The Seizure Was the Easy Part

$Abstract medical-illustration motif: a stack of lumbar vertebral bodies in teal and bone-white on a deep navy field, one vertebra crushed and split by an amber fracture line with a fragment pushed backwards into the spinal canal, faint concentric pressure rings in the corner$

The convulsion is the part everyone prepares for. A patient goes rigid inside a sealed steel or acrylic vessel, the oxygen is cut, the run is aborted, the seizure stops on its own, and the incident is written up as a near miss. In the November issue of Clinical Practice and Cases in Emergency Medicine, five emergency physicians describe a case in which the seizure did almost no harm and the body’s response to it did a great deal.

Their patient was a 35-year-old man with beta-thalassemia major who had been receiving hyperbaric oxygen for chronic venous ulcers of the legs. Mid-treatment he had a witnessed, first-time generalised tonic-clonic seizure. What followed over the next hours was severe low back pain and midline tenderness over the lumbar spine. Computed tomography showed an acute burst fracture of the third lumbar vertebra with retropulsion — bone pushed backwards into the spinal canal. He then developed progressive weakness of the right leg and acute urinary retention. Magnetic resonance imaging put the loss of vertebral height at 70 per cent, with severe narrowing of the thecal sac at L2-L3 and L3-L4 and a ventral dural haematoma compressing L4-L5. The diagnosis was cauda equina syndrome, and he went to theatre for emergent neurosurgical decompression.

What the report establishes, and what it leaves out

The abstract is unusually specific about imaging and unusually quiet about the treatment that preceded it. It does not report the pressure the chamber was run at, the gas schedule, which session this was, whether air breaks were in use, or how the seizure was managed inside the chamber. It does not say how long elapsed between the convulsion and the first scan, what operation was performed, or how the patient recovered. That is not a criticism of the authors, who wrote up what an emergency department can see. It is a caution for anyone tempted to read causation into the settings. What the paper does establish is that the fracture was acute: a fresh injury on CT, not an old deformity unmasked by a seizure. A vertebral body that had been sitting there, quiet, for years did not simply announce itself. It broke.

A burst fracture is not a compression fracture

The distinction sounds like radiology trivia and is not. Generalised convulsive activity loads the thoracolumbar spine along its axis while the paraspinal muscles contract violently, and the usual result, when there is a result at all, is an anterior wedge or compression fracture — common after seizures, frequently asymptomatic, and often found incidentally weeks later. A burst fracture is the same force taken one step further: the vertebral body fails in all directions, the posterior wall gives way, and fragments are driven into the canal that carries the cord and the nerve roots.

That is why the literature on seizure-related burst fractures reads less like a description of a fracture and more like a description of a missed diagnosis. In a 2020 paper in Epilepsy and Behavior Reports, Aaron Robichaud and colleagues at the State University of New York counted only eleven reported cases of thoracolumbar burst fracture after a generalised tonic-clonic seizure — six in the lumbar spine, five in the thoracic — and found that diagnosis was delayed in eight of the eleven. Deficits attributable to the fracture appeared in four. The authors added two cases of their own, both at L1, both initially missed: a 71-year-old whose leg weakness was put down to a possible myositis until he became incontinent on day six, and a 53-year-old whose ascending weakness was treated as suspected Guillain-Barré, including a trial of intravenous immunoglobulin, before imaging found the crushed vertebra. Notably, his scapulae were broken too, and nobody had noticed.

Their recommendation is unglamorous and specific. Every post-ictal patient should get a focused neurological examination of the lower limbs. If the patient cannot cooperate, spinal precautions should be maintained until the spine is cleared clinically or radiographically. And if the patient complains of musculoskeletal pain that might be coming from the spine, imaging is warranted — because untreated burst fractures are one of the few places in this field where the harm compounds by the hour.

The bone underneath

Beta-thalassemia major is not a neutral detail. Ineffective erythropoiesis drives marrow expansion, the cortex thins, and the skeleton pays for two decades of that. A 2021 systematic review and meta-analysis by Nipith Charoenngam and colleagues pooled 25 studies and 4,934 patients and put the prevalence of fracture in thalassemia at 16 per cent, rising to 17 per cent in beta-thalassemia and 18 per cent in transfusion-dependent disease, against 4 per cent in alpha-thalassemia. A 2025 review in the Journal of Endocrinological Investigation describes a condition that is multifactorial, must be managed from childhood and still lacks good trial evidence for the drugs used against it.

Where the fragility argument runs out

It would be tidy to file this case as a story about a fragile skeleton meeting an unlucky force. The seizure-fracture literature will not allow it. Robichaud and colleagues went through the eleven prior cases and concluded the opposite: three of the eleven were women, the age range ran from 20 to 54 with most between 35 and 45, and the authors stated plainly that osteopenia and osteoporosis did not appear to be a major causal factor. Their own two patients had no evidence of reduced bone density on imaging.

The mechanism does not need help. A generalised convulsion is a whole-body isometric contraction, and a vertebral body is a hydraulic structure being loaded far off its design axis. Fragile bone lowers the threshold; it is not the threshold. That matters for how the risk is pictured, because it means the population at risk is not the frail elderly patient in the corner of the ward. It is anyone in a chamber whose first seizure happens to occur there.

