The human physiological impact of global deoxygenation
There has been a clear decline in the volume of oxygen in Earth’s atmosphere over the past 20 years. Although the magnitude of this decrease appears small compared to the amount of oxygen in the atmosphere, it is difficult to predict how this process may evolve, due to the brevity of the collected records. A recently proposed model predicts a non-linear decay, which would result in an increasingly rapid fall-off in atmospheric oxygen concentration, with potentially devastating consequences for human health. We discuss the impact that global deoxygenation, over hundreds of generations, might have on human physiology. Exploring the changes between different native high-altitude populations provides a paradigm of how humans might tolerate worsening hypoxia over time. Using this model of atmospheric change, we predict that humans may continue to survive in an unprotected atmosphere for ~3600 years. Accordingly, without dramatic changes to the way in which we interact with our planet, humans may lose their dominance on Earth during the next few millennia.
KeywordsOxygen Hypoxia Acclimatization Physiological adaptation
Human dominion over planet Earth is driving profound changes that may culminate in extinction. Loss of natural vegetation and the burning of fossil fuels are altering our atmosphere at an alarming rate . Two interconnected themes have received the most attention: the accelerated rise in atmospheric carbon dioxide concentration and the escalation of global temperatures. These changes are accompanied by natural phenomena with potentially catastrophic consequences, such as increasingly unpredictable climate subsystems and rising sea levels from polar ice cap recession [2, 3, 4]. If such environmental hazards were not a sufficient threat to the survival of Earth’s 7 billion plus human inhabitants, there is yet another concerning change already underway, global deoxygenation. Although the current volume of oxygen in our atmosphere is vast, it is diminishing inexorably, and yet this does not appear to be a priority for environmental concern. While the dynamics of oxygen decline is highly contentious, a new model of its nonlinear nature predicts total oxygen depletion within several thousand years . This disruption of Earth’s fragile ecosystem could be the final straw for humans and the many other forms of life that rely on oxygen to generate energy. Here we discuss the biological significance of atmospheric oxygen, the proposed model of its decline, and its potential impact on human survival.
Oxygen and Earth’s atmosphere
The decline in global oxygen concentration
From the original Scripps Institute data, it was hypothesized that the decline in atmospheric oxygen concentration was linear, however, as the dataset grows, it has been proposed that this may not be the case. One possible scenario is a parabolic decline . Because the observed records span only a few decades, projection of this decay is highly uncertain and the complexity of the biogeochemical interactions makes such a projection a challenging task. However, this new mathematical model proposes that the horizon of oxygen decline will be reached much sooner than previously estimated from a linear model, leaving only a few thousand years until total oxygen depletion . If we apply this parabolic projection to the currently available observed data, it predicts that the concentration of oxygen in the atmosphere will reach zero in ~4400 years, passing beyond the halfway point in ~3600 years (Fig. 2).
Why is atmospheric oxygen concentration declining?
Biological impact of atmospheric deoxygenation
The biological effect of a gas is determined by its partial pressure, which, according to Dalton’s law, is equal to the product of barometric pressure and the fractional concentration of the gas in the mixture (Eq. 4). For example, at sea level, PatO2 is ~21 kPa (sea level barometric pressure 101 kPa × fractional concentration of oxygen 0.21). Thus, atmospheric hypoxia may result from either a decline in oxygen concentration (normobaric hypoxia) or a reduction in barometric pressure (hypobaric hypoxia). Differences in the biological responses to these situations are subtle but not completely insignificant . For humans, the principal consequence of a fall in PatO2 is hypoxemia (a lack of oxygen in the blood), resulting in reduced delivery of oxygen to the tissues (tissue hypoxia). This occurs during ascent to high altitude, due to the exponential decline in barometric pressure. Above ~1500–2500 m, depending upon the individual, hypoxemia can lead to altitude-related illnesses, such as acute mountain sickness (AMS). Hypoxemia and tissue hypoxia can also result from many pathophysiological states that impair oxygen transport, such as respiratory and cardiovascular diseases. Humans have the ability to adapt to hypoxemia, through a process known as acclimatization, but the extent to which adaptation can compensate for the oxygen deficit depends on the magnitude of the deficit, and the time over which it occurs.
