Five Reports, One Research Group: the Thin Hyperbaric Record in Guillain-Barré Syndrome

Abstract illustration of a nerve fibre crossing a dark navy field, wrapped in pale beaded myelin segments with one segment stripped away so the bare axon glows amber

On 9 November the Journal of International Medical Research published a narrative review with a title that already contains its verdict. Yuan, Yang and Bu, neurologists at Zaozhuang Municipal Hospital in Shandong, set out to ask whether oxygen-tension biology gives anyone a reason to study hyperbaric oxygen in Guillain-Barré syndrome. Their answer is yes, and it is hedged in almost every sentence that follows.

The number doing most of the work is five. Five is the entire published record they could assemble: one case report of a mechanically ventilated patient, two case reports of acute motor axonal neuropathy, one small retrospective series, and one animal model. Four of the five come from a single research group in a single clinical setting, and none has been independently reproduced. The review calls oxygen-tension modulation “biologically plausible but remains unproven in severe or axonal Guillain-Barré syndrome”, then states that hyperbaric oxygen “should not replace intravenous immunoglobulin, plasma exchange, respiratory support, or rehabilitation.”

It is unusually disciplined for the field, and read closely it is an argument about what evidence is missing.

What the five reports actually contain

The ventilated case is the most striking, published in Diving and Hyperbaric Medicine in 2020 by Song and colleagues at the First Hospital of Jilin University. Their patient had severe Guillain-Barré syndrome and was ventilated inside a multiplace chamber at 220 kPa — 2.2 atmospheres — for 40 minutes, with 25 minutes each way for compression and decompression, in three courses of ten sessions separated by two-day rests. Motor function improved after the first session and ventilation stopped after eight. After thirty, the patient had normal breathing and speech, upper-limb strength of 4 on the Medical Research Council scale, lower-limb strength of 3, and went home. Song’s group then wrote the sentence that belongs above every hyperbaric case report: “studies are required to separate spontaneous recovery rates from treatment benefit.”

The other four trace to a group at the Naval Health Institute and Dr. Ramelan Navy Hospital in Surabaya. Two are case reports of acute motor axonal neuropathy, the Guillain-Barré variant in which the target is the axon rather than myelin. In the 2021 report in International Medical Case Reports Journal, a patient with progressive upper-limb weakness and difficulty breathing and swallowing had improved little by day five of conventional therapy; during hyperbaric treatment there were daily gains until day twenty, when the patient could walk slowly unaided. In the 2023 report in the Journal of Neurosciences in Rural Practice, immunoglobulin was given in a high-care unit and “an optimal recovery was not obtained” before hyperbaric oxygen was added. That paper concedes the regulatory position without flinching: “Although it has not been indicated for peripheral neuropathy.”

The animal work is the same group’s and is small: twenty-one New Zealand rabbits, in three groups, in which the neuropathy was induced and then treated with 2.4 atmospheres of 100% oxygen for 90 minutes daily for ten days. Glutathione activity rose significantly, interleukin-1β rose without reaching significance, and motor scores beat untreated animals. Plausible result, plausible mechanism, twenty-one rabbits.

Then the retrospective series, published in 2021 in the Open Access Macedonian Journal of Medical Sciences: the piece of human data clinics quote.

The Indonesian detail that reframes the biggest series

Twenty-five patients were treated with hyperbaric oxygen at the Lakesla facility between 2016 and 2019. Most were men in their twenties with a preceding diarrhoeal illness. First subjective change came three to ten days in; they walked with assistance after two to three weeks and without it after four to twelve. The authors conclude that hyperbaric oxygen “show[s] clinical improvement” and should become an adjunct in Indonesia.

The stated motivation is the important part. Standard therapy is immunoglobulin and plasma exchange, and the authors write, without embarrassment, that “in Indonesia, the availability and cost of these therapies are constraints.” Their aim was to show what hyperbaric oxygen could do in patients “who did not undergo standard therapy.”

Read that against the disease’s natural history and the series becomes much harder to interpret. In the 2024 Nature Reviews Disease Primers overview, roughly 20% of treated patients are still unable to walk after six months and about 5% die; the great majority recover substantially on a timeline measured in weeks to months. A single-arm series of twenty-five, assessed without a comparator, in a condition that improves over exactly the months being described, cannot separate a drug effect from the shape of the recovery curve. The Surabaya group never claimed otherwise; the citation of that paper in clinic marketing does.

There is a genuine moral argument buried here. Where immunoglobulin costs more than a family can pay, an inexpensive adjunct that might shorten disability is worth studying seriously. But that is an argument for a trial, and the opposite of an argument for a protocol.

The bar in this disease is low. Hyperbaric oxygen is still underneath it.

What makes Guillain-Barré a clean test case is that the surrounding literature is so well graded.

Plasma exchange has been under randomised study since the 1980s. The Cochrane review, last updated in 2017, pools six trials with 649 participants and reports moderate-quality evidence that exchange increases the proportion improving by at least one disability grade at four weeks, risk ratio 1.64, and reduces the need for artificial ventilation, risk ratio 0.53. Immunoglobulin, in the companion review covering twelve trials, hastens recovery as much as exchange in severe disease at the same certainty. The effects are modest, replicated, and consistent enough that both are standard care worldwide.

Now the instructive comparison. A Cochrane review published in 2020 examined every other drug tried in acute Guillain-Barré syndrome — interferon beta-1a, brain-derived neurotrophic factor, cerebrospinal fluid filtration, a Chinese herbal extract, eculizumab. Six randomised trials, five interventions, 151 participants. Every outcome was graded very low certainty, and the authors’ conclusion is worth quoting: “All six RCTs were too small to exclude clinically important benefit or harm from the assessed interventions.”