Why the diagnosis is usually late

Robichaud’s two cases are a catalogue of the reasons. Sedation and intubation make a neurological examination meaningless for the first day. An indwelling catheter conceals the urinary retention that would otherwise have been the loudest signal. A competing diagnosis — a meningioma in one case, suspected Guillain-Barré in the other — absorbs the clinical attention, and post-ictal weakness is easy to attribute to the seizure itself or to Todd’s paralysis.

A hyperbaric chamber adds its own version of every one of those problems. The patient is inside a pressure vessel, observed through a porthole or on a monitor, and cannot be examined properly until the run ends and decompression is complete. He has just had a convulsion, so he is expected to be sore and slow, and back pain after a seizure is common enough to read as background noise. The tell in this case was the combination the authors emphasise: acute spinal pain after convulsive activity, followed by a progressive neurological deficit.

What the safety material covers

The FDA’s August 2025 letter to health care providers about hyperbaric oxygen devices is the most authoritative safety document in the field and it is entirely about fire. Its recommendations run: follow the manufacturer’s instructions for use, ensure fire prevention measures, ensure proper grounding, train staff, monitor the patient for the whole treatment, follow cleaning and maintenance intervals, avoid prohibited electrical and static-producing items, and dress patients in hyperbaric-compatible fabric such as cotton. It states that the agency is aware of fires that caused serious injuries and deaths and that the root cause is not known, and it notes that while serious adverse events with these devices are believed to be rare, they can and do occur. The words seizure, convulsion, fracture and bone appear nowhere on that page.

That is not a gap the FDA created. Its remit in that letter is device safety, and a vertebral fracture caused by a patient’s own musculature is not a device malfunction. The Undersea and Hyperbaric Medical Society’s indications document is similarly framed around what the therapy is and when it works: physician prescription and oversight, a hard-sided chamber built to ASME PVHO-1 and NFPA 99, at least 2.0 atmospheres absolute, medical-grade oxygen for typically 90 to 120 minutes. Both documents are about doing the treatment correctly rather than about what to do with the patient an hour after it goes wrong.

And the seizure itself is genuinely the smaller risk. The largest published complication audit remains Plafki and colleagues’ prospective review of 782 patients and 11,376 sessions in a multiplace chamber at 240 to 250 kPa. It recorded four generalised seizures from central nervous system oxygen toxicity and noted that none of those four had any recognisable risk factor or prodrome — and that none suffered recurrences or sequelae. Ear discomfort affected more than 17 per cent of patients and verified barotraumatic lesions 3.8 per cent; the neurological event was the rare one, and it was survivable. What Plafki could not count was what happened to a body that convulsed while lying on a rigid surface.

The indication at the centre of it

There is an uncomfortable detail in the case report’s first sentence, and it is not clinical. The treatment was being given for chronic venous ulcers.

Medicare’s national coverage determination for hyperbaric oxygen therapy, NCD 20.29, lists fifteen covered conditions, and venous or stasis ulcers are not among them. They appear instead on the non-covered list, which reads, in the agency’s own words, that no program payment may be made for HBO in the treatment of “cutaneous, decubitus, and stasis ulcers,” and follows with the general rule that “no program payment may be made for any conditions other than those listed in §270.4(A).” Even the covered diabetic wound indication requires 30 days of standard wound therapy without measurable healing first, and payment stops if no healing appears within any later 30-day window.

The society and the payer do not agree here. The UHMS indications document does address venous leg ulcers, describing them as the most common lower extremity wound seen in ambulatory wound care centres and discussing hyperbaric oxygen as an adjunct alongside optimised local wound care, inside its entry on selected problem wounds. The randomised evidence underneath that position is thin: the Cochrane review of hyperbaric oxygen for chronic wounds found twelve trials and 577 participants, but ten of those trials enrolled diabetic foot ulcers, and exactly one, with sixteen participants, considered venous ulcers. It reported a reduction in ulcer area at six weeks and nothing at eighteen.

So the harm in this case landed in the least-evidenced, least-reimbursed corner of the field. That is not an argument that the patient should not have been treated; it is an observation about where risk and evidence sit relative to each other, and about how a rare catastrophic complication can attach itself to a use that the guidelines treat as marginal.

What one case can and cannot tell you

A single case report cannot produce a rate. There is no denominator here, no comparison group, and no way to know whether the pressure used, the number of sessions, the patient’s haematological disease or the design of the chamber table contributed. Journals publish the spectacular and not the unremarkable, so the published record of hyperbaric convulsions that ended in a fractured spine is certainly an undercount of the ones that ended in a sore back and an ordinary X-ray.

What it can do is document a chain of events that does not depend on the specifics of hyperbaric medicine to be plausible, because the intermediate step is a convulsion and convulsions break vertebrae in epilepsy units and electroconvulsive therapy suites too. The authors’ own conclusion is modest and worth repeating in their terms: physicians should keep a high index of suspicion for vertebral fracture and monitor for evolving neurological deficits when a patient has acute back pain after seizure activity. In a chamber, that means treating the end of the run as the start of an assessment rather than the end of an incident.

None of this changes what hyperbaric oxygen is good for. For decompression sickness, carbon-monoxide poisoning, gas gangrene, crush injury, compromised grafts and flaps, refractory osteomyelitis, late radiation tissue injury and the rest of the cleared list, it remains a standard of care with a complication rate that large audits describe as acceptable. The point of a case like this one is narrower and more useful: the field has spent years teaching itself to watch for the seizure, and comparatively little time thinking about what the seizure does to the skeleton of the person having it.

Sources

← Back to blog