High-altitude cerebral edema: from stubborn headache to a race against a swelling brain in a fixed skull
High-Altitude Cerebral Edema (HACE) does not appear out of nowhere. It sits at the extreme, life-threatening end of a single continuous disease process — the AMS–HACE spectrum — that begins with an ordinary-seeming altitude headache. Recognizing the moment a "bad headache" becomes a warning sign is the first and most important triage decision in the mountains.
Acute Mountain Sickness and HACE are not separate diseases — they are stages of the same hypoxia-driven cerebral process. Both begin with the same triggers: rapid ascent to altitude, insufficient acclimatization time, and individual susceptibility (which is only weakly predicted by fitness level). At the mild end, AMS presents as headache plus nausea, fatigue, dizziness, or poor sleep — miserable, but self-limited if ascent stops.
At the severe end of the same continuum sits HACE: the same hypoxic cascade, but now with enough cerebral fluid accumulation to produce objective neurological signs. There is no sharp biological boundary between "bad AMS" and "early HACE" — the distinction is clinical, defined by the appearance of ataxia or altered mental status. This is why a headache that refuses to respond to rest, hydration, and standard analgesics must always be treated as a potential prelude to something worse.
The Lake Louise AMS Score defines AMS as a headache at altitude plus at least one of: GI symptoms, fatigue/weakness, dizziness, or difficulty sleeping, with a total score ≥3. A headache that persists or worsens despite analgesics and adequate rest is the single strongest predictor of progression toward HACE.
As barometric pressure falls with altitude, the partial pressure of inspired oxygen falls with it — at 4500 m, arterial oxygen saturation in an unacclimatized person can drop into the low-to-mid 80s (%), compared with ~98% at sea level. The brain, which consumes roughly 20% of total body oxygen despite being only ~2% of body mass, responds to this hypoxia with compensatory cerebral vasodilation to maintain oxygen delivery.
This vasodilation is protective in the short term but carries a cost: it increases cerebral blood volume and vascular wall stress at exactly the moment the vessel lining is already being biochemically stressed by hypoxia. Hypoxia-inducible factors upregulate vascular endothelial growth factor (VEGF), which loosens the tight junctions of the blood-brain barrier. This is the molecular hinge point — the same mechanism that produces a mild vascular headache in simple AMS is the mechanism that, if hypoxia continues, tips into frank fluid leakage and HACE.
The single most dangerous mistake in altitude medicine is treating a refractory headache as "just AMS" and continuing to ascend, or waiting to see if it resolves overnight. Standard AMS management — stop ascending, rest at the same altitude, hydrate, treat with ibuprofen/paracetamol and consider acetazolamide — is appropriate only while symptoms are mild and improving.
The moment a headache fails to respond to analgesics, or any new symptom appears (especially unsteadiness or confusion), the correct action is unambiguous: descend. Waiting is the variable that determines whether a climber walks down under their own power or is evacuated unconscious. Every stage that follows in this simulator is what happens when that decision is delayed.
Beneath a still-normal neurological exam, a dangerous process is already underway. Increased blood-brain barrier permeability allows fluid and plasma protein to leak into brain tissue, and because the skull is a rigid, fixed-volume container, that fluid has nowhere to go. This stage is silent to the patient but visible on imaging — and it is the last opportunity to reverse course before objective signs appear.
HACE edema is classified as vasogenic — it results from breakdown of the blood-brain barrier allowing protein-rich plasma fluid to leak from capillaries into the extracellular space of brain tissue. This differs mechanistically from cytotoxic edema (cellular swelling from failed ion pumps, seen in stroke) and from the high-altitude retinal and pulmonary counterparts, though all share the same root cause: hypoxia-driven capillary stress.
On MRI, this fluid characteristically accumulates in the white matter, and a strikingly consistent finding in HACE case series is edema concentrated in the splenium of the corpus callosum — a finding specific enough that it is considered a near-diagnostic imaging signature of the disease when available.
The Monro-Kellie doctrine, first articulated in the late 18th and early 19th centuries, states that the skull is a rigid, non-expansible container holding three components in volume equilibrium: brain tissue (~80%), cerebrospinal fluid (~10%), and blood (~10%). If one compartment's volume increases, one or both of the others must decrease to compensate, or intracranial pressure rises.
Early compensation works by displacing CSF into the spinal canal and reducing venous blood volume. But this buffering capacity is limited and roughly exponential in shape: pressure stays deceptively near-normal for a while as edema accumulates, then rises steeply once compensatory reserve is exhausted. This is precisely why HACE can seem to "suddenly" become severe — the intracranial pressure curve is not linear, and by the time it is rising fast, the safety margin is already gone.
