There is a CO2 tolerance story repeated throughout the functional breathing world that sounds physiologically convincing.

Most people are chronically overbreathing.

This causes chronically low CO₂.

Low CO₂ happens because we have poor “CO2 tolerance” or heightened chemosensitivity.

We can measure this with a breath-hold test.

Then, by breathing less, particularly through the nose, we increase CO₂, improve our “CO2 tolerance”, activate the Bohr effect and dramatically improve oxygen delivery and utilisation throughout the body.

It is a very compelling story because almost every individual component contains some real physiology.

It was convincing enough that when I first came across people speaking about breathing in this way, I went all in. I studied their programmes, invested in their work and started applying many of these ideas into my own coaching.

Then two things started to happen.

Firstly, I noticed a difference between what was being spoken about and some of the measurable outcomes I was actually seeing. Clients were improving in certain areas, sometimes substantially, but other physiological changes that should have occurred if the explanation was correct simply weren’t showing up in the way I expected.

Secondly, I started going deeper into the research.

I am grateful that this work opened me up to an area of physiology that is genuinely underappreciated in health and performance, because carbon dioxide, respiratory control, breathing behaviour and the perception of breathing are all fascinating and important. But because there are pieces of genuine physiology within this story, it is also very easy to connect those pieces together in ways that sound scientifically coherent without actually having evidence for the entire chain.

And that is where I think parts of the functional breathing industry have gone wrong.

In 2023 I put together my own Breath Science Practitioner Certification, which subsequently developed into an ongoing professional community, the School of Breath Science, and later another applied 12-week programme, the Breath Resilience Instructors Training. Since then I have run more than 16 cohorts and trained more than 250 students.

A major goal of that work has been to raise the standard of scientific education within the breathing space, because although there are many passionate educators within it, there is still relatively little depth in respiratory physiology and sometimes a large gap between identifying an interesting physiological mechanism and demonstrating that the outcome being marketed actually occurs.

The deeper I went into the literature, the more I realised that several of the most commonly repeated claims were much less certain than I had previously believed.

Are most people actually chronically overbreathing?

The first question seems obvious, but it is rarely asked.

Where is the evidence that 70, 80 or 90% of the population is chronically overbreathing?

Dysfunctional breathing certainly exists. Hyperventilation syndrome exists. Hypocapnia exists. I work with these things and I suspect they are under-recognised in many areas of healthcare.

But that does not mean almost everybody has them.

Reviews of dysfunctional breathing and hyperventilation syndrome have generally estimated prevalence somewhere around 6–10% of the general population, while simultaneously acknowledging that the true prevalence is difficult to establish because we lack universally agreed diagnostic criteria and a gold-standard assessment. Rates can become considerably higher in particular clinical populations, including people with asthma, which is one reason very large percentages can appear when numbers are taken out of the populations in which they were originally measured.

I would not be surprised if dysfunctional breathing as a broader phenomenon turns out to be more common than 6–10%. But there is an enormous difference between saying we probably underestimate a poorly defined clinical phenomenon and telling people that 70, 80 or 90% of humanity is chronically overbreathing.

If someone describes themselves as evidence-based or science-informed, then that matters. Otherwise statistics stop being tools for public education and start becoming tools for marketing.

There is another problem as well.

Dysfunctional breathing does not necessarily mean persistent hypocapnia, and the mechanisms involved are considerably more complex than simply saying that someone has become “too sensitive to CO₂”.

A recent prospective case-control study by Pauwen and colleagues investigated people with idiopathic hyperventilation syndrome in considerable physiological detail. Rather than finding a simple story of universally exaggerated central CO₂ responsiveness, their findings pointed towards differences in peripheral chemoreceptor behaviour, plant gain, CO₂ stores and ventilatory control, while central hypercapnic chemosensitivity was not simply globally increased.

That is much more interesting physiology, but it is also considerably messier than:

Overbreathing → low CO₂ → increased CO₂ sensitivity.

