Introduction
In 2007, an unusual scientific expedition travelled to Mount Everest. The objective was not simply to reach the world’s highest summit. Mount Everest was being used as a vast natural laboratory in which healthy human beings could be studied while progressively exposed to increasingly severe hypoxia.
The Caudwell Xtreme Everest (CXE) expedition systematically studied 222 healthy volunteers. The underlying research question had implications extending far beyond mountaineering:
Why do some human beings tolerate severe hypoxia remarkably well while others, particularly critically ill patients, deteriorate or die?
The investigators recognised an important opportunity. Hypoxia is central to many forms of critical illness, yet studying its effects independently is difficult because ICU patients are simultaneously affected by disease, drugs, mechanical ventilation, inflammation, circulatory abnormalities and multiple organ dysfunction.
Everest removes many of those confounding variables.
Healthy humans can be observed as environmental oxygen availability progressively declines, providing an opportunity to examine how respiratory, cardiovascular, haematological, microcirculatory and cellular systems respond.
The expedition therefore represented something unusual in medical research:
an investigation into critical-care physiology conducted on a mountain.
The Purpose of Caudwell Xtreme Everest
Caudwell Xtreme Everest was organised by researchers associated with University College London’s Centre for Altitude, Space and Extreme Environment Medicine.
Following baseline investigations at sea level, 198 trekker-subjects followed a standardised ascent towards Everest Base Camp at approximately 5,300 metres. Twenty-four investigator-subjects also reached Base Camp, with a smaller group subsequently conducting research higher on Everest.
Measurements included cardiopulmonary exercise testing, cerebral and skeletal-muscle oxygenation, neurocognitive testing, spirometry, blood biomarkers, arterial blood gases, cerebral blood-flow measurements, microcirculatory imaging and, in selected subjects, muscle biopsies.
Measurements were ultimately obtained at altitudes of up to 8,400 metres.
The investigators explicitly described their objective as understanding human adaptation to cellular hypoxia because cellular hypoxia represents an important mechanism of injury in critical illness.[1,2]
The mountain therefore provided a progressively more extreme physiological experiment.
Why Everest?
The percentage of oxygen in the atmosphere remains approximately 20.9% whether standing in London or on Mount Everest.
What changes dramatically is atmospheric pressure.
As altitude increases:
Barometric pressure ↓
which causes:
Inspired partial pressure of oxygen ↓
leading to:
Alveolar PO₂ ↓
then:
Arterial PO₂ ↓
and ultimately:
Available oxygen for tissue metabolism ↓
This is hypobaric hypoxia.
The enormous advantage for physiological research is that healthy subjects can be studied repeatedly as the severity of hypoxia progressively increases.
The process also occurs over days and weeks.
This allows something that most critically ill patients never receive:
time to acclimatise.
The Extraordinary Blood Gases of Everest
The most famous results from CXE came from arterial blood samples obtained from climbers at approximately 8,400 metres during descent from the summit.
The results were remarkable.
Among four subjects breathing ambient air:
Mean PaO₂: 3.28 kPa (24.6 mmHg)
with individual values ranging from:
2.55–3.93 kPa (19.1–29.5 mmHg).
Mean PaCO₂ was:
1.77 kPa (13.3 mmHg).
The lowest measured PaO₂ was therefore approximately 2.55 kPa.[3]
From the perspective of an intensive-care unit, these figures are extraordinary.
A critically ill patient with a PaO₂ approaching 3 kPa would ordinarily provoke immediate concern about profound respiratory failure.
Yet these individuals were conscious and functioning.
They had climbed at extreme altitude and were descending from Everest.
The obvious question is:
How can a human being remain functional with arterial oxygen tension this low?
The answer requires separating oxygen tension from oxygen content and, ultimately, oxygen delivery.
PaO₂ Is Not Oxygen Delivery
PaO₂ describes the partial pressure of oxygen dissolved in arterial blood.
It does not directly describe the total quantity of oxygen being transported.
Most oxygen is carried bound to haemoglobin.
Arterial oxygen content can approximately be expressed as:
CaO₂ = (Hb × 1.34 × SaO₂) + (0.003 × PaO₂)
while systemic oxygen delivery can be represented as:
DO₂ = Cardiac Output × CaO₂
This distinction is fundamental.
A patient can have a reasonable PaO₂ but inadequate oxygen delivery because haemoglobin concentration or cardiac output is insufficient.
Conversely, an acclimatised mountaineer can have an extremely low PaO₂ while compensatory mechanisms preserve useful oxygen transport.
CXE demonstrated this particularly well.
Although PaO₂ progressively decreased with altitude, haemoglobin concentration increased sufficiently for arterial oxygen content to remain at or above sea-level values until approximately 7,100 metres.
