the regeneration of neurons. The list of conditions for which those same biological processes are disrupted — but for which insurance will not pay for HBOT — is considerably longer.
This is the story of what hyperbaric oxygen therapy actually does, why it works where it works, and why the boundary between what is covered and what is not is drawn by insurance categories rather than by the underlying science.
Hyperbaric oxygen therapy involves breathing pure oxygen inside a pressurized chamber. The pressure — typically between 1.5 and 2.5 times normal atmospheric pressure — forces oxygen into the blood plasma at levels far beyond what normal breathing achieves. This oxygen-saturated plasma can reach tissues that red blood cells cannot, including tissues where circulation has been compromised by injury, infection, disease, or radiation damage. The immediate effect is the delivery of oxygen to cells that are starving for it. The longer-term effects are the more interesting part.
A 2015 review by Yan and colleagues documented four categories of biological activity triggered by HBOT: inhibition of apoptosis (programmed cell death), reduction of inflammation, regulation of oxygen free radicals, and activation of neural stem cells. A 2006 study by Thom and colleagues found that HBOT increased the number of circulating stem cells in the bloodstream eightfold — a finding with implications for any condition in which tissue repair is impaired. Sureda and colleagues, reviewing wound healing specifically, documented that HBOT activates immune cells, modifies cytokine production, modulates inflammatory mediators, and promotes the growth of new blood vessels via increases in vascular endothelial growth factor. These are not effects that apply only to wounds. They are systemic biological responses to a specific environmental stimulus.
What it is approved for, and why
Carbon monoxide poisoning works by displacing oxygen from hemoglobin and disrupting mitochondrial function in the brain and heart, leading to cell death in the hippocampus, basal ganglia, and cerebellum. HBOT outcompetes carbon monoxide for the binding site, restores mitochondrial function, and reduces the inflammation that causes the delayed neurological damage that can appear weeks after initial exposure. The biology is well understood and the approval is uncontroversial.
Diabetic foot wounds are approved because diabetes impairs circulation, reduces the oxygen delivery necessary for wound healing, creates an environment hospitable to anaerobic bacteria, and blunts the immune response. HBOT addresses all of these simultaneously: it delivers oxygen directly through plasma to hypoxic tissue, kills or inhibits anaerobic bacteria, stimulates growth factors, and enhances the oxidative killing capacity of white blood cells. A systematic review by the Health Quality Ontario program in 2017 found that the overall trend of available studies favored HBOT in reducing amputations and improving wound healing, with meaningful results in the subset of patients most likely to respond.
Necrotizing soft tissue infections — the flesh-eating bacterial infections that kill by destroying tissue faster than surgery can remove it — are approved because the bacteria responsible thrive in low-oxygen environments. A 2014 study by Shaw and colleagues found that patients with these infections who received HBOT had a tenfold increase in survival rate. Stephens documented a twofold reduction in mortality in clostridial gas gangrene specifically. The mechanism is direct: oxygen kills the bacteria and restores the immune system’s ability to fight them.
For each of these approved conditions, the biological logic is traceable and specific. Tissue is hypoxic, or inflamed, or infected by anaerobes, or undergoing apoptosis, or failing to heal — and HBOT addresses these processes through documented mechanisms.
The off-label argument: same biology, different diagnosis
Traumatic brain injury is not on the FDA-approved list for HBOT. But the biology of TBI overlaps substantially with the biology of conditions that are approved. A traumatic brain injury produces inflammation, disrupts the blood-brain barrier, causes mitochondrial dysfunction, triggers apoptosis in neurons, impairs circulation to injured areas, and reduces neurogenesis — the production of new neurons. These are precisely the processes that HBOT modulates in CO poisoning, in stroke, in diabetic tissue damage. The mechanism of action is the same. The diagnosis code is different.
Lin and colleagues, studying TBI in rats, found that HBOT stimulated both angiogenesis and neurogenesis while increasing IL-10, an anti-inflammatory cytokine, and reducing markers of inflammation in injured brain tissue. Lee and colleagues found that long-course HBOT stimulated neurogenesis and attenuated inflammation after ischemic stroke. A phase I clinical study by Harch and colleagues of military veterans with blast-induced TBI and post-traumatic stress disorder found improvements in symptoms following low-pressure HBOT. Wolf and colleagues found improvements in symptoms after mild TBI in a randomized trial. Eve and colleagues, reviewing the evidence in 2016, concluded that HBOT represents a credible candidate treatment for PTSD in the context of TBI specifically because the neurological mechanisms of PTSD-associated TBI overlap with the mechanisms HBOT is known to address.
Stroke is another example. Stroke produces ischemic damage to brain tissue through the same mechanisms as several approved indications: oxygen deprivation, inflammation, apoptosis, mitochondrial failure. Efrati and colleagues published a randomized, prospective trial in PLOS ONE in 2013 demonstrating that HBOT induced late neuroplasticity in post-stroke patients — improvements in function that appeared months to years after the initial injury, during a period when conventional medicine considers neurological recovery complete. The mechanism proposed was reactivation of metabolically dormant neurons in the penumbra surrounding the infarct zone. This is the same neuroplasticity and neurogenesis mechanism documented in CO poisoning recovery and TBI animal models.
Spinal cord injury produces ischemia, inflammation, and blocked neurological function through mechanisms that overlap with TBI and stroke. Falavigna and colleagues published a 2017 study showing improved locomotor recovery in a rat model of thoracic spinal cord injury following HBOT. The pathophysiology — ischemia, inflammation, blocked neurogenesis — is the same as in approved TBI-adjacent indications.