Then we get to breath-hold time

During a breath hold, CO₂ rises. Obviously. Oxygen also falls, respiratory motor drive increases, respiratory sensations become progressively more intense and eventually the urge to breathe becomes difficult to ignore. Your response to that internal state contributes to when you terminate the breath hold. The problem is that somewhere along the way this has been simplified into the idea that a short breath hold demonstrates “poor CO2 tolerance”, which demonstrates heightened chemosensitivity, which demonstrates low resting CO₂, which demonstrates chronic overbreathing.

The breath hold does not directly measure any of those variables.

During a breath hold several things are happening simultaneously. CO₂ rises, oxygen falls, respiratory motor drive increases, afferent respiratory sensations intensify and the urge to breathe becomes progressively stronger. The point at which someone chooses, or feels compelled, to breathe therefore depends on far more than their starting CO₂. Chemosensitivity matters, but so do lung volume, oxygen stores, metabolic rate, respiratory mechanics, previous experience, expectations, anxiety, interoceptive sensitivity, motivation, the protocol being used and willingness to tolerate discomfort. Which makes this common inference problematic:

Short breath hold = poor CO2 tolerance = low resting CO₂ = chronic overbreathing.

A breath hold may tell us something useful about an individual’s response to respiratory discomfort. It does not, by itself, demonstrate chronic hyperventilation or hypocapnia.

If we actually want to establish whether someone is overventilating relative to their metabolic requirements, then at some point we need to measure ventilation and carbon dioxide rather than infer them from a stopwatch.

This assumption has actually been tested

Courtney and Cohen specifically investigated one of the central claims made within the Buteyko model, which was that breath-hold time could be used to predict resting alveolar or end-tidal CO₂.

It couldn’t.

In 83 participants, breath-hold time did not predict resting end-tidal CO₂ in the direction proposed by the model. In fact, the small relationship they observed went in the opposite direction, with longer breath holds associated with slightly lower rather than higher ETCO₂.

That does not mean breath-hold testing is useless. It means we need to stop pretending that it measures something it does not directly measure. There is another problem with treating breath-hold time as a surrogate for someone’s habitual CO₂ level. The score itself is remarkably context dependent.

Bosco and colleagues repeatedly measured end-expiratory breath-hold time and found that after only eight days of repeated testing, morning breath-hold duration had increased to approximately 160% of the original value. The amount of CO₂ accumulated before the breaking point also increased, suggesting that participants had become capable of tolerating a greater respiratory disturbance before terminating the hold.

The same person had become dramatically “better” at the test over eight days. That does not mean their resting respiratory chemistry had somehow improved by 60%. It means practice, familiarity and tolerance of the breaking-point sensations matter.

Starting lung volume matters. Starting O₂ and CO₂ matter. Metabolic rate matters. Time of day matters. Previous breath-hold exposure matters. The intensity of air hunger matters. How threatening that air hunger feels matters.

A breath-hold score can therefore improve considerably without demonstrating that someone’s habitual breathing pattern has been “corrected” or their resting PaCO₂ has meaningfully increased.

There is research connecting breath-holding behaviour under experimentally controlled hypercapnic conditions with ventilatory responsiveness to CO₂, and of course chemosensory drive contributes to the breaking point. But this is very different from taking someone’s ordinary Control Pause, BOLT score or breath-hold time and declaring that you have measured their “CO2 tolerance” or chemosensitivity.

That doesn’t mean breath-hold testing has no value

One study frequently cited in support of breath-hold testing is Kiesel et al. in 2017, and I actually think this study is useful when interpreted correctly.

Using a breath-hold time below 20 seconds alone, sensitivity for identifying the study’s definition of dysfunctional breathing was only 54%. In simple terms, if 100 people actually met the researchers’ criteria for dysfunctional breathing, a breath hold below 20 seconds would identify only around 54 of them and miss approximately 46.

Increasing the threshold improved sensitivity, but the breath-hold test alone still missed a meaningful proportion of those classified as having dysfunctional breathing.

