Back to Home

← All mechanisms

Activating breathwork — rapid breathing and breath holding

Reviewed November 2026

Research strength — Strong for acute activation and for the inflammatory response, Moderate for mood and anxiety with regular practice.

Traditional convergence — Strong (bhastrika, kapalabhati and tummo, in continuous use for centuries).

In short: these practices rouse the system immediately and measurably — that part is physiology, not belief.1 More surprisingly, they give people a degree of voluntary control over their own inflammatory response,2 which is one of the more remarkable findings in this whole evidence base. And three weeks of daily practice has produced substantial reductions in depression, anxiety and stress that were still there three months later.3 The safety cautions on this page are not decorative, and they are the reason it is the longest such section in the library.1


What it is

Breathing deliberately faster and harder than the body needs, often followed by holding the breath. The common element is intentionally shifting the balance of carbon dioxide and oxygen rather than letting the body regulate it.

Rapid breathing drives carbon dioxide down. Breath holding then allows oxygen to fall further than it otherwise would, because the urge to breathe is driven mostly by rising carbon dioxide rather than falling oxygen, so it arrives late. Both halves of this have been measured directly: hyperventilation dropped end-tidal carbon dioxide from around 29 to 17 mmHg, delayed the breaking point of a breath hold, and let blood oxygen fall progressively lower across repeated holds.1

That single fact explains both the effects and the dangers.


What it does for you

It wakes the system up, immediately and reliably

Not controversial, and not subtle. Fifteen seconds of rapid breathing drops end-tidal carbon dioxide measurably — in one study from around 29 to 17 mmHg.1 That shift produces the familiar tingling, light-headedness and heightened alertness directly, through known physiology.

Training in the best-known modern protocol also produces a sharp rise in adrenaline.2 A measurable adrenaline surge is, straightforwardly, an activating effect.

If what you want is to be more awake and more alert within a couple of minutes, this is the best-evidenced thing in the library for that purpose — a comparison you can check against the other mechanism pages, none of which has acute activation evidence of this quality.

It gives you measurable control over your own inflammatory response

This is the finding worth knowing about.

In a randomised trial, healthy young men trained in the method for ten days were then injected with bacterial endotoxin — a standard laboratory challenge that provokes a controlled inflammatory response. Compared with untrained controls, the trained group showed:2

MarkerChange
TNF-α (inflammatory)−53%
IL-6 (inflammatory)−57%
IL-8 (inflammatory)−51%
IL-10 (anti-inflammatory)+194%

They also reported fewer flu-like symptoms and recovered from fever faster.2 The adrenaline surge produced by the breathing appears to be what mediates it.2

Voluntary, trained modulation of the innate immune response was not thought to be possible. Ten days of breathing practice produced it.

Regular practice reduces anxiety and lifts mood

A randomised controlled trial of Bellows Breath (indexed as bhastrika pranayama) trained 30 healthy young adults over four weeks against an active control doing cognitive games.4 State anxiety fell, negative affect fell, and positive affect rose in the breathing group. Brain imaging showed changes in the amygdala and anterior insula when viewing negative images, and connectivity changes that tracked with the anxiety improvement.4

And over three weeks it produced substantial, lasting improvement in depression

The largest trial in this area randomised 84 women with moderate-to-high depressive symptoms and high stress to 21 days of practice — about 15 minutes each morning: four cycles of rapid breathing with breath holds, plus a cold finish to the shower.3

MeasureAfter 21 daysAt 3 months
Depression−20%−30%
Anxiety−21%−39%
Perceived stress−17%−25%

All statistically significant, and larger at three months than at the end of the intervention — the benefit kept developing after the programme stopped.3 At follow-up, 46% of the women who had started with high symptoms had mild or no depression.3 Reactive cortisol also fell.3

Retention was 93%3 — which tells you something useful on its own: people were willing to keep doing it.

It has improved disease activity in inflammatory arthritis

In patients with axial spondyloarthritis, eight weeks of added training reduced disease activity scores (BASDAI, p = 0.012; ASDAS-CRP, p = 0.044) and inflammatory markers.2 Nineteen patients, so a signal rather than a settled result — but a signal in the same direction as the immune finding.


How much you need to do

  • The immune result came from 10 days of training, most intensive over the first four days, then daily home practice.2
  • The depression result came from 21 days at about 15 minutes each morning.3 That is the clearest dose instruction available for this mechanism.
  • The anxiety trial ran four weeks: three supervised 30-minute sessions weekly plus at least two at home, with 25 minutes of continuous practice per session.4 A heavier dose than most people will do.
  • No study establishes an optimal number of rounds, breaths or sessions.

So: around 15 minutes in the morning for three weeks is the dose with the best outcome evidence behind it.


Practices that rest on it

Bellows Breath · Skull-Shining Breath · Cyclic Hyperventilation · Breath Retention · The Great Lock

Where the evidence stops

It has not been shown to beat slow breathing. The 84-woman trial compared it against an active control doing slow breathing at eight a minute with warm showers — and both groups improved equally.3 There was no no-treatment arm, so some of the change in both groups could be time or regression to the mean.3

Read carefully, this is useful rather than discouraging: two different breath practices, both in this library, produced the same substantial result. What it does not support is a claim that the activating form is superior. If someone finds rapid breathing unpleasant, the evidence says slow breathing gets them to the same place.

