The FDA's Own Hyperbaric Ledger: 25 Reports in 2025, Four Deaths, and a Clearance Pipeline That Has Gone Quiet

Technical illustration of a bolted hyperbaric chamber porthole end-cap in brushed steel with a dark glass window, a pale pressure gauge with a red needle at upper left, and a rising bar chart of report counts at lower right above amber hazard striping

A woman could not reach her husband by phone, so she drove to his health clinic and found the building filled with smoke. Emergency responders arrived at about 10:50 that evening in August 2025 and concluded that a flash fire had occurred while he was inside a sealed hyperbaric chamber at 2.2 atmospheres, treating himself. The chamber held. The man did not. Inside it, investigators found the debris of a handheld communications device, a headset, earphones, a brain-tapping gadget, battery power banks, a phone, a tablet, charging cables, a fan, and possibly a red-light therapy panel; blackened lumps that looked like swollen lithium-ion packs were suspected as the ignition source.

That account is not a newspaper story. It is a narrative field in the Food and Drug Administration’s own adverse-event database, filed under product code CBF, the code that covers hyperbaric chambers, and publicly searchable in about a minute. Written in the flat, redacted, blameless register these filings use, it is one of 25 hyperbaric-chamber reports the agency logged in 2025 — more than in any other year the database covers — and one of four that were classified as deaths.

What the database is, and what it counted

The FDA’s Medical Device Reporting system obliges manufacturers to tell the agency when a device may have caused or contributed to a death or serious injury, or has malfunctioned in a way that would do so if it recurred. Those reports land in MAUDE, the Manufacturer and User Facility Device Experience database, and from there into openFDA, where anyone can query them by product code. For hyperbaric chambers the code is CBF, the classification is Class II, and the governing regulation is 21 CFR 868.5470.

Run that query and 157 reports come back, stretching from 1992 to 13 October 2026. Ninety-six are malfunctions, 42 injuries, 12 deaths, with a handful unclassified. The distribution over time is the part worth pausing on: 13 reports in 2023, 13 in 2024, and then 25 in 2025. Through the first ten and a half months of 2026 there are 13 more — 38 in 22 months, against 26 in the two years before them.

There are honest ways to deflate that number. MAUDE is a complaint box, not a verdict: a filing records an allegation, frequently redacted beyond usefulness — the patient above is “(B)(6)” throughout — and the agency does not verify the narratives before publishing them. There is no denominator anywhere in the database, no count of treatments delivered in the same period, so a rising count cannot be converted into a rate. And 2025 is the year the FDA published a safety letter about these devices, which tends to make manufacturers report more carefully. A reporting artefact is a real possibility, though not a complete explanation: many of the 2025 filings describe events that happened that same year, and a clinic that never reported before has no backlog to clear.

Two deaths inside a Fortius 420

The August record was not the first. A filing dated 15 July 2025 states, in its opening line, that “a male clinic founder died in a flash fire inside an OxyHealth Fortius 420 hyperbaric chamber during self-treatment,” and that the fire’s cause was still under investigation. The same filing notes that the man had recently completed a safety and certification course run by the International Hyperbaric Association and the International Board of Undersea Medicine, a non-profit affiliated with the chamber’s manufacturer. It then attaches photographs, and makes a claim the agency has not adjudicated: that the chambers carried the FDA logo on the exterior control panel and on an interior wall, which the filer says does not appear to be authorised and may violate federal law.

A third record, dated 6 February 2025, is a single sentence from someone who does not know which device was involved: “The complainant alleges that a child was killed as the result of a hyperbaric chamber fire.” A fourth, from 31 January 2025, concerns a Sechrist 3300HR chamber and records both a death and an injury, with the note that the incident was under investigation by officials in Troy, Michigan, and that the manufacturer had not yet been formally notified.

None of these filings establishes a cause, and none should be read as a finding of fault. What they establish is a pattern of setting: self-treatment, a clinic chamber rather than a hospital suite, and — in the August record — a sealed vessel at 2.2 atmospheres occupied by a man surrounded by consumer electronics. The FDA’s own position, in its August letter, is that “the root cause of these events is not known,” while adding that the agency believes serious adverse events with these devices are rare.

The parts that fail are the parts that save you

The bulk of the database is duller and, in a way, more useful: it is a maintenance ledger. The reported failures cluster on the systems that exist only for emergencies.

A customer found the emergency vent toggle “barely responding” during daily start-up. A technician opened the emergency vent button and found the three-way valve sticking, so that decompression needed four or five pushes to begin. On a chamber more than twelve years old, emergency decompression was found out of specification and traced to a clogged flow restrictor orifice. A door locking pin stopped engaging until 4.2 psi and began retracting at 7.0, caused by a bent safety-block nipple that a gurney had struck. A chamber reached treatment pressure in four minutes instead of the normal seven, and a patient came out with an ear injury. At another facility, multiple patients developed barotrauma after the chamber was serviced by a third party. In one Canadian case a patient suffered cardiac arrest during treatment, reported to the manufacturer a month later.

Sechrist appears in roughly 22 of the 38 reports filed since January 2025, and attributes almost every one of them to component wear or user error, repeatedly stating that no manufacturing non-conformity was found and no corrective action was necessary. That is a defensible position for a company whose devices are expected to last decades, and the concentration under its name reflects market share and reporting diligence as much as anything else. But the aggregate effect is a fleet in which the safety-critical hardware is ageing, the emergency systems are the failure point, and the recorded remedy is a service visit.

That matters more than it would in another device class, because the fleet is not being renewed.