Where the approach became considerably more useful was when context was added. The researchers combined the breath hold with several simple questions around symptoms and breathing behaviour, substantially increasing the sensitivity of the screening process. That is an important distinction. The study gives some support to breath-hold time contributing to a broader screening process for dysfunctional breathing. It does not demonstrate that a breath hold below 25 seconds independently identifies chronic overbreathing, hypocapnia, heightened chemosensitivity or “poor CO2 tolerance”.

In fact, the test became more useful when it stopped being treated as a standalone physiological marker and became part of a multidimensional assessment. That is exactly what dysfunctional breathing appears to be.

Multidimensional.

Calling all of this “CO2 tolerance” risks turning a complicated interaction between respiratory physiology, perception and behaviour into a single marketing metric, and there may also be consequences to repeatedly telling otherwise healthy people that a 15 or 20 second breath hold means something is fundamentally wrong with their breathing, particularly when the sensations associated with breathing disorders can themselves become amplified by attention, interpretation, expectation and threat.

And then there is the Bohr effect

The Bohr effect is absolutely real.

As CO₂ rises and hydrogen ion concentration increases, lowering pH, haemoglobin’s affinity for oxygen decreases and the oxygen-haemoglobin dissociation curve shifts to the right, facilitating oxygen unloading into metabolically active tissues. But even here the functional breathing story often becomes far too CO₂-centric. CO₂ is not the only variable changing haemoglobin’s affinity for oxygen during exercise. Hydrogen ions are critically important. Increasing H⁺ lowers pH and reduces haemoglobin’s affinity for oxygen. Temperature also matters substantially. Working muscle becomes warmer during exercise and this increase in tissue temperature further shifts the oxygen dissociation curve to the right. Other variables, including 2,3-DPG, influence haemoglobin affinity as well.

These factors do not operate in isolation.

During exercise, metabolically active muscle produces more CO₂, H⁺ increases, pH falls and temperature rises. Together, these changes make it easier for haemoglobin to unload oxygen precisely where metabolic demand is increasing.

That is an elegant physiological system.

But it also means that reducing the whole process to “more CO₂ equals a bigger Bohr effect” leaves out quite a lot of what is actually happening. Most importantly:

Oxygen unloading is not the same thing as oxygen consumption.

VO₂ represents the amount of oxygen actually being consumed by the body.

From the Fick principle:

VO₂ = cardiac output × arterial-venous oxygen difference.

So whole-body oxygen consumption depends on how much oxygenated blood reaches the tissues and how much oxygen those tissues actually extract, which itself is linked to metabolic demand. Changing haemoglobin affinity can facilitate unloading. It does not mean that tissues automatically start consuming substantially more oxygen.

We therefore cannot simply say:

More CO₂ → stronger Bohr effect → more oxygen released → greater oxygen consumption → better performance.

There are several missing physiological steps in that chain. And nasal-breathing research gives us a useful way of looking at whether the predicted advantage actually appears.

Acute nasal breathing: CO₂ goes up, but VO₂ does not

A useful recent study here is the 2025 BreathWISE study by Mapelli and colleagues because the participants were not specifically selected as people who had spent months adapting themselves to maximal exercise while breathing exclusively through their nose.

Twelve healthy adults completed maximal cardiopulmonary exercise tests under several breathing conditions. When they were forced to breathe exclusively through the nose, peak end-tidal CO₂ increased markedly from approximately 37.2 to 45.0 mmHg. Ventilation fell from approximately 88.2 to 55.2 L/min. And peak VO₂ fell from approximately 33.4 to 28.0 mL/kg/min, alongside a reduction in peak workload. The authors concluded that exclusive nasal breathing impaired peak exercise capacity because the nasal airway became a ventilatory limitation at high intensity.

That is important.

If the simple explanation were correct, then creating substantially higher CO₂ and supposedly amplifying the Bohr effect should provide some observable advantage in whole-body oxygen utilisation.

Instead:

CO₂ went up.

VO₂ went down.

Performance went down.

That does not mean the Bohr effect stopped operating. Of course it did not. It means that increasing CO₂ does not automatically translate into greater whole-body oxygen consumption or improved performance.

But what happens after someone adapts to nasal breathing?

