A person pulled out of a burning building can look reassuring in the emergency department. The burns on the arms are partial thickness, the voice sounds ordinary, the chest X-ray is unremarkable. Then the airway narrows, sometimes within the first four hours, sometimes while the fluid resuscitation is still running, and the problem stops being the skin.
Among patients admitted with burns, the pooled prevalence of inhalation injury is 15.7% — with a 95% confidence interval of 13.4 to 18.3 — and its presence carries a pooled adjusted in-hospital mortality odds ratio of 3.2 (2.5 to 4.3). Those figures come from a new paper published this week, and they are worth carrying into any conversation about what hyperbaric oxygen can do for someone who has breathed a fire.
The paper is a 16-page narrative review by Metin Akgün and Benoit Nemery in the Balkan Medical Journal (volume 43, issue 11, pages 605 to 620, formally published 2 November 2026). Nemery is an occupational and environmental toxicologist at KU Leuven, and the review is organised by mechanism rather than by setting: house fires, industrial releases, wildfire smoke, fumigated shipping containers, and the gas that comes off a waterproofing spray. It carries 159 references.
Its verdict on hyperbaric oxygen takes one sentence, and that sentence is not promotional.
One exposure, three injuries
A fire attacks the lung in three ways that only partly overlap. Heat damages the upper airway, but thermal injury rarely travels far below the vocal cords because the gas cools as it descends. Chemical irritation does travel: the site of injury is set by how soluble the gas is in water, so highly soluble irritants hit the nose, pharynx and trachea, while poorly soluble ones — phosgene, oxides of nitrogen, some pyrolysis products — pass deeper into the alveoli before reacting. Then come the systemic asphyxiants, which do their damage without touching the airway lining at all: carbon monoxide binding haemoglobin some 200 to 250 times more avidly than oxygen does, and hydrogen cyanide shutting down cytochrome c oxidase at the end of the mitochondrial chain.
The review adds a fourth pathway that gets less airtime: ultrafine particles and chemical vapours can cross the alveolar-capillary membrane and enter the circulation, one explanation for how a lung exposure produces systemic inflammation and late neurological harm. The authors are careful with the epidemiology. They note that the mortality association weakened and lost significance once severity was accounted for, and that no single surveillance system captures acute inhalation injury at all; the systematic review that would pin its frequency down is registered but not yet reported.
The sequelae list is the part clinicians should read twice. Most survivors recover. A clinically important minority do not, developing reactive airways dysfunction syndrome, bronchiolitis obliterans, organising pneumonia, bronchiectasis, or an accelerated decline in lung function — and structural airway lesions, sinonasal and ocular damage and late neurological effects are, in the authors’ words, systematically underrecorded.
What the review does with a chamber
Akgün and Nemery build a table that separates clinical consensus from persistent uncertainty. In the consensus column sit early endotracheal intubation for impending upper-airway obstruction and 100% oxygen by high-flow non-rebreather for anyone with suspected smoke or carbon monoxide exposure. In the uncertainty column sit, among other things, the long-term respiratory surveillance schedule and — precisely — “hyperbaric oxygen for CO poisoning: patient selection, timing, and availability.”
Their summary sentence on adjunctive therapy is blunt. Nebulised anticoagulants, hyperbaric oxygen and corticosteroids, they write, “remain uncertain.” Management, they note, has rested on observational data, small case series and expert opinion, while only ventilatory support has a firmer footing from the acute respiratory distress syndrome guidelines.
Two things follow. The first is that it is not a swipe at hyperbaric medicine: the review also gives conditional support to early bronchoscopy for diagnosis and removal of carbonaceous casts, to nebulised unfractionated heparin with N-acetylcysteine, and to empirical hydroxocobalamin for suspected cyanide poisoning with unexplained haemodynamic instability or severe lactic acidosis — then attaches the counterweight, a multicentre cohort of 739 smoke-inhalation admissions in which hydroxocobalamin was associated with more acute kidney injury and more renal replacement therapy. Adjuvants in this field arrive with an enthusiasm deficit on the other side of the ledger.
The second is that this is a narrative review, and it says so: no pooled estimate, no risk-of-bias table. It is an expert map of a messy literature, and its value lies in the mapping.
Twenty minutes at three atmospheres
Where the case for a chamber is strongest, and where this review is most generous, is pharmacokinetics. Breathing room air, carboxyhaemoglobin has an elimination half-life of roughly 300 minutes. On 100% oxygen at one atmosphere that falls below 90 minutes. At three atmospheres absolute it is about 20 to 25 minutes. The review prints all three numbers and cites the Undersea and Hyperbaric Medical Society’s 2024 chapter on carbon monoxide poisoning, which reaffirms hyperbaric oxygen as an option for severe CO poisoning while conceding that the evidence is mixed and the optimal time window uncertain.