A complement-blocking monoclonal antibody, randomised in 41 patients, could not clear that threshold. Hyperbaric oxygen has never been randomised in this disease at all. The review’s most revealing admission is mechanical rather than statistical: “human studies of Guillain-Barré syndrome have not yet measured nerve-root oxygenation or microvascular changes before and after hyperbaric oxygen therapy.” The mechanism has never been demonstrated in the tissue it is proposed to help; what supports it is a rabbit model and evidence borrowed from other diseases.

Neither list includes it

The regulatory picture is not ambiguous, and hyperbaric claims tend to blur it.

The Undersea and Hyperbaric Medical Society’s Hyperbaric Oxygen Therapy Indications, fifteenth edition, lists seventeen numbered entries covering fifteen accepted indications. Searched in full, it contains no occurrence of Guillain-Barré, polyradiculoneuritis, neuropathy, or peripheral nerve.

Medicare’s National Coverage Determination 20.29 is narrower: reimbursement is limited to fifteen named conditions, from carbon monoxide intoxication and decompression illness through refractory osteomyelitis, osteoradionecrosis and diabetic wounds meeting three criteria. It then closes the door: “No program payment may be made for any conditions other than those listed in §270.4(A).” The explicit non-covered list names multiple sclerosis, arthritic diseases and chronic peripheral vascular insufficiency. Guillain-Barré syndrome appears on neither list, which for a bill means the same outcome as being named in the second one. The determination was last reviewed in April 2017 and carries a quarterly coding maintenance update dated October 2026.

The trial registry adds the third empty column: a search of ClinicalTrials.gov this week for Guillain-Barré syndrome with a hyperbaric intervention returns no records at all. The distance between a review arguing for a trial and a registered trial is the difference between a hypothesis and a research programme.

In this disease, a chamber is not a neutral addition

Most hyperbaric safety discussion concentrates on barotrauma and oxygen toxicity. Here the relevant hazard is different and it is documented.

A Mayo Clinic review of 187 patients treated between 2000 and 2017 found that 71 of them — 38% — had at least one manifestation of autonomic dysfunction: ileus in 42%, hypertension in 39%, hypotension in 37%, fever in 29%, tachycardia or bradycardia in 27%, urinary retention in 24%. Quadriparesis, bulbar weakness and mechanical ventilation were associated with it. Those patients had more cardiac complications, higher disability scores, and 6% mortality against 2% for the cohort.

Those are, almost by definition, the patients in whom an adjunct looks most urgently needed: the paralysed, the ventilated, the ones not recovering. They are also the patients whose blood pressure swings and whose heart rate destabilises, being asked to tolerate controlled pressure change and Valsalva against a closed glottis inside a multiplace chamber with a ventilator, monitoring lines and a support team.

The FDA’s letter to health care providers of 25 August 2025 covers the equipment side. It reminds facilities that hyperbaric devices are Class II, cleared through the 510(k) pathway under product code CBF, and that the agency “is aware of recent reports of fires that occurred with HBOT devices that resulted in serious injuries and deaths” while “the root cause of these events is not known.” Its recommendations are mundane: follow the instructions for use, ground electrical and static-generating equipment, dress patients in cotton rather than wool or synthetics. None of that makes hyperbaric oxygen impossible in critical care; it makes it an intervention that has to be engineered, monitored and justified rather than an add-on offered because the disease is frightening.

What would count as an answer

The review’s own prescription is modest and specific. Hyperbaric oxygen “should be tested only as an adjunctive intervention, with phenotype selection, standardized protocols, independent replication, blinded outcome assessment, safety monitoring, and a candidate biomarker panel whose analytical feasibility is established before it is considered a core panel.”

Every clause is a criticism of the existing literature. Phenotype selection means separating the axonal variant from the demyelinating form, because the four Surabaya papers all concern the former and the case series does not resolve the distinction. Standardised protocols confronts regimens running from 2.2 to 2.4 atmospheres, 40 to 90 minutes, ten to thirty sessions, with no head-to-head comparison anywhere. Independent replication means outside one institution in one country; blinded assessment addresses the obvious problem with case reports in which everyone knows when the chamber door closed.

The authors also applied SANRA, a six-item scale for narrative-review quality, to their own paper. It is a transparency gesture worth understanding precisely: SANRA scores how well a review explains its importance, aims, search strategy and evidence levels. It says nothing about whether the underlying evidence is any good — and here that evidence is five reports, four from one group.

The idea is not new either. The review’s bibliography reaches back to Soviet neurology of the late 1980s, where Neretin and colleagues reported that hyperbaric oxygenation in “polyradiculoneuritis” accelerated regression of neurological deficits, acted mainly on motor and autonomic-trophic disturbances, and allowed lower steroid doses and shorter hospital stays. Thirty-eight years and two continents later, the randomised question has still not been asked.

What patients actually get

The treatments that work are unglamorous and established: immunoglobulin or plasma exchange within two weeks of onset, intensive-care support when the respiratory muscles fail, rehabilitation that continues long after discharge. The review says none of this should be displaced, and its authors are right.

The useful thing about the paper is that it does not pretend otherwise. It assembles the record honestly, names the methodological failures, notes that nobody has measured the oxygen tension it theorises about in an actual patient’s nerve roots, and asks for a trial with blinded assessment and independent replication. That is what an investigational use looks like when described by people who intend to find out.

Which leaves the gap patients and families walk into. On one side sits a review recommending that hyperbaric oxygen be tested only as an adjunct, under protocol, with monitoring and a biomarker panel. On the other sits a chamber, a fee, and a disease that leaves people terrified and immobile for months. Everything between them — the consent conversation, the claim about what twenty-five Indonesian patients prove, the question of who is supervising — is where the difference between research and marketing gets decided. The literature in this condition, such as it is, does not decide it for anyone.

Sources

← Back to blog