This is where the story becomes much more interesting, because acute restriction and adaptation are not the same physiological experiment.

Dallam et al studied ten recreational runners who had already spent an extended period, at least six months, training predominantly with nasal-only breathing. They completed maximal graded exercise and high-intensity steady-state running under nasal and oral breathing conditions.

Unlike people who are suddenly forced to breathe exclusively through the nose, these adapted runners were able to reach similar maximal running performance and time to exhaustion while ventilating considerably less. At maximal exercise, ventilation averaged approximately 90.5 L/min with nasal breathing compared with 117.8 L/min orally, while PETCO₂ was higher at approximately 44.7 versus 40.2 mmHg.

That is genuinely interesting physiology. These runners appeared to have adapted to a substantially different ventilatory strategy while maintaining the ability to exercise at very high intensities. But there is an important detail in the data that often disappears when the paper gets reduced to “nasal breathing produced the same VO₂max”.

VO₂max was approximately:

Nasal: 2.55 L/min

Oral: 2.75 L/min

That is roughly 7% lower during nasal breathing. The p-value was 0.09.

Because this was above the conventional statistical threshold of p < 0.05, the difference was reported as not statistically significant. But “not statistically significant” does not mean “the two conditions were identical”. A p-value of 0.09 means that, if there were genuinely no difference between the breathing conditions, an observed difference this large or larger would occur around 9% of the time under the statistical model because of sampling variation. It does not mean there is a 9% chance that the difference is real, and it certainly does not mean there is a 91% chance that nasal and oral breathing produce identical VO₂max.

The study also contained only ten participants, which means there is considerable uncertainty around the estimate. This is where the effect size becomes useful. The effect size for VO₂max was approximately Cohen’s d = 0.60, which is conventionally considered a moderate effect. Effect size asks a different question from the p-value. Rather than focusing mainly on whether the result clears a statistical threshold, it gives us an indication of how large the observed difference was relative to the variability in the sample. A d of 0.60 is not a trivial difference. Practically, what we saw was a moderate-sized difference in maximal oxygen consumption and the direction favoured oral breathing, not nasal breathing. The study was simply too small to determine with much confidence whether that difference would persist in a larger population.

So I think the scientifically accurate interpretation is not:

“Nasal breathing produces exactly the same VO₂max.”

It is:

“No statistically significant difference in VO₂max was demonstrated, but the observed difference was moderate in size and actually pointed towards lower maximal oxygen consumption during nasal breathing.”

That becomes particularly relevant when the study is used to support the Bohr-effect narrative.

PETCO₂ was higher with nasal breathing.

Yet VO₂max did not increase despite oxygen extraction increasing.

It trended approximately 7% lower.

So once again, higher CO₂ did not translate into evidence of greater maximal whole-body oxygen utilisation.

Perhaps the most interesting adaptation is happening centrally

This is where my own thinking around the nasal-breathing data started moving in a different direction.

When Kathryn Raphael, George Dallam and I were developing our 2024 theoretical paper on nasal breathing, cardiac perfusion, fibrosis and arrhythmia, one question was how someone actually becomes capable of sustaining nasal-only breathing during increasingly intense exercise.

A straightforward explanation is that repeated exposure to relatively higher CO₂ reduces chemosensitivity. The respiratory control system becomes less responsive to a given rise in CO₂, the ventilatory response becomes smaller, and the individual can therefore tolerate the higher CO₂ associated with the ventilatory constraint.

That remains one possible mechanism and was discussed in the paper. Although my confidence in true chemosensitivity shifts are lower than when i first came across functional breathing training.

When we were developing that work, I was particularly interested in whether the adaptation might also be central, meaning that what changes is not simply the sensitivity of the chemoreceptors to CO₂, but the way the brain predicts, interprets and responds to the respiratory sensations produced by that physiology. This is important because air hunger is not CO₂ itself. CO₂ contributes strongly to respiratory drive, but the conscious experience of needing to breathe is an emergent perceptual experience created from respiratory motor drive, sensory feedback, expectations, context and previous experience.