Normal intracranial pressure is roughly 7–15 mmHg in a supine adult. Because the cranial vault cannot expand, even a modest absolute increase in brain water content can produce a disproportionate rise in pressure once compensatory CSF and venous displacement are exhausted — the defining danger of any intracranial process, HACE included.
The compensatory reserve described above means a climber at this stage may report only a persistent headache — the same complaint as simple AMS — while their intracranial pressure is already measurably elevated and edema is expanding on imaging. Gait, coordination, and mental status typically remain normal.
This is the central diagnostic challenge of field altitude medicine: there is no bedside test that reliably detects this silent phase. The only reliable signal available to a non-physician companion or guide is the trajectory of symptoms — a headache that is static or worsening despite rest and treatment should be assumed to represent this stage until proven otherwise, because by the time ataxia appears, the disease has already crossed into unambiguous HACE.
This is the stage that defines HACE. Simple AMS, however severe the headache, does not produce objective neurological deficits. The moment truncal ataxia or altered mental status appears in a person at altitude, the diagnosis changes: this is now HACE, a medical emergency, not a headache to be managed with rest.
The single most useful bedside test in high-altitude medicine is the tandem (heel-to-toe) gait test. The subject is asked to walk a straight line, placing the heel of one foot directly against the toe of the other with each step, ideally with eyes open and then closed. A healthy, acclimatizing person performs this without difficulty.
Failure — stepping off the line, wide stance to avoid falling, staggering, or inability to complete the sequence — is considered a positive test for ataxia and, in the context of altitude exposure, is sufficient by itself to diagnose HACE even without headache. Its power lies in its simplicity: it requires no equipment, can be performed by any team member in minutes, and directly probes cerebellar and proprioceptive pathways that are among the first to be affected by rising intracranial pressure and edema.
A positive tandem gait test in anyone who has been at altitude is HACE until proven otherwise. Guides and expedition protocols worldwide use this single test as the trigger for an immediate, non-negotiable descent order — no further observation period is required.
Alongside ataxia, altered mental status is the second defining feature of HACE. Early changes are subtle: slowed responses, difficulty concentrating, uncharacteristic irritability, poor short-term memory, or a flattened, apathetic affect that companions often describe as the person "not being themselves." These changes reflect rising intracranial pressure affecting cortical and subcortical function broadly, rather than any single localized deficit.
Because the affected individual's own judgment is compromised by the very process being assessed, self-report is unreliable at this stage — a hallmark and dangerous feature of HACE is that patients frequently underestimate their own impairment and resist evacuation. This is why expedition protocols assign a companion to make objective observations and hold pre-agreed authority to insist on descent regardless of the patient's objections.
HACE and High-Altitude Pulmonary Edema (HAPE) share the same underlying hypoxic trigger and frequently occur together — roughly half of HACE cases have concurrent HAPE. A climber presenting with ataxia should always be assessed for cough, dyspnea disproportionate to exertion, and reduced exercise tolerance, since coexisting pulmonary edema further compromises oxygenation and accelerates cerebral deterioration.
When both are present, the treatment priorities compound rather than compete: descent addresses both conditions simultaneously, oxygen benefits both, and nifedipine (for HAPE) can be added alongside dexamethasone (for HACE) without conflict. Recognizing the overlap matters clinically because pulmonary compromise can silently worsen cerebral hypoxia even while attention is focused on the neurological signs.
Once ataxia and altered mental status are established, HACE can progress with alarming speed. Confusion deepens into hallucination and disorientation, ataxia becomes severe enough that walking unassisted is impossible, and drowsiness can progress through stupor toward coma. This is the phase where delay converts a survivable emergency into a fatal one.
Recall the Monro-Kellie compensatory curve: pressure rises slowly at first while CSF and venous blood are displaced, then steeply once that reserve is exhausted. Progressive neurological decline is the clinical face of that steep part of the curve. Small further increases in cerebral edema now produce large increases in intracranial pressure, and rising pressure itself begins to compromise cerebral perfusion, creating a self-reinforcing cycle of worsening hypoxia and worsening edema.
Clinically this manifests as rapid, sometimes hour-to-hour deterioration: a climber who could stand with assistance in the morning may be unable to sit upright by afternoon. Hallucinations, severe disorientation, and inability to perform even simple commands reflect widespread cortical dysfunction, while increasing drowsiness reflects brainstem and reticular activating system compromise as pressure rises further.
Without intervention, progressive obtundation can proceed to stupor (arousable only with vigorous stimulation) and then coma. The lethal endpoint of unchecked intracranial pressure is brain herniation — displacement of brain tissue across rigid dural folds or through the foramen magnum, compressing the brainstem and its vital cardiorespiratory centers.