Adaptation to acute hypoxia
No clear definitions exist to define time-related exposures to hypoxia, but attempts have been made to unify the language used . A significant and abrupt fall in PatO2 cannot be tolerated for more than a few minutes before cerebral hypoxia results in unconsciousness. Descriptions of death following sudden oxygen deprivation were common amongst early high-altitude aviators during World War II . The “time for useful consciousness” on sudden exposure to a simulated altitude of 7620 m above sea level (PatO2 ~8 kPa) is ~4–5 min) . The rapidity of the fall in PatO2 prevents any meaningful adaptation beyond hyperventilation and tachycardia, a desperate attempt to increase circulating blood oxygen levels. This almost immediate physiological change is brought about through oxygen sensing in the carotid bodies and their subsequent effect on the respiratory and cardiovascular centers of the brain.
Adaptation to subacute hypoxia
Lifelong hypoxic exposure
Chronic disease or long-term residence at high altitude can expose humans to a lifetime of hypoxemia. ~140 million people live permanently at high altitude (conventionally defined as >2500 m: the elevation that most people demonstrate a drop in the oxygen saturation of hemoglobin, SpO2) ; and whilst this is compatible with life, it is not without consequences for many. With increasing altitude of residence, chronic mountain sickness (CMS) and intrauterine growth retardation (IUGR) become more prevalent and an acceleration of pre-existing chronic respiratory diseases is observed . Global atmospheric deoxygenation would lead to such complications being encountered at progressively lower elevations, and we would expect their incidence and severity increase on a global scale. CMS is characterized by excessive polycythemia, progressing to pulmonary hypertension, right ventricular failure, and death . There is limited information about the frequency of and mortality from this disease in the present day as it is not a recognized classification in death certificates , but it will undoubtedly generate a significant and progressive disease burden as deoxygenation continues. The impact on human reproduction may have even more grave ramifications for population expansion and health. Fetal hypoxia, due to reduced maternal oxygenation and uteroplacental blood flow, reduces birth weight by an average of 100 g for every 1000 m above sea level [20, 23]. IUGR has wide-ranging and severe consequences throughout life. Low birth weights are linked to higher mortality in infancy, childhood, and later in life. Late (adult) morbidity and mortality may be due to the heightened risk of systemic hypertension, coronary heart disease, and diabetes observed in low-birth weight groups . Decreasing PatO2 at altitude is also associated with an increased prevalence of pre-eclampsia, a syndrome of maternal hypertension and proteinuria, which can progress to life-threatening seizures, as well as IUGR . Chronic lung disease follows an accelerated course in hypoxic environments, resulting in a shorter interval between onset and death, and further increasing the incidence of right heart failure . In short, lifetime exposure to low PatO2 exerts detrimental effects that may limit longevity, increase morbidity, and impair human reproduction. However, some populations have thrived at altitude and perhaps the survival of future generations of humans depends on the long-term adaptations observed in these people.
Adaptation to hypoxia over generations
Populations that have occupied hypoxic environments for hundreds of generations appear to have undergone genetic adaptation leading to the expression of phenotypes that convey an enhanced ability to survive and reproduce under chronic hypoxic stress. Long-resident populations enjoy reduced incidence and severity of the high-altitude complications such as CMS and IUGR  compared to non-ancestral high altitude residents, and their superior physical performance at altitude is widely reported anecdotally and demonstrated by increased maximal oxygen consumption on exercise testing [27, 28].