Someone repeatedly exercising while restricting themselves to nasal breathing is repeatedly entering a predictable internal state in which ventilation is constrained, PETCO₂ may remain relatively higher, respiratory effort increases and the urge to open the mouth becomes progressively stronger.

Over time they may become capable of continuing to work despite those signals. The common language would be that they have developed greater “CO2 tolerance”. But perhaps that description is too crude.

Perhaps the chemoreceptors remain perfectly capable of detecting the changing chemistry, while the brain becomes better at predicting the respiratory disturbance, matching the sensation to the underlying physiology and maintaining behaviour without immediately responding by dramatically increasing ventilation.

In other words:

the respiratory signal may still be strong, but the behavioural response to the signal has changed.

After working through these ideas, I later came across the work of Olivia Harrison, previously Olivia Faull, and colleagues, and it struck me how closely some of their findings aligned with the central component I had been thinking about.

In 2016, Faull, Cox and Pattinson compared endurance athletes with sedentary controls and found that during a hypercapnic challenge the athletes’ perception of breathlessness tracked changes in ventilation more closely. The athletes were not simply “less sensitive” to breathing. Their perception appeared to be more tightly calibrated to the physiological respiratory signal. They then took this further using 7 Tesla functional MRI. In their 2018 study, endurance athletes and sedentary participants were exposed to inspiratory resistance while the researchers looked at how the brain anticipated and processed respiratory sensations.

Athletes showed different anticipatory activity within key interoceptive and sensorimotor regions, including the insula, thalamus and primary sensorimotor cortices, alongside altered connectivity between interoceptive attention networks and sensorimotor areas. The authors suggested that exercise training may alter anticipatory representations of respiratory sensations.

That is a fascinating finding because it suggests that adaptation to respiratory stress may involve changes in how the brain predicts what is coming and interprets the incoming respiratory signal. Then Harrison, Russell and Pattinson revisited hypercapnic chemosensitivity more directly in 2022. They compared endurance athletes and sedentary controls using both ventilatory and perceptual responses to increasing CO₂. And importantly, they found no group differences in hypercapnic chemosensitivity slopes. The endurance athletes were not simply less responsive to CO₂. The authors instead suggested that athletes may employ additional strategies during exercise that reduce the influence of chemosensitivity on their ventilatory and perceptual responses. That aligns remarkably well with the possibility I had been interested in.

Perhaps what looks behaviourally like greater “CO2 tolerance” does not necessarily require a large reduction in CO₂ chemosensitivity.

Perhaps part of adaptation is central. Prediction may change. Interoceptive accuracy may change. The relationship between respiratory sensation and behavioural response may change. Respiratory motor control may change. The sensation itself may acquire a different meaning through repeated experience. And this has important implications for the way breathwork describes adaptation.

Someone may become substantially better at functioning while CO₂ rises without having substantially altered their resting PaCO₂ or globally switched down their chemoreceptors. The physiology produces a signal. The brain still has to interpret that signal and decide what to do with it. I want to be very clear that Harrison’s work does not demonstrate that nasal-breathing training causes these brain adaptations. Her studies compared trained endurance athletes with sedentary individuals. They were not longitudinal nasal-breathing intervention studies. So this remains a hypothesis.

But the work provides an intriguing independent line of evidence showing that long-term exposure to respiratory stress can be associated with differences in respiratory perception and central processing without a corresponding reduction in hypercapnic chemosensitivity.

For me, that makes the adaptation story much more interesting than simply saying:

“They improved their CO2 tolerance.”

And the adapted physiology may still offer a real advantage

There is another finding from Dallam et al. that I actually find more interesting than trying to force the results into a Bohr-effect explanation.

During the steady-state run at 85% of maximal running velocity, VO₂ was approximately:

Nasal: 2.64 L/min

Oral: 2.76 L/min

The difference sat at around the conventional significance threshold, p = 0.05, with an effect size of approximately d = 0.71. So at the same external running speed, the participants were using somewhat less oxygen when breathing nasally.  That potentially points towards improved exercise economy. And notice how different that is from the common story.