Sixty-nine clearances, and two this decade

The FDA’s 510(k) premarket notification database lists every hyperbaric chamber cleared for marketing under product code CBF. As of its snapshot on 2 November 2026 there are 69 of them: one from the 1970s, nine from the 1980s, twenty from the 1990s, twenty-nine from the 2000s, eight from the 2010s, and two from the 2020s. The most recent is K240569, cleared on 21 November 2024 for Fink Engineering’s FESL, FEDL and FETL chamber family. Nothing in 2025. Nothing in 2026.

Read carefully, that is less sinister than it sounds. A hyperbaric chamber is a pressure vessel expected to serve for twenty or thirty years, the American market is small, and a clearance covers a device family rather than a unit. Nor does a 510(k) mean much clinically: it is a finding of substantial equivalence to a predicate device, not a demonstration of efficacy, and it says nothing about whether anyone should get in.

The detail that does carry weight is quieter. Two of the clearances in the last seven years belong to the Revitalair line — the 430F in 2019, the 430+ in 2023 — a mild-pressure chamber sold into wellness settings. The pathway therefore exists for exactly the kind of device that dominates the consumer market, and the agency’s letter even names the way to check: cleared devices can be identified by searching product code CBF in the 510(k) database. What that search cannot find is not cleared, whatever a website says. “FDA registered” is not “FDA cleared”; a facility registration is paperwork, not permission.

What the August letter does, and does not do

The FDA’s letter of 25 August 2025 is seven recommendations long. Follow the manufacturer’s instructions for use. Ensure fire prevention and safety measures, and note the heightened fire risk of high-concentration oxygen. Ground equipment properly. Train staff. Monitor the patient for the entire treatment. Keep to the manufacturer’s cleaning, maintenance and safety-check intervals. Avoid prohibited items, including electrical and static-generating devices, and dress patients in hyperbaric-compatible fabrics such as cotton rather than wool or synthetics. It closes with two references: NFPA 99-2024, the health care facilities code, and ASME PVHO-1-2023, the standard for pressure vessels for human occupancy.

What the letter does not do is recall anything, mandate a design change, or impose a new reporting threshold. It is an advisory, addressed to “health care providers and facilities” — a phrase that fits a hospital hyperbaric unit and, by any reasonable reading, a clinic in a strip mall. That mismatch is where the data gets thin. A hospital is a user facility with defined reporting obligations; a wellness business selling sessions in a soft-sided chamber may sit outside that definition entirely, and its incidents may reach the database only when a patient, a relative or a fire marshal writes in. Several of the 2025 and 2026 records are exactly that: first-person filings by injured patients or their spouses, not manufacturer reports. The 25 reports for 2025 should be read as a floor.

Chambers in houses

The clearest signal in the recent records is the disappearance of the attendant. In a hospital multiplace chamber an inside attendant rides with the patient — this year’s literature in Diving and Hyperbaric Medicine is still working out how to schedule that attendant’s oxygen periods so emergency decompression stays possible late in a session — and a monoplace chamber has a nurse at the console, watching. Remove both and what remains is a pressure vessel with a user manual.

The records show what that looks like. A nurse practitioner treating her own mastectomy wound in a chamber in her home reported that the door blew off at nearly 2 atmospheres, injuring her leg; she told the filer she wanted to add the modality to her office practice. A man became trapped alone inside a chamber in his own house and had to telephone the manufacturer from inside, enter administrative passwords and recalibrate the pressure before the door would release. A patient who took unlimited 90-minute sessions daily for three months reported headaches and lethargy every time she came out. In October 2025 a husband described his wife being sealed into a chamber while the technician on duty failed to start the oxygen flow, with a dead walkie-talkie, no monitoring in the room and no licensed clinician on site; she was found by coincidence. That record is a complaint, not an inspection. Read plainly, it is also a description of a business running medical equipment without any of the systems that make the equipment survivable.

Where the record is strong, and where the marketing is not

None of this argues against hyperbaric oxygen therapy. It argues for knowing which thing you are looking at, because the regulated version has a real evidence base and a real coverage framework. The Undersea and Hyperbaric Medical Society’s fifteenth-edition indications list runs to fifteen accepted conditions, from air or gas embolism and decompression sickness through carbon monoxide poisoning, crush injury, compromised grafts and flaps, refractory osteomyelitis, delayed radiation injury and sudden sensorineural hearing loss. Medicare’s National Coverage Determination 20.29 covers a comparable list with conditions attached: diabetic wounds, for example, only as an adjunct after a month of failed standard care. The evidence behind those indications is uneven but not imaginary — the 2015 Cochrane review of chronic wounds pooled 12 trials and 577 participants and found healing improved at six weeks, while the reduction in amputations did not reach significance, and middle-ear barotrauma remains the commonest side effect at roughly 5 to 10% of patients treated.

The UHMS says something about the other half of the market that is worth quoting, because it is harsher than anything the agency has published: mild chambers in wellness centres and spas are “usually not physician-prescribed or supervised,” often sited in commercial buildings without fire-safety or chamber-construction compliance, “putting those exposed at risk for serious injury and even death.”

Which leaves the question a reader can actually use, and it is not whether hyperbaric oxygen works. It is whether the vessel in front of you is cleared under product code CBF, whether the pressure and gas mix match a published protocol for a named condition, who is supervising, whether anyone is monitoring the patient for the whole session, and when the emergency vent was last tested. The answers are public, or should be. The FDA’s database is imperfect, redacted, unverified and missing a denominator — and it is still the only ledger of this industry that nobody in it writes.

Sources

← Back to blog