The wider literature on the modern protocol is weak. A systematic review found nine papers covering eight trials, samples of roughly 13 to 48 people, with seven of eight rated high concern for risk of bias and overall quality described as "very low."2 Around 86% of participants were male, so effects in women were largely untested until the trial above.2 Outcomes were too varied to pool.

That review's own summary overstates its findings. Its discussion says the method "can probably be used" in lifestyle medicine, while the same paper rates the evidence very low quality and says more trials are needed before recommending it to unwell populations.2 The first sentence is the one usually quoted.

Unknowns: whether the Bellows Breath effect comes from the breathing rather than the postures and relaxation practised alongside it;4 whether the immune finding generalises beyond a laboratory challenge in young men; whether it improves exercise performance, where findings were inconsistent and the reviewers call them "too early" to judge;2 anything about older people or people with health conditions.

And one thing to actively disbelieve: that intense sensations indicate benefit. They indicate altered blood chemistry, which is a different thing.


Who should be careful

This is the section that matters most on this page.

Never do these in or near water, and never before swimming or diving. This is not a general caution — it is the specific mechanism by which hyperventilation kills people. Lowered carbon dioxide delays the urge to breathe and lengthens a breath hold, so oxygen falls further before the body objects.1 Across repeated breath holds, blood oxygen dropped progressively lower only in the hyperventilation condition, and four of eighteen participants reached a saturation of 75% or below.1 The authors describe hyperventilation as "a major risk factor for hypoxic blackout during breath-hold diving."1 Someone who blacks out in water drowns without struggling.

The same study found hyperventilation gave no oxygenation benefit at equal breath-hold durations.1 It lets you hold longer without making it safer to.

Practise sitting or lying down, never standing. Light-headedness and fainting are common.

Avoid in pregnancy, and with high blood pressure, heart conditions, epilepsy or glaucoma. These cautions are standard in the traditional literature and reasonable given the cardiovascular changes involved. They are precautionary rather than trial-derived — no study has tested these practices in those groups, which is itself the reason for caution.

Expect side effects. Tingling in hands and face, light-headedness and muscle tightening are the normal consequence of low carbon dioxide, not signs of progress. In one trial, a third of participants reported deafness and heaviness.2

Stop if you feel unwell. These are the only practices in the library where that needs saying.


Traditional convergence

Strong. Forceful rhythmic breathing appears in Indian yogic practice as bhastrika and kapalabhati, and in Tibetan practice as tummo, from which the best-known modern protocol is ultimately derived.5 Both are documented over centuries, and both are treated in their own traditions as advanced practices requiring supervision — a caution that long predates any physiological understanding of why it was warranted.5

These are the only breathing practices that come with that condition attached, in every tradition that has them.


References

APA 7th.

Footnotes

  1. Pernett, F., Bergenhed, P., Holmström, P., Mulder, E., & Schagatay, E. (2023). Effects of hyperventilation on oxygenation, apnea breaking points, diving response, and spleen contraction during serial static apneas. European Journal of Applied Physiology, 123(8), 1809–1824. https://doi.org/10.1007/s00421-023-05202-7 ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8

  2. Almahayni, O., & Hammond, L. (2024). Does the Wim Hof Method have a beneficial impact on physiological and psychological outcomes in healthy and non-healthy participants? A systematic review. PLOS ONE, 19(3), e0286933. https://doi.org/10.1371/journal.pone.0286933 — reporting Kox, M., et al. (2014), Proceedings of the National Academy of Sciences, 111, 7379 (https://doi.org/10.1073/pnas.1322174111), and Buijze, G. A., et al. (2019). ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12

  3. Blades, R., Mendes, W. B., Don, B. P., Mayer, S. E., Dileo, R., O'Bryan, J., Fromer, E., Guan, J. Y., Cheng, S. S., Mason, A. E., Prather, A. A., Epel, E. S., et al. (2024). A randomized controlled clinical trial of a Wim Hof Method intervention in women with high depressive symptoms. Comprehensive Psychoneuroendocrinology, 100272. https://doi.org/10.1016/j.cpnec.2024.100272 ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9

  4. Novaes, M. M., Palhano-Fontes, F., Onias, H., Andrade, K. C., Lobão-Soares, B., Arruda-Sanchez, T., Kozasa, E. H., Santaella, D. F., & de Araujo, D. B. (2020). Effects of yoga respiratory practice (Bhastrika pranayama) on anxiety, affect, and brain functional connectivity and activity: A randomized controlled trial. Frontiers in Psychiatry, 11, 467. https://doi.org/10.3389/fpsyt.2020.00467 ↩ ↩2 ↩3 ↩4

  5. Primary textual sources for the traditional claim: Svātmārāma, Haṭha Yoga Pradīpikā, ch. 2 (for bhastrikā and kapālabhāti, including its instruction that these be learned under a teacher); the Six Yogas of Nāropa corpus for tummo. Precise verse locations are pending verification against critical editions and are logged as outstanding. ↩ ↩2