The potential advantage was not:

More CO₂ caused the body to consume more oxygen.

It was closer to:

The same external task was performed while consuming less oxygen.

Those are almost opposite physiological claims. This is why I think the adaptation data deserve more attention.

Long-term nasal-breathing practice appears capable of changing ventilatory behaviour, lowering ventilation and respiratory frequency, maintaining relatively higher PETCO₂ and allowing some people to perform substantial exercise while operating within that respiratory constraint. Whether the mechanisms include altered respiratory motor control, central perceptual adaptation, changes in breathing mechanics, economy, chemosensory adaptation or some combination remains unresolved. But there is interesting physiology there without needing to claim that the Bohr effect suddenly unlocks substantially more oxygen.

I have explored another possible CO₂ mechanism myself

I also want to be transparent here because I am not standing outside this field throwing stones at everyone else.

In 2024 Kathryn Raphael, George Dallam and I published a brief literature review and theoretical analysis asking whether nasal breathing during exercise could potentially influence the development of cardiac fibrosis and arrhythmia associated with long-term endurance exercise. The mechanism we were interested in was not simply that a larger Bohr effect would create more whole-body oxygen consumption. We were interested in blood flow.

Nasal breathing during exercise can reduce ventilation and maintain relatively higher PETCO₂ compared with oral breathing at equivalent workloads in people adapted to it. Because arterial CO₂ is a potent regulator of vascular tone, particularly within the cerebral and coronary circulation, we proposed that maintaining relatively higher CO₂ could potentially help preserve myocardial perfusion during intense exercise.

The theoretical concern was that excessive relative hyperventilation and falling arterial CO₂ could contribute to myocardial vasoconstriction, whereas relatively higher PaCO₂ during nasal breathing might help maintain myocardial blood flow. That could theoretically alter myocardial oxygen supply without increasing whole-body VO₂.

This is a completely different mechanism from:

higher CO₂ → bigger Bohr effect → more oxygen consumption.

And it remains exactly what we described it as.

A hypothesis.

We did not demonstrate that nasal breathing prevents myocardial fibrosis.

We did not measure coronary blood flow directly during nasal versus oral exercise.

We did not demonstrate reduced arrhythmia.

The paper proposed a physiological mechanism that deserves direct investigation. It’s important to state that here, that this is a speculate hypothesis, mechanisms are useful for generating hypotheses. They are not proof that an outcome has occurred.

Oxygen delivery is also about where the blood goes

There is another piece of physiology that helps show why reducing oxygen delivery to the Bohr effect alone is far too simplistic. Look at prolonged breath holding and breath-hold diving. During severe apnoea, oxygen falls and CO₂ rises, but the body does not simply allow progressively greater oxygen unloading everywhere. It actively redistributes blood flow. The human diving response includes peripheral vasoconstriction, bradycardia, changes in cardiac output and preferential preservation of cerebral and myocardial oxygen supply as systemic oxygen availability becomes threatened. This is why breath-hold diving has been proposed as an in vivo model of the human brain survival response. I would not suggest that nasal breathing during ordinary exercise produces the diving response. It doesn’t. The physiological conditions are entirely different. But the diving response illustrates an important principle.

Oxygen delivery is not determined solely by how easily haemoglobin releases oxygen. It is also determined by where the circulation sends the blood.

The body can preserve oxygen delivery to critical organs while simultaneously reducing peripheral perfusion and constraining oxygen use elsewhere. That is an integrated response involving autonomic regulation, regional vascular tone, cardiac output, arterial gases, metabolic demand and haemoglobin chemistry. The Bohr effect is one component of that system.

So what is the Bohr effect actually doing?

The Bohr effect is physiologically important, and I do not think we should downplay it simply because some people in the breathing industry have overstated what deliberately manipulating CO₂ can achieve.