That is a real and well-measured physiological effect. It is also a surrogate. Moving a toxin off a protein faster is not the same as preventing delayed neurological sequelae, and the trials that tried to bridge the gap have not settled it.
The largest positive trial remains Weaver and colleagues in the New England Journal of Medicine in 2002, which reported a reduction in cognitive sequelae at six weeks — 25.0% against 46.1%, adjusted odds ratio 0.45 (0.22 to 0.92). It was stopped after the third of four planned interim analyses, a design fact the Cochrane reviewers later treated as a reason to discount the effect size. The French paired trials reported by Annane and colleagues in Intensive Care Medicine in 2011, covering 385 patients, went the other way: adding a hyperbaric session to normobaric oxygen conferred no benefit in patients whose loss of consciousness had been transient (58% versus 61% complete recovery at one month), and in comatose patients two hyperbaric sessions produced worse recovery than one (47% versus 68%; unadjusted odds ratio 0.42, 0.23 to 0.79).
Pooling six trials and 1,361 participants, the Cochrane review by Buckley and colleagues found no significant reduction in neurological deficits, odds ratio 0.78 (0.54 to 1.12). It cautioned that heterogeneity was substantial, that design or analysis flaws were evident in every included trial, and that both positive trials had been stopped early for benefit, which likely inflated the observed effect. That review was last updated in 2011. Fifteen years of subsequent practice have not replaced it, and the review’s own framing is that a strong professional-society recommendation coexists with an unresolved trial literature.
Thirteen thousand burn patients, sixty-seven chambers
The one large human dataset that looks specifically at burns plus inhalation injury is a registry study, and its result is uncomfortable. Ryan Nygaard and Frederick Endorf, using the American Burn Association’s National Burn Repository, identified 13,044 admissions coded with both a cutaneous burn and an inhalation injury. Sixty-seven of them — about half a percent — received hyperbaric oxygen.
Those 67 patients were older, mean age 51.7 against 42.8 years, but had smaller burns, so their Baux scores — a standard mortality predictor combining age and burn size — came out nearly identical, 66.6 against 65.2. Mortality was 29.9% in the hyperbaric group against 17.5% in the rest, p = 0.01. In a multivariable model adjusting for age, sex, race, payer, burn mechanism, total body surface area, procedure count and ventilator days, hyperbaric oxygen was an independent predictor of death, odds ratio 2.484, p = 0.004. The paper appeared in the Journal of Burn Care and Research in 2021.
The authors do not conclude that oxygen under pressure kills people, and neither should anyone reading them. They write that severity of inhalation injury is hard to determine in the registry, so the patients selected for a chamber may well have had worse inhalation injury — a textbook case of confounding by indication, in which the treatment marks the sickest patients rather than causing their deaths. The 67 also represent 0.5% of the cohort, and the authors cannot tell whether that half-percent reflects genuine rarity of use or burn centres that simply do not record hyperbaric sessions in the database.
What the study does establish is that the therapy is uncommon in American burn care, and that the largest observational look at this question produced a signal pointing away from benefit. That is not proof. It is the shape of the evidence a clinician would have to argue against.
Where the indication lists stop
Professional and payer documents are more categorical than the literature, and their wording is worth quoting.
The Undersea and Hyperbaric Medical Society’s fifteenth-edition indications list numbers entries 01 through 15, with lettered subdivisions at 02 and 06 — seventeen in all. Entry 02a is carbon monoxide poisoning. Entry 02b is carbon monoxide poisoning complicated by cyanide poisoning. Entry 13 is acute thermal burn injury. There is no entry for inhalation injury, smoke inhalation, or chemical pneumonitis, and none covers the airway lesion itself.
Medicare’s National Coverage Determination 20.29, whose current version took effect on 3 April 2017, splits the question in a way that maps almost exactly onto the physiology. The covered list includes, as separate items, acute carbon monoxide intoxication and cyanide poisoning. The non-covered list includes skin burns (thermal) and then, in a line that deserves to be better known, “acute thermal and chemical pulmonary damage, i.e., smoke inhalation with pulmonary insufficiency.”
Read carefully, that is a coherent position rather than a contradiction. The two things a chamber demonstrably changes within minutes — carboxyhaemoglobin and, by a different mechanism, cyanide — are payable. The thing it does not repair, a chemically burned and oedematous airway, is explicitly excluded, along with thermal skin burns, which are FDA-cleared and UHMS-listed but not Medicare-covered. The rule is a payment judgement, not a prohibition, and patients in the United States do receive hyperbaric treatment in this situation, usually as part of an emergency CO pathway rather than as airway therapy.