Historical case reports and expedition mortality data consistently describe a pattern: individuals who continue ascending or delay descent after ataxia onset can deteriorate to coma and death within as little as 12–24 hours, while individuals who descend promptly at the first neurological sign typically recover fully within days. The difference in outcome is determined almost entirely by the speed of the descent decision, not by any drug.
Case series of fatal HACE consistently show the same pattern: death occurs not because no treatment was available, but because descent was delayed — whether due to weather, terrain, summit-focused decision-making, or underestimation of the patient's own impairment. Time, not altitude reached, is the variable most tightly linked to survival.
A particularly dangerous feature of this phase is that the patient's own capacity for self-advocacy and risk judgment is being destroyed by the same process threatening their life. Confused or hallucinating climbers frequently resist help, insist they are fine, or even attempt to continue climbing — a phenomenon well documented in expedition medicine.
This is why pre-established, non-negotiable team protocols matter more than in-the-moment persuasion: expedition leaders and guide services train teams to treat a positive tandem gait test or altered mental status as an automatic descent trigger decided in advance, precisely because relying on real-time consensus with an already-impaired patient is unreliable and can cost the hours that separate survival from death.
HACE has exactly one definitive treatment: descent. Every other intervention — dexamethasone, supplemental oxygen, portable hyperbaric therapy — is supportive, buys time, or treats the swelling chemically, but none of them substitutes for reducing the hypoxic drive that caused the edema in the first place. Getting this hierarchy right, and executing it immediately, is what separates full recovery from disaster.
Immediate descent is the only treatment that addresses the root cause of HACE: hypoxia. Every hour spent at altitude while edema-generating hypoxia continues works against every other therapy given simultaneously. Guidelines are explicit that descent should not be delayed to administer other treatments — oxygen and dexamethasone should be given during descent or while evacuation is organized, never as a substitute for moving down.
The target is a minimum of 500–1000 m of vertical descent, or further until neurological signs clearly improve; some patients require considerably more. Even modest descent meaningfully raises inspired oxygen partial pressure and can produce noticeable clinical improvement within hours — the reverse of how quickly symptoms can worsen with continued ascent.
Unlike simple AMS, where rest at the same altitude may be sufficient, HACE management guidelines are unambiguous: descent cannot be substituted with rest, and it cannot be delayed to see if drugs work first. Descent begins immediately, with other treatments layered on top of it.
Dexamethasone is a potent corticosteroid that reduces cerebral vasogenic edema, most likely by stabilizing the blood-brain barrier and reducing vascular permeability, along with general anti-inflammatory effects. The standard regimen is an 8 mg loading dose (oral, IM, or IV depending on the patient's ability to swallow and available equipment), followed by 4 mg every 6 hours.
Critically, dexamethasone treats the swelling but does nothing to improve the body's acclimatization to hypoxia — it does not increase oxygen delivery or correct the underlying hypoxic drive. A patient who feels better on dexamethasone and is tempted to continue ascending remains just as physiologically unacclimatized as before, and symptoms can rebound sharply once the drug wears off if descent has not occurred. Dexamethasone buys time and improves the odds of the descent going smoothly; it is a bridge, never a destination.
Supplemental oxygen directly raises arterial oxygen saturation, reducing the hypoxic drive behind cerebral vasodilation and edema formation, and is given at 2–4 L/min or higher as available during descent and evacuation. It is fast-acting and complements dexamethasone's slower anti-edema effect.
When actual descent is impossible or must be delayed — due to darkness, severe weather, avalanche risk, or impassable terrain — a portable hyperbaric chamber (commonly known by the brand name Gamow bag) is the standard bridge therapy. The patient is sealed inside and the bag is pressurized by a foot pump, simulating a descent of roughly 1500–2500 meters in equivalent barometric pressure. Sessions typically last 1–2 hours and can produce significant symptomatic improvement, but relief is temporary: symptoms often recur once the patient is removed from the bag, and actual descent must still occur as soon as conditions allow. The Gamow bag has saved lives in situations where evacuation was delayed by hours to days, but it has never been shown to eliminate the eventual need for descent.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Immediate Descent | 500–1000 m minimum, or until improved | Restores inspired oxygen partial pressure, reverses hypoxic drive at its source | Definitive — does not substitute for anything else |
| Dexamethasone | 8 mg load, then 4 mg q6h | Stabilizes BBB, reduces vasogenic edema and inflammation | Does NOT aid acclimatization — descent still required |
| Supplemental Oxygen | 2–4+ L/min during descent/evacuation | Directly raises arterial SaO2, reduces cerebral vasodilation | Fast-acting bridge — not a substitute for descent |
| Portable Hyperbaric (Gamow) Bag | 1–2 h sessions, ~1500–2500 m simulated descent | Pressurizes patient environment to mimic lower altitude | Bridge only when real descent is impossible — symptoms recur after |