When metabolically active tissue produces more CO₂, hydrogen ion concentration rises, pH falls and temperature increases, haemoglobin’s affinity for oxygen decreases and oxygen is released more readily where it is being demanded. That is elegant physiology. It is part of the way the body matches oxygen unloading with local metabolic demand. You could reasonably describe that as a homeostatic and survival-supporting oxygen-delivery mechanism. But that is very different from demonstrating that deliberately retaining additional CO₂ creates an extra aerobic performance reserve.

At sea level, in otherwise healthy people with normal arterial oxygenation, I cannot find convincing evidence that intentionally increasing CO₂ through nasal breathing or reduced ventilation produces a meaningful ergogenic advantage specifically through enhancement of the Bohr effect.

The acute data do not show it.

The adapted Dallam data do not show it.

In BreathWISE, CO₂ increased while maximal VO₂ and performance decreased.

In the adapted Dallam runners, PETCO₂ was higher while maximal VO₂ numerically moved around 7% lower with a moderate effect size.

None of that means the Bohr effect is irrelevant during exercise. Without appropriate oxygen unloading, exercise performance would clearly suffer.

The distinction is between these two statements:

“The Bohr effect supports normal oxygen unloading during exercise.”

and:

“If I deliberately raise my CO₂, I can amplify the Bohr effect enough to meaningfully increase oxygen utilisation and performance.”

The first statement is well-established physiology. I cannot see convincing evidence for the second.

And when we remember that H⁺ and temperature are also shifting haemoglobin affinity during exercise, alongside changes in blood flow and metabolic demand, it becomes even harder to justify treating CO₂ as a single master variable controlling oxygen availability.

None of this means nasal breathing is useless

Far from it.

The nose warms, humidifies and filters inspired air. Nasal breathing alters airway physiology, changes ventilatory behaviour and can provide a useful respiratory constraint during exercise. Adaptation to nasal breathing appears capable of substantially reducing ventilation while allowing high workloads to be maintained, and the Dallam data raise interesting questions around exercise economy. Can this prolong an athletes career span? Reduce overall allostatic load and lower injury risk? These are all interesting questions that come to mind.

There may also be interesting central adaptation.

Repeated exposure to respiratory discomfort may alter prediction, interoception and the relationship between respiratory sensation and behaviour in ways that we are only beginning to understand. Our myocardial-perfusion hypothesis raises another potential avenue through which maintaining relatively higher CO₂ could be relevant, although that requires direct experimental testing.

CO₂ itself matters enormously. It influences respiratory control, acid-base chemistry, cerebral and coronary vascular tone and haemoglobin’s affinity for oxygen. Breath holds can also be useful. They can expose how someone responds to air hunger, provide an interoceptive challenge and potentially contribute to a broader assessment when interpreted alongside symptoms, breathing pattern and actual physiological measurements.

The issue is not that these tools and mechanisms are worthless. The issue is what we claim they measure and what conclusions we make from them. A short breath hold is not a blood gas. “CO2 tolerance” is not synonymous with resting PaCO₂. Breath-hold time is not a direct measurement of chemosensitivity. Becoming better at tolerating respiratory discomfort does not necessarily mean the chemoreceptors have become less sensitive to CO₂. Dysfunctional breathing does not demonstrate that most humans are chronically hypocapnic. CO₂ is not the only determinant of oxygen unloading. H⁺, temperature, metabolic demand and blood-flow distribution matter too. A rightward shift of the oxygen dissociation curve does not automatically mean greater whole-body oxygen consumption. Higher CO₂ during nasal breathing has not been shown to produce a meaningful performance advantage through enhancement of the Bohr effect at sea level. And invoking a genuine physiological mechanism does not demonstrate that a breathing technique produces the outcome being attached to it. Perhaps most importantly, physiology should not be flattened into simple cause-and-effect stories simply because those stories are easier to teach, market and remember.

Breathing science is far more interesting than that.

Measure the variable you claim to be measuring.

