Intermittent hypoxia is the deliberate practice of cycling between brief periods of reduced blood oxygen and periods of normal oxygen. In breathwork it is produced by short breath holds, most often after an exhale. The pattern is repeated for a few minutes. Researchers study it as a controlled stressor that may prompt cellular adaptation rather than harm.
That last word is the whole argument, and it is where most popular writing on the subject goes wrong.
The distinction that matters: intermittent versus chronic low oxygen
Low oxygen has a bad reputation, and for good reason. Chronic or sustained hypoxia is genuinely damaging. Obstructive sleep apnoea is the clearest example: the airway collapses repeatedly through the night, oxygen falls, the person partially wakes, and the cycle repeats for hours across years. That pattern is associated with cardiovascular strain, oxidative stress and metabolic disruption. Nobody in the research literature thinks that is good for you.
So how can the same broad stimulus be studied as beneficial?
The answer sits in three variables: dose, duration and recovery. Sleep apnoea delivers hundreds of uncontrolled desaturation events per night, for years, with no recovery window and no consent. Intermittent hypoxia as practised in breathwork involves a handful of short holds inside a session lasting a few minutes, once a day, with full normal breathing in between.
Serebrovska and colleagues put this directly in the title of their 2016 paper in Frontiers in Physiology: “Fitness and intermittent hypoxia training: does the dose matter?” That question is not rhetorical. Their wider work, along