Only at extreme altitude did this compensation substantially fail.
At 8,400 metres, mean arterial oxygen content was approximately 26% lower than at 7,100 metres.[3]
The implications extend beyond altitude medicine.
A low PaO₂ is important.
But it does not, in isolation, describe oxygen delivery to the tissues.
Hyperventilation: A Remarkable Adaptation
The extraordinarily low PaCO₂ measured at 8,400 metres demonstrates another fundamental component of acclimatisation.
Hypoxaemia stimulates peripheral chemoreceptors, particularly the carotid bodies, increasing ventilation.
Therefore:
Hypoxia → increased ventilation → increased CO₂ elimination → PaCO₂ falls
Reducing alveolar CO₂ permits alveolar oxygen tension to rise relative to what it otherwise would have been.
The immediate consequence is respiratory alkalosis.
With continued exposure, renal bicarbonate excretion partially compensates for that alkalosis, facilitating continued hyperventilation.
The result at extreme altitude is physiology that would appear profoundly abnormal at sea level but is essential for survival on Everest.
A PaCO₂ around 1.8 kPa is not evidence of physiological failure in this situation.
It is evidence of extreme physiological compensation.
This introduces an important concept:
Normality is contextual.
From the Artery to the Cell
Critical-care monitoring naturally concentrates on variables that can be measured reliably:
- oxygen saturation,
- PaO₂,
- haemoglobin,
- blood pressure,
- cardiac output,
- lactate,
- acid-base status.
But none independently tells us whether every cell is receiving and successfully utilising adequate oxygen.
The complete pathway is considerably longer:
Atmosphere
↓
Lungs
↓
Arterial blood
↓
Haemoglobin
↓
Cardiac output
↓
Regional circulation
↓
Microcirculation
↓
Cell
↓
Mitochondrion
↓
ATP
The ultimate purpose of oxygen delivery is oxidative phosphorylation within mitochondria.
The monitor observes only selected points along this pathway.
The Mitochondrial Response
CXE researchers subsequently examined skeletal-muscle mitochondrial adaptation to prolonged hypoxia.
This produced another fascinating observation.
Adaptation was not simply a case of the body producing more mitochondria and consuming more oxygen.
Changes occurred in mitochondrial function and metabolic pathways during acclimatisation, consistent with the organism adapting energy metabolism to an increasingly oxygen-limited environment.[4]
This introduces a concept with potentially important implications for critical illness:
survival may depend not only upon how much oxygen reaches a tissue, but upon how efficiently that tissue uses the oxygen available.
The distinction is crucial.
Increasing oxygen delivery cannot necessarily correct a problem occurring at the level of cellular oxygen utilisation.
The Microcirculation
CXE also investigated the microcirculation using sublingual imaging.
Researchers demonstrated alterations in microvascular flow and vessel density during ascent.[5]
This matters because oxygen ultimately reaches tissues through capillary networks rather than through the large vessels represented by conventional haemodynamic monitoring.
An ICU patient can therefore theoretically have:
acceptable arterial pressure
- ●
acceptable cardiac output
- ●
acceptable arterial oxygen content
while still experiencing abnormal regional or microvascular oxygen delivery.
This is particularly relevant to sepsis and distributive shock, where microcirculatory dysfunction can become partly disconnected from apparently satisfactory systemic haemodynamics.
Restoring blood pressure does not necessarily prove that cellular perfusion has been restored.
The Parallel With Sepsis
The comparison with sepsis is particularly interesting.
A septic patient may have an adequate PaO₂.
They may have reasonable haemoglobin.
Their blood pressure may have been restored using vasopressors.
Their cardiac output may even be normal or elevated.
Yet organ dysfunction can continue.
Why?
Part of the explanation may involve abnormalities beyond the macrocirculation:
endothelial dysfunction,
heterogeneous microvascular blood flow,
impaired oxygen extraction,
inflammatory cellular injury,
and potentially
altered mitochondrial function.
This concept has sometimes been described as cytopathic hypoxia: oxygen may be available, but cellular utilisation becomes abnormal.[6]
Everest provides almost the reverse experiment.
The available oxygen becomes extraordinarily low, yet healthy cells progressively adapt to use what remains.
Everest Is Not ARDS
There is an important limitation to the comparison.
An acclimatised Everest climber and a patient with acute respiratory distress syndrome are not physiologically equivalent.
On Everest, the lungs are generally structurally healthy.
The problem originates outside the body:
Reduced atmospheric pressure → reduced inspired PO₂.
ARDS is fundamentally different.