References

Bosco, G., Ionadi, A., Data, P. G., & Mortola, J. P. (2004). Voluntary breath-holding in the morning and in the evening. Clinical Science, 106(4), 347–352. https://doi.org/10.1042/CS20030260

Boulding, R., Stacey, R., Niven, R., & Fowler, S. J. (2016). Dysfunctional breathing: A review of the literature and proposal for classification. European Respiratory Review, 25(141), 287–294. https://doi.org/10.1183/16000617.0088-2015

Courtney, R., & Cohen, M. (2008). Investigating the claims of Konstantin Buteyko, M.D., Ph.D.: The relationship of breath holding time to end tidal CO₂ and other proposed measures of dysfunctional breathing. Journal of Alternative and Complementary Medicine, 14(2), 115–123. https://doi.org/10.1089/acm.2007.7204

Dallam, G. M., McClaran, S. R., Cox, D. G., & Foust, C. P. (2018). Effect of nasal versus oral breathing on VO₂max and physiological economy in recreational runners following an extended period spent using nasally restricted breathing. International Journal of Kinesiology and Sports Science, 6(2), 22–29. https://doi.org/10.7575/aiac.ijkss.v.6n.2p.22

Dujic, Z., Breskovic, T., & Ljubkovic, M. (2011). Breath hold diving: In vivo model of the brain survival response in man? Medical Hypotheses, 76(5), 737–740. https://doi.org/10.1016/j.mehy.2011.02.012

Faull, O. K., Cox, P. J., & Pattinson, K. T. S. (2016). Psychophysical differences in ventilatory awareness and breathlessness between athletes and sedentary individuals. Frontiers in Physiology, 7, 231. https://doi.org/10.3389/fphys.2016.00231

Faull, O. K., Cox, P. J., & Pattinson, K. T. S. (2018). Cortical processing of breathing perceptions in the athletic brain. NeuroImage, 179, 92–101. https://doi.org/10.1016/j.neuroimage.2018.06.021

Harrison, O. K., Russell, B. R., & Pattinson, K. T. S. (2022). Perceptual and ventilatory responses to hypercapnia in athletes and sedentary individuals. Frontiers in Physiology, 13, 820307. https://doi.org/10.3389/fphys.2022.820307

Jensen, F. B. (2004). Red blood cell pH, the Bohr effect, and other oxygenation-linked phenomena in blood O₂ and CO₂ transport. Acta Physiologica Scandinavica, 182(3), 215–227. https://doi.org/10.1111/j.1365-201X.2004.01361.x

Kiesel, K., Rhodes, T., Mueller, J., Waninger, A., & Butler, R. (2017). Development of a screening protocol to identify individuals with dysfunctional breathing. International Journal of Sports Physical Therapy, 12(5), 774–786.

Mapelli, M., Salvioni, E., Mattavelli, I., Grilli, G., Zerboni, G., Nava, A., Capra, N., Galotta, A., Biroli, M., Bellini, G., Dall’Asta, M., Pasini, E., De Paola, A., Torzolini, L., Mani, N., Turri, S., Campodonico, J., & Agostoni, P. (2025). Nasal vs. oral BREATHing WIn Strategies in healthy individuals during cardiorespiratory exercise testing (BreathWISE). PLOS ONE, 20(7), e0326661. https://doi.org/10.1371/journal.pone.0326661

Pauwen, N. Y., Bruyneel, M., Herpeux, A., Sergysels, R., Ninane, V., & Faoro, V. (2026). Peripheral chemoreceptors, plant gain, and CO₂ stores as drivers of resting ventilatory control in idiopathic hyperventilation: A prospective case-control study. Journal of Applied Physiology, 140(1), 262–278. https://doi.org/10.1152/japplphysiol.00744.2025

Puigcerver Pérez, M., & Serrano Rosa, M. Á. (2026). Breath-hold time and anxiety-related vulnerability: A systematic review and meta-analysis. Biological Psychology, 209, 109316. https://doi.org/10.1016/j.biopsycho.2026.109316

Raphael, K., McPhilimey, M., & Dallam, G. (2024). Could nasal breathing during exercise inhibit the development of cardiac fibrosis and arrhythmia associated with endurance training? A brief literature review with theoretical analysis. International Journal of Physical Education, Fitness and Sports, 13(4), 10–20. https://doi.org/10.54392/ijpefs2442