One Word, Two Devices: When Topical Oxygen Wears the Hyperbaric Label

Stylised illustration of two mismatched oxygen devices divided by a boundary line: a lit pressurised hyperbaric chamber with its gauge on one side, and a dim soft limb-encasing topical oxygen boot on the other

A patient with a diabetic foot ulcer is offered “hyperbaric oxygen” and goes home with a plastic boot. Another patient, with a clinically similar ulcer, is sealed into a chamber, breathes pure oxygen at two atmospheres for ninety minutes, and is watched by an attendant trained to spot a seizure or a perforated eardrum. Both have been told they are receiving oxygen therapy. Only one of them is receiving hyperbaric oxygen therapy.

It is the difference between a device category the Food and Drug Administration moved from class III to class II in 2011 — a limb-encasing device delivering humidified oxygen “at a pressure slightly greater than atmospheric pressure” — and a pressure vessel cleared through the 510(k) pathway under Product Code CBF and covered by a national coverage determination. The confusion has been formal enough that the Undersea and Hyperbaric Medical Society has issued position statements on it twice, in 2005 and 2018, and blunt enough that a Medicare contractor compressed the argument into one sentence of a coverage document.

This year the argument acquired fresh material. A review published this month in the Journal of Wound Care contends that cyclically pressurised topical wound oxygen — the boot-and-compression combination sold as TWO2 — produces not merely closure but durable closure, and should be adopted earlier in refractory wounds. That is a real claim backed by a real randomised trial. It also arrives in a country where a large Medicare contractor still classifies the category as non-covered, rates the evidence “poor quality,” and calls its own recommendation “weak.” Both things are true at once, and holding them together is the only honest way to read this literature.

What the name is doing

The UHMS statement, revised in May 2018 by Dag Shapshak and Matthew Kelly, is unusual for how explicitly it legislates vocabulary. Devices designed to increase external wound exposure to oxygen “should not be termed hyperbaric oxygen therapy,” because doing so “intentionally or unintentionally suggests that topical oxygen delivery methods are equivalent or identical to hyperbaric oxygen.” Study results for hyperbaric oxygen “should not be used to support topical oxygen therapy.” Topical oxygen lacks HBO2’s systemic effects. And the fourth conclusion has aged awkwardly: topical oxygen “may be a promising treatment based on some recent studies, but it cannot be recommended for routine clinical care at this time due to a restricted volume and quality of supporting scientific evidence.”

The naming problem is not hypothetical. The Medicare determination lists, by name, “TOPOX portable HBO extremity and sacral chambers” among examples of the category — devices marketed with the letters HBO on them while operating nowhere near hyperbaric pressure. The same document records the drift plainly: this delivery method “was known by some as topical HBO therapy, a term that is less commonly used today due to the low pressures delivered.”

The low pressures are the point. Hyperbaric oxygen, as the UHMS defines it, means the patient is entirely enclosed in a pressure vessel breathing oxygen at elevated pressure, typically above 1.4 atmospheres absolute. Topical devices operate at 1.004 to 1.013 atmospheres absolute. At those pressures the physics underwriting hyperbaric therapy — dissolving meaningful quantities of oxygen into plasma so it can travel past a diseased or occluded vessel — does not apply. The UHMS credits HBO2 with mobilising stem cells, altering cellular signalling, and restoring the oxygen content needed for angiogenesis and collagen formation. None of that is a claim about a boot.

What 2026 actually added

Set the vocabulary aside and the clinical question narrows: does blowing oxygen across a wound surface help it heal, and does the help last?

The new Journal of Wound Care review, led by John Steinberg, makes durability its organising idea — a defensible choice, because these ulcers recur, and closing a wound for three months only to lose it by twelve has not delivered much. The numbers it assembles are not trivial. In the double-blind randomised TWO2 trial, adjunctive therapy doubled the proportion of ulcers closed at twelve months, 56 per cent against 27, and roughly tripled twelve-week closure, 41.7 per cent against 13.5. In refractory venous ulcers, recurrence at 36 months ran at 5.9 per cent against 46.7 without; a cohort of 3,126 previously refractory wounds reported 64.8 per cent complete healing; a propensity-matched Veterans Affairs cohort showed 73 per cent fewer amputations and 82 per cent fewer hospitalisations.