Here the atmosphere contains perfectly adequate oxygen, but diseased lungs cannot transfer it normally:
Alveolar injury + inflammation + oedema + collapse + shunt + V/Q abnormalities → hypoxaemia.
The ARDS patient may simultaneously have infection, systemic inflammation, fever, shock, renal dysfunction, cardiac dysfunction and altered cellular metabolism.
The mountaineer has also reached extreme hypoxaemia gradually.
The ARDS patient may reach it within hours.
Consequently:
PaO₂ 5 kPa in an acclimatised climber ≠ PaO₂ 5 kPa in septic ARDS.
The laboratory value may be identical.
The physiological situation is not.
The Importance of Time
This may be one of the most important lessons Everest provides.
Acclimatisation occurs progressively.
During ascent, the body has time to alter:
ventilation,
acid-base regulation,
haemoglobin concentration,
vascular responses,
microcirculation,
metabolism,
and
cellular function.
Critical illness frequently provides no such opportunity.
A patient can deteriorate from relatively normal oxygenation to profound hypoxaemia within hours.
Haemorrhage can compromise oxygen delivery within minutes.
Septic shock can produce rapidly developing circulatory and metabolic abnormalities.
Pulmonary oedema can develop over a short period.
The physiological significance of hypoxia may therefore depend not merely upon:
how low the oxygen level is,
but also:
how quickly it became low.
The Survival Phenotype
One of the original purposes of CXE was to examine inter-individual variation in the response to hypoxia.[2]
This is perhaps the most intriguing parallel with intensive care.
People exposed to essentially the same altitude do not respond identically.
Some acclimatise exceptionally well.
Others experience substantial deterioration at much lower elevations.
Subsequent Xtreme Everest research involving Sherpa populations strengthened the concept that successful adaptation involves more than haemoglobin concentration alone. Sherpas demonstrated differences in microcirculatory behaviour and oxygen handling compared with altitude-naïve lowlanders.[7]
The parallel with intensive care is striking.
Two patients of similar age can develop apparently similar illnesses.
They can receive comparable treatment.
Their initial physiological measurements can look remarkably similar.
Yet one recovers rapidly while another progresses towards multiple-organ failure.
Why?
The answer may partly involve individual differences in:
genetics,
ventilatory response,
vascular regulation,
microcirculation,
oxygen extraction,
mitochondrial function,
and
metabolic efficiency.
CXE therefore raises the possibility of what might broadly be described as a hypoxia-tolerant phenotype.
Should We Always Normalise Oxygen?
This is where the implications for intensive care become more controversial.
CXE does not demonstrate that critically ill patients should be permitted to become profoundly hypoxaemic.
Nor does the ability of acclimatised mountaineers to survive with a PaO₂ around 3 kPa establish a safe oxygen target for ARDS.
That would be an inappropriate extrapolation.
What Everest does demonstrate is something subtler.
Human physiology can function considerably outside conventional sea-level reference ranges when adaptation has occurred.
This raises an important question about intensive-care medicine:
Should treatment always attempt to restore physiological variables to normal values, or should the objective sometimes be adequate rather than normal physiology?
There are consequences to aggressively pursuing apparently perfect numbers.
High inspired oxygen concentrations can produce hyperoxia and oxidative stress.
Excessive fluid administration can cause pulmonary and tissue oedema.
Blood transfusion carries risks and does not automatically guarantee improved cellular oxygen utilisation.
Excessive catecholamine exposure can increase myocardial workload and metabolic demand.
The objective therefore cannot simply be to maximise every component of oxygen delivery.
The objective must be to provide sufficient oxygen delivery for cellular metabolic requirements while minimising the harm created by the interventions used to achieve it.
Oxygen Supply Versus Oxygen Demand
Everest also illustrates another principle.
When oxygen becomes scarce, survival cannot depend indefinitely upon increasing oxygen supply.
The organism must also adapt its demand.
The physiological response therefore becomes a balance between:
oxygen acquisition,
oxygen transport,
oxygen distribution,
oxygen extraction,
and
oxygen utilisation.
This provides a different way of considering critical illness.
The question is no longer simply:
“How do we increase oxygen delivery?”
It becomes:
“How much oxygen does this patient actually require, where does it need to go, and can their cells use it?”
That is a substantially more sophisticated physiological question.
What Caudwell Xtreme Everest Actually Demonstrated
It is important to separate established findings from clinical extrapolation.
CXE demonstrated that:
1. Humans can remain conscious and functional despite extraordinarily low arterial oxygen tensions following progressive acclimatisation.