The device side has constraints that apply regardless of indication. Hyperbaric chambers are class II devices cleared through the 510(k) process under product code CBF. In August 2025 the FDA issued a letter to health care providers stating that it was aware of reports of fires occurring with hyperbaric oxygen therapy devices that resulted in serious injuries and deaths, that the root cause was not then known, and that providers should follow the manufacturer’s instructions for use, verify grounding, train and maintain staff, monitor the patient for the whole treatment, avoid prohibited items including electrical or static devices, and dress patients in hyperbaric-compatible fabrics such as cotton rather than wool or synthetics. None of that is specific to smoke inhalation. All of it travels with the chamber.
The hazards the review wants noticed
The most useful part of the new paper has nothing to do with chambers. It is a catalogue of exposures that clinicians and the public routinely fail to connect to inhalation injury: chloramines generated by mixing household cleaning agents, fumigants used in ships, freight containers and greenhouses, fluorine-based agents from waterproofing sprays and polymer pyrolysis, and pesticides that injure the lung by inhalation or by systemic absorption.
The scale-setting numbers are startling. Household air pollution from solid-fuel combustion was associated with an estimated 2.8 million deaths in 2023, mostly in low- and middle-income countries. Wildfire-related fine particulate matter accounts for more than 10,000 excess deaths annually in the United States alone. The State of Global Air 2025 report attributed 7.9 million deaths in 2023 to air pollution overall, and the Global Burden of Disease study for 2021 put particulate-matter pollution at 8.0% of global disability-adjusted life-years. Those are measures of exposure scale, not of acute injury incidence, and the authors say as much.
For anyone reading this with a chamber in mind, the practical sequence the review describes is unglamorous. Assess the airway and intubate early if obstruction is coming. Give 100% oxygen by high-flow non-rebreather and keep giving it until carboxyhaemoglobin is below 5%. Consider hydroxocobalamin if cyanide exposure is plausible, weighing the renal signal. Bronchoscope early if there is carbonaceous debris or a cast forming. Everything else — including the pressurised oxygen — sits in the column marked uncertain, and the most honest version of the offer is that a chamber is being used for the carbon monoxide and cyanide component of a smoke exposure, not for the injured airway, and that the payer rule treats those as different things because the evidence does.
Two boundaries are worth stating plainly, because this is a field where they blur. Most wellness, recovery and longevity uses of hyperbaric oxygen are not FDA-cleared indications, however confidently they are sold, and the fire and electrical hazards in the 2025 FDA letter do not become smaller because a session is elective. And in a fire victim, the intervention with the best evidence behind it remains a secured airway, high-flow oxygen and a burn centre — none of which needs a pressure vessel, and all of which are available at three in the morning.
Sources
- Akgün M, Nemery B. Inhalation Injury: Update on Pathogenesis, Emerging Hazards, and Management. Balkan Med J 2026;43(11):605-620 (published 2 Nov 2026; e-pub 25 Sep 2026).
- Nygaard RM, Endorf FW. Hyperbaric Oxygen and Mortality in Burns With Inhalation Injury: A Study of the National Burn Repository. J Burn Care Res 2021;42(5):900-904.
- Weaver LK, Hopkins RO, Chan KJ, et al. Hyperbaric oxygen for acute carbon monoxide poisoning. N Engl J Med 2002;347:1057-1067.
- Annane D, Chadda K, Gajdos P, Jars-Guincestre MC, Chevret S, Raphael JC. Hyperbaric oxygen therapy for acute domestic carbon monoxide poisoning: two randomized controlled trials. Intensive Care Med 2011;37(3):486-492.
- Buckley NA, Juurlink DN, Isbister G, Bennett MH, Lavonas EJ. Hyperbaric oxygen for carbon monoxide poisoning. Cochrane Database Syst Rev 2011;2011(4):CD002041.
- Dépret F, Hoffmann C, Daoud L, et al. Association between hydroxocobalamin administration and acute kidney injury after smoke inhalation: a multicenter retrospective study. Crit Care 2019;23(1):421.
- Weaver LK. Carbon monoxide poisoning (reprinted from the 2023 hyperbaric indications manual, 15th edition). Undersea Hyperb Med 2024;51(3):253-276.
- Undersea and Hyperbaric Medical Society. Hyperbaric Oxygen Therapy Indications, 15th edition.
- Centers for Medicare & Medicaid Services. National Coverage Determination 20.29, Hyperbaric Oxygen Therapy (version effective 04/03/2017).
- US Food and Drug Administration. Follow Instructions for Safe Use of Hyperbaric Oxygen Therapy Devices — Letter to Health Care Providers, 25 August 2025.
- US Food and Drug Administration. Product Classification: Chamber, Hyperbaric — Product Code CBF, 21 CFR 868.5470, Class II.