Hypoxic ventilatory response
Similar to sea level (high)
Arterial oxygen saturation
No increase (up to 4000 m)
No increase (up to 4000 m)
Pulmonary arterial pressure
Nitric oxide levels
Many approaches have been applied to uncover the genetic basis of these hypoxia-adapted phenotypes. High-altitude populations appear to have undergone positive selection in many genes that are involved in the HIF signaling cascade, which co-ordinates the cellular and systemic response to hypoxia. Examples of such genes are summarized in Fig. 4. In most instances, the precise function these genetic variants is yet to be revealed, but in some instances, putative mechanisms are beginning to emerge. For example, a variant of the EPAS1 gene (which encodes the alpha subunit of the HIF-2 transcription factor) has been demonstrated at increased frequency in high-altitude Tibetans. The selected variant actually down-regulates HIF targets, including erythropoietin, and is associated with lower hemoglobin concentrations . It has thus been proposed that it may promote survival in hypoxic conditions by protecting against CMS and improving microcirculatory flow and local oxygen delivery due to reduced blood viscosity. The Tibetans inherited this gene from an ancient human race called the Denisovans, prior to their extinction 40,000 years ago . It has been identified in only one other population on Earth, the Han Chinese, from which the Tibetans split less than 3000 years ago. In this time, the frequency of the gene in the two populations has diverged significantly: it is present in only 9% of Han but in 87% of Tibetans, the fastest known example of Darwinian evolution of humans . The timeframe over which this significant genetic population change occurred is roughly equivalent to the time over which the model predicts global PatO2 to halve, and offers some insight into how quickly humankind might be able to adapt to the oncoming hypoxic selection pressure.
Any predictions about the nature of the human race in an oxygen-deplete future using the genetics of present-day high altitude populations is hampered by the fact that different genes appear to have undergone positive selection in each, with no overlap in the variants expressed by the Tibetans, Andeans, or Ethiopian highlanders . Genomic analysis of Andean populations has revealed at least 40 candidate genes involved in the HIF pathway or hypoxia-related genes, including PRKAA1 (which codes for a subunit of adenosine monophosphate-activated protein kinase, and may influence fetal growth) [33, 34]. Natural selection in many genes involved in the same pathways has been demonstrated in high-altitude Tibetans, but the specific genes are at different loci or constitute different variants, such as EGLN1 (which encodes prolyl hydroxylase 2, the oxygen-dependent modulator of the HIF alpha subunits ). Ethiopian highlanders show positive selection in different genes again, this time including BHLHE41, which may be both a target and a modifier of HIF-1 alpha [33, 36]. One possibility is that different populations have followed different paths towards hypoxic adaptation, influenced by other environmental variables in each location (such as temperature or food availability), and population factors such as the genetic variation in the original settlers (contributing to genetic drift) and access to other gene pools (contributing to genetic flow). A second explanation is that they represent different time points on the same journey towards an optimally adapted phenotype, with duration and degree of hypoxic exposure different in each region. If we accept the second explanation, then the Tibetans, exposed to the greatest degree of hypoxic stress for the longest time, would represent the current pinnacle of long-term hypoxic adaptation. This is corroborated by the fact that Tibetans have a lower incidence of CMS and IUGR than their shorter-resident Andean counterparts . The nature and rate of human adaptation to future atmospheric hypoxia will depend on stochastic events and making predictions is dogged by uncertainty, but the rate of oxygen decline that is projected by the parabolic model (PatO2 falling by 50% over the next 3500 years) may not provide sufficient time for the development of a Tibetan phenotype, but perhaps just enough to allow an Andean pattern of traits to emerge.
Hypoxia survival limits and human extinction
Even with genetic and phenotypic adaptation, the parabolic decline described by this mathematical model predicts a scenario in which atmospheric oxygen concentration falls to levels below the threshold where human survival and reproduction may be sustained. Defining this point in terms of oxygen concentration is difficult, and our hypothesis is based on the highest elevations known to sustain lifelong human habitation. The highest permanent settlement in the present day is the Peruvian village of La Riconada, at an altitude of 5100 m, which has around 30,000 inhabitants . Native villagers have survived there for at least 40 years and current residents have successfully gone through child birth to create the next generation at this altitude , however, it is not known whether the birth rate can sustain this population indefinitely. The highest permanent settlement on record is the (now abandoned) Chilean mining village of Quilcha (5340 m), which was discovered by the 1935 International High Altitude Expedition to Chile . It has been argued that this represents the upper limit of long-term human habitation, because the residents chose to sleep at this elevation and make a daily ascent to the mine above. PatO2 at the Quilcha settlement is 11.3 kPa (slightly higher than 50% of the current PatO2 at sea level). The parabolic deoxygenation model described here predicts that PatO2 at sea level will reach this threshold in ~3600 years from now. During this time, the human species is likely to undergo further positive selection for physiological phenotypes conveying survival advantage in hypoxic conditions. Studies of high-altitude residents tell us that while such adaptations may enable us to function relatively well in an atmosphere that contains just over half the oxygen we breathe today; many will suffer the long-term consequences. Higher rates of maternal pre-eclampsia and death, increased perinatal mortality, low birth-weights (and the myriad consequences of this in adulthood) and escalating pulmonary disease will curtail life expectancy and population growth. Those individuals with independent comorbidities, particularly chronic respiratory and cardiac disease, may suffer exacerbation of their symptoms, reduced function, and reduced length of life. Highlanders may be forced to descend as life becomes intolerable at hypobaric elevations, therefore reducing the surface of Earth that we can populate. The burden of ill health will begin to overwhelm the capability of healthcare services. The last prevailing human phenotypes may resemble those of current high-altitude populations: with enhanced abilities to extract precious oxygen from the atmosphere or deliver it to the tissues, and perhaps superior cellular mechanisms to improve efficiency of oxygen use and defend against hypoxic stress.