2. Increasing haemoglobin concentration can preserve arterial oxygen content despite major reductions in PaO₂.
3. Extreme hyperventilation and profound hypocapnia form an important component of successful high-altitude adaptation.
4. The microcirculation changes during exposure to progressive hypoxia.
5. Skeletal-muscle mitochondrial and metabolic adaptations occur during prolonged exposure to altitude.
6. There is considerable inter-individual variation in the ability to adapt to hypoxia.
What it did not demonstrate is that similar levels of hypoxaemia are safe during acute critical illness.
That distinction is essential.
From Everest to the Intensive Care Bed
The real value of Xtreme Everest lies not in providing an ICU oxygen prescription but in changing how we conceptualise oxygen physiology.
A pulse oximeter measures saturation.
An arterial blood gas measures oxygen tension.
Haemoglobin helps determine oxygen content.
Cardiac output contributes to global oxygen delivery.
But the mitochondrion ultimately determines whether oxygen can be converted into useful cellular energy.
Between the arterial blood gas and the mitochondrion lies an enormous physiological territory that routine monitoring only partially describes.
That territory includes:
regional perfusion,
microcirculatory function,
oxygen diffusion,
cellular extraction,
and
mitochondrial metabolism.
It may be within these poorly monitored processes that some of the differences between survival and organ failure are ultimately determined.
Conclusion
Caudwell Xtreme Everest transformed Mount Everest into an extraordinary laboratory for investigating one of medicine’s most fundamental problems:
What happens to the human organism when oxygen becomes critically scarce?
The arterial samples obtained at 8,400 metres remain one of the most dramatic demonstrations of human physiological adaptation ever recorded.
Climbers remained conscious and functional with a mean PaO₂ of approximately 3.28 kPa — a value that, encountered unexpectedly in an intensive-care patient, would indicate an immediate and extremely serious clinical situation.
But the lesson is not that oxygen does not matter.
It is almost the opposite.
Oxygen matters at every stage between the atmosphere and the mitochondrion.
PaO₂ represents only one stage of that journey.
Everest demonstrates what happens when healthy humans are given sufficient time to adapt their entire physiology to progressively diminishing oxygen availability.
Intensive care frequently confronts the opposite situation: sudden hypoxia imposed upon an organism already compromised by disease, inflammation, circulatory dysfunction and metabolic disturbance.
That difference may help explain why an acclimatised mountaineer can remain functional at oxygen tensions incompatible with survival in many acutely ill patients.
The enduring contribution of Caudwell Xtreme Everest is therefore not a new oxygen target.
It is a different question.
Instead of asking only:
“What is the patient’s oxygen saturation?”
perhaps critical-care physiology ultimately requires us to ask:
“Is sufficient oxygen reaching the cells, and are those cells still capable of using it?”
Everest was the laboratory.
The intensive-care patient was always the reason for the experiment.
References
1. Grocott MPW, Martin DS, Wilson MH, et al. Caudwell Xtreme Everest: a field study of human adaptation to hypoxia. Critical Care. 2007.
2. Levett DZH, Martin DS, Wilson MH, et al. Design and conduct of Caudwell Xtreme Everest: an observational cohort study of variation in human adaptation to progressive environmental hypoxia. BMC Medical Research Methodology. 2010;10:98. doi:10.1186/1471-2288-10-98.
3. Grocott MPW, Martin DS, Levett DZH, McMorrow R, Windsor J, Montgomery HE; Caudwell Xtreme Everest Research Group. Arterial blood gases and oxygen content in climbers on Mount Everest. New England Journal of Medicine. 2009;360:140–149. doi:10.1056/NEJMoa0801581.
4. Levett DZH, Radford EJ, Menassa DA, et al. Acclimatization of skeletal muscle mitochondria to high-altitude hypoxia during an ascent of Everest. FASEB Journal. 2012;26:1431–1441. doi:10.1096/fj.11-197772.
5. Martin DS, Goedhart P, Vercueil A, Ince C, Levett DZH, Grocott MPW. Changes in sublingual microcirculatory flow index and vessel density on ascent to altitude. Experimental Physiology. 2010;95:880–891. doi:10.1113/expphysiol.2009.051656.
6. Fink MP. Bench-to-bedside review: cytopathic hypoxia. Critical Care. 2002;6:491–499. doi:10.1186/cc1824.
7. Gilbert-Kawai E, Coppel J, Court J, et al. Studies arising from the Xtreme Everest programme examining Sherpa and lowlander responses to hypobaric hypoxia, including preservation of microcirculatory perfusion and oxygen unloading.
8. Grocott MPW, Levett DZH, Martin DS, et al. Caudwell Xtreme Everest: An Overview. Advances in Experimental Medicine and Biology. 2016;903:427–437. doi:10.1007/978-1-4899-7678-9_28.