Post-operative cognitive dysfunction — the confusion, memory impairment, and cognitive decline that can follow major surgery and anesthesia, particularly in older patients — shares the same oxidative stress and neuroinflammation profile as CO poisoning. Gao and colleagues found in a 2017 study that hyperbaric oxygen preconditioning improved post-operative cognitive dysfunction by reducing oxidative stress and inflammation. The biology is the same. The insurance coverage is not.
Vascular dementia, which accounts for a significant proportion of cognitive decline in older Americans, involves impaired blood flow to brain tissue, neuronal death, and impaired neurogenesis — the same triad present in approved stroke indications. Zhang and colleagues found in 2010 that HBOT improved neurogenesis in the piriform cortex in rats with vascular dementia. Mu and colleagues, in a 2011 review in Medical Gas Research, documented the state of evidence for HBOT in neurogenesis broadly and identified vascular dementia as a plausible candidate for HBOT intervention based on shared pathophysiology.
Chronic refractory infections — including late-stage Lyme disease, some biofilm-forming bacterial infections, and other conditions involving anaerobic microbial communities in poorly vascularized tissue — share the pathophysiology of the approved indications for HBOT in osteomyelitis and intracranial abscess. In both approved conditions, HBOT is effective because anaerobic bacteria cannot survive high-oxygen environments and because poor circulation to infected tissue limits antibiotic delivery. These are not properties specific to bone or brain tissue. They are properties of any poorly vascularized anaerobic infection.
The insurance barrier
A typical HBOT session costs between $150 and $300 out of pocket. A standard course of treatment for TBI or post-stroke neurological recovery involves 40 to 80 sessions. The total cost of a full treatment course for a condition like TBI, even at the lower end of those estimates, runs to $6,000 to $24,000. For a military veteran with blast-induced TBI and PTSD, or a stroke survivor with residual cognitive impairment, or a child with severe hypoxic brain injury, that cost is not theoretical. It is the difference between access and exclusion.
Medicare and most private insurers cover HBOT for the fourteen FDA-approved indications. For everything else, including TBI, stroke recovery, vascular dementia, spinal cord injury, and post-operative cognitive dysfunction, coverage is denied on the grounds that the evidence for efficacy in those specific indications is insufficient. This is a narrower standard than the one the FDA uses for its approved indications, and a narrower standard than the one implied by the biological evidence. A therapy that is approved because it reduces apoptosis, stimulates neurogenesis, and reduces inflammation in CO poisoning does not require separate proof that it reduces apoptosis, stimulates neurogenesis, and reduces inflammation in TBI. The biology is identical. The clinical trials for the specific diagnostic code are not.
The argument for maintaining the current coverage standard is a serious one. Clinical trials matter because the specific population, protocol, and outcome measures for each condition determine whether a treatment works in that setting. Animal models and shared pathophysiology are hypothesis-generating, not definitive. The history of medicine includes many treatments that worked in theory and failed in practice. Requiring randomized controlled trials for each specific application before mandating coverage is a reasonable precautionary standard, particularly for expensive treatments where the placebo effect in subjective outcome measures is real.
The argument against the current standard is distributional. The patients who bear the cost of the evidentiary gap are those who cannot afford $200 per session out of pocket. They are people with TBI, many of them veterans. People with post-stroke cognitive impairment. People with spinal cord injuries. The trial evidence that would close the gap requires funding, which requires either industry interest — which HBOT, as an unpatentable treatment, does not attract — or public funding through NIH or DARPA, which has been episodic and insufficient. The consequence of the funding gap is not scientific neutrality. It is the perpetuation of a coverage decision that is most costly to the people with the least resources to absorb it.
The FDA’s own approved indications for HBOT constitute a map of the conditions it can address at the level of basic biology. That map includes inflammation, ischemia, anaerobic infection, mitochondrial dysfunction, impaired neurogenesis, and disrupted wound healing. These processes are not exclusive to the fourteen conditions on the approved list. They are present in dozens of conditions for which coverage is denied. The scientific question of whether HBOT is effective in each of those conditions requires specific study and specific data. The scientific question of whether the underlying biology is the same has largely been answered. The gap between those two questions is where people are spending money they often do not have, or going without treatment they may need.
The research needs to happen. Until it does, the access gap is a policy choice, not a scientific necessity.
Sources: Yan S et al., Medical Gas Research, 2015. Thom SR et al., Am J Physiol Heart Circ Physiol, 2006. Sureda A et al., PLOS One, 2016. Health Quality Ontario, Ont Health Technol Assess Ser, 2017. Shaw JJ et al., Surg Infect (Larchmt), 2014. Stephens MB, Postgrad Med, 1996. Lin KC et al., J Trauma Acute Care Surg, 2012. Lee YS et al., Mediators Inflamm, 2013. Harch PG et al., J Neurotrauma, 2012. Wolf G et al., J Neurotrauma, 2012. Eve DJ et al., Neuropsychiatr Dis Treat, 2016. Efrati S et al., PLOS One, 2013. Falavigna A et al., Spine, 2017. Gao ZX et al., Neural Regen Res, 2017. Zhang T et al., Brain Inj, 2010. Mu J et al., Medical Gas Research, 2011.
American Life is independent, advertisement-free analysis. Subscribe to receive every article in your inbox.