It is important to stress that the parabolic model described here is mathematical rather than geophysical . Other authors have disputed the idea that global deoxygenation on a catastrophic scale is possible . One of the key reasons cited for this is that the determining factor in global oxygen decline is fossil fuel usage and current estimates predict that oil, coal, and gas stocks will last 35, 107, and 37 years, respectively . Thus, is it plausible that the increased fossil fuel usage in recent years has caused a temporary acceleration of the deoxygenation phenomenon, which will resolve once reserves have been exhausted. This scenario would predict a very different decline in atmospheric oxygen from the one we have described, with a fall of only a fraction of a percent in 4400 years . Consensus in this area has not yet been achieved, but the need to understand the limits of long-term human survival under progressively hypoxic conditions cannot be questioned, whether we are considering the persistence of the human race in the mathematical model discussed here, or other contexts in which atmospheric oxygen may become scarce, such as future long-term space expeditions. Other environmental changes may also impact the ability of humans to acclimatize to hypoxia during global deoxygenation and these include the rise in both temperature and concentration of carbon dioxide. It is hard to predict the precise effect of these additional physiological stressors but both are likely to reduce further our chances of long-term survival. In particular, rising carbon dioxide levels could lead to metabolic problems if individuals fail to adapt to this adequately. High blood carbon dioxide levels (hypercarbia) can cause acidosis, hypertension, and tachycardia. Those with underlying lung chronic disease may suffer greatest from this. In a simple, short-term experimental model of Earth’s atmosphere, a novel experiment that used plants to generate oxygen and consume carbon dioxide in a sealed hypoxic chamber noted a high carbon concentration (0.66%) towards the end of the 48-h experiment .
The atmospheric changes will also impact other animals on Earth, and failure to adapt will result in extinction both on land and in the seas and oceans. All aerobic life forms will suffer as oxygen is removed from the atmosphere. In addition, plant metabolism may also be detrimentally affected. Already, the rising carbon dioxide concentration has been predicted to reduce the rate of photorespiration, and a falling oxygen concentration may exacerbate this. While the overall effect of a reduction in photorespiration remains unclear, we do know that complete removal of this pathway could lead to metabolic disaster for the plants that use it .
Progressive asphyxiation of the planet would ultimately lead to the demise of humankind through escalating infant mortality and eventually complete failure to reproduce. Perhaps technological advancement could permit the continuation of life within biospheres in the short term , but beyond this, the outside world would become a barren and inhospitable place. It is not possible to predict with certainty the threshold value at which mass extinction becomes inevitable, but we have no evidence that humans can persist for more than a generation in an atmosphere containing half the amount of oxygen currently available at sea level, a situation that, according to a new model, could be upon us in a few thousand years. Unless the process of global deoxygenation is reversed, either by increasing oxygen production or by reducing its consumption, the human race, as obligate aerobes, will be left behind forever, our domination of this planet a brief footnote in its history.
Compliance with ethical standards
This article does not contain any studies with human participants performed by any of the authors.
Conflict of interest
DM provides remunerated consultancy for Siemens Healthcare and Masimo, and has received payment for lectures from Deltex Medical.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.