Independent pooling points the same direction, less dramatically. George Theodorakopoulos and David Armstrong, reviewing only randomised trials with a twelve-week assessment, found three eligible studies covering 171 treated and 154 control patients. Topical oxygen alongside standard care increased complete closure, risk ratio 1.53 with a 95 per cent confidence interval of 1.14 to 2.05, heterogeneity low. Their Hartung-Knapp sensitivity analysis — the conservative one reviewers apply precisely when an evidence base is thin — widened the interval. They say so: “with few trials, uncertainty remains.”

That is a genuine signal. It is also, by their own count, three trials.

The trial that stopped a third of the way in

The largest of those three explains both the size of the effect and the width of the interval around it.

The TWO2 study, published in Diabetes Care in 2020 under Robert Frykberg, was a multinational, multicentre, double-blind, sham-controlled trial with a group sequential design and hard stopping rules at predetermined analyses of 73, 146 and 220 completed patients. At the first point, with 73 analysed, the active arm was already superior, and the trial stopped. Twelve-week closure was 41.7 per cent against 13.5, an odds ratio of 4.57 with a 97.8 per cent confidence interval running from 1.19 to 17.57. Adjustment for ulcer grade pushed the estimate to 6.00, interval 1.44 to 24.93.

A trial that stops early for benefit is not a flawed trial — that is what sequential designs are for, and continuing to randomise to a losing arm would raise its own ethical problem. But an effect measured at the first interim look, in 73 people, is an estimate, and early stopping systematically inflates effect sizes. An interval reaching from 1.19 to 24.93 says what the authors said: the direction is credible, the magnitude is not pinned down.

A 2026 network meta-analysis in the Journal of Clinical Medicine, pooling 27 randomised trials and 1,673 participants, found gas therapies as a class roughly doubled complete healing against standard care, risk ratio 2.17 — with heterogeneity of 75.7 per cent, a warning rather than a finding. When the network was ranked, standard care came out highest at a SUCRA of 93.9 per cent, with ozone and topical oxygen next and hyperbaric oxygen mid-range; the authors blame network structure and between-study variability, and recorded small-study effects on both Egger and Begg tests with certainty graded moderate. Their conclusion is the sentence worth keeping: “HBOT remains the modality supported by the broadest evidence base.”

A separate systematic review of 22 topical-oxygen studies found the benefit did not extend to everything measured. Case-control pooling favoured the therapy, odds ratio 4.48, but there was no significant effect on percentage wound-area reduction, on ulcer grade scores, or on pain — and most included studies were unblinded, single-arm, or case reports.

One national decision, many local ones

Whether any of this gets paid for depends on where the patient lives, and the reason is a quiet piece of administrative history.

Medicare once addressed topical oxygen inside its national coverage determination for hyperbaric oxygen treatment, NCD 20.29, as a separate section. In response to a reconsideration request seeking to remove the coverage exclusion for continuous diffusion of oxygen, CMS decided “that no NCD was appropriate concerning the use of topical oxygen for the treatment of chronic wounds,” amended NCD 20.29 by removing Section C, and handed the question to its local contractors. National policy did not become permissive; it became absent.

What fills the gap varies. Palmetto GBA’s determination L37873, effective in May 2019 and revised as recently as April 2026, is a non-coverage document. Its appraisal is unusually direct: quality, “poor”; strength of recommendation, “weak.” It notes that a randomised comparative trial exists and that results “may be promising,” but concludes that further trials are needed, that no population or wound characteristic predicting response could be identified, and that “the evidence is insufficient in determining the effects of this technology on net health outcomes.” It quotes the 2005 UHMS statement approvingly, including the line that the Society “does not recommend third party payor reimbursement” — a position the 2018 revision did not repeat in those words, and did not retract.

One supporting argument has since gone stale, and saying so is fair. The determination states that “there are no society guidelines that provide recommendations regarding the use of TOT.” The 2023 International Working Group on the Diabetic Foot guideline does now make a conditional supportive recommendation for topical oxygen in diabetes-related foot ulcers — and, in the same list and the same conditional register, for hyperbaric oxygen. Both are recommended only where best standard of care has failed and resources allow. That is no ringing endorsement of either, but the payer document’s claim about guideline silence no longer holds as written.

The asymmetry with hyperbaric oxygen is nevertheless sharp. NCD 20.29 names the conditions Medicare covers when a chamber is used — decompression illness, gas embolism, acute carbon monoxide intoxication, chronic refractory osteomyelitis, diabetic lower-extremity wounds meeting Wagner grade III and failing an adequate course of standard therapy, among others. There is no equivalent national list for a boot.

Where the oxygen actually goes

Underneath the coverage fight sits a mechanistic question nobody has settled, and the two sides are not arguing about the same evidence.

The payer determination asserts that “in vitro evidence suggests that TOT does not increase tissue oxygen tension beyond the superficial dermis.” Against that sits a human-skin model published in the Journal of Surgical Research in 2010, in which David Roe and colleagues interposed viable human skin between a topical oxygen source and a fluid-filled sensing chamber. Topically applied dissolved oxygen penetrated more than 700 micrometres of human skin, better through dermis than epidermis, and dissolved-oxygen devices outperformed gaseous ones. Seven hundred micrometres is not the deep compartment — but it is not nothing, and it undercuts the flat claim that topical oxygen never gets past the surface.

The strongest counter-evidence is a sham-controlled study the coverage determination itself cites. In cardiac surgery patients, insufflating oxygen locally into an occlusive dressing did not raise wound tissue oxygen tension at all — 24 against 25 millimetres of mercury, p = 0.88 — while raising the fraction of inspired oxygen raised both arterial and wound oxygenation. That is a specific, well-controlled negative result, and it should temper any mechanistic story.

The commercial context deserves the same scrutiny. The 2026 Journal of Clinical Medicine review of TWO2 discloses that one author is chief medical officer of AOTI, the device’s manufacturer, that three others received speaking and consulting honoraria from the company, and that one was an investigator in the TWO2 trial and received research support from it. The Journal of Wound Care durability review declares no conflicts of interest — and it is a narrative synthesis drawing heavily on one device family’s own trials that closes by urging “earlier, durability-focused adoption.” Its observation that “most meta-analyses pool mechanistically different oxygen devices” is a fair methodological point; its recommendation runs ahead of three randomised trials, however promising. None of this is misconduct. It is the ordinary texture of device research, and it is why pooled, independently conducted estimates matter more than any single enthusiastic review.

What a patient is actually choosing between

The regulatory split has a consequence that has nothing to do with efficacy. Hyperbaric oxygen therapy, because it involves a pressure vessel and a pressurised oxygen atmosphere, carries a fire risk serious enough that the FDA issued a device-safety letter in August 2025 stating that it is “aware of recent reports of fires that occurred with HBOT devices that resulted in serious injuries and deaths,” and reminding providers about training, continuous monitoring, prohibition of electrical and static-generating items, and hyperbaric-compatible cotton clothing. Topical devices, patient-administered at home at near-ambient pressure, do not carry that hazard — the coverage determination notes they “may be provided in the home setting by a well-trained patient or caregiver” — but they carry others: possible desiccation of an open wound, unsuitability for eschar-covered or deep penetrating wounds, and contraindication where perfusion is inadequate to support healing.

So the two categories differ in mechanism claim, in pressure, in FDA pathway, in payer treatment, in where they are administered, and in what can go wrong. What they do not differ in is the phrase printed on the brochure, which is the thing a patient is most likely to read.

The reasonable position, given all of it, is narrower than either camp would like. Topical oxygen is not hyperbaric oxygen, and the UHMS is right that the names must stay apart. It is also no longer accurate to call the evidence absent: three randomised trials, a sham-controlled study that stopped early, a durability signal no hyperbaric trial has matched for recurrence, and conditional guideline support in both directions. What remains absent is a large, fully recruited, independently replicated trial with blinded outcome verification and a standardised endpoint — the thing the 2018 UHMS statement asked for and the thing that would make a coverage determination easy to write. Until it exists, a patient offered “hyperbaric oxygen” for a wound is entitled to ask one clarifying question: how many atmospheres, in what device, and is anyone other than me in the room while it runs.

This article reports published research and regulatory documents. It is not medical advice, and decisions about wound treatment belong with a treating clinician. HBOToday’s position remains that the cleared indications for hyperbaric oxygen are those named in FDA device listings and CMS coverage policy; uses outside them — including wellness, longevity, and unproven neurological claims — are not FDA-cleared.

Sources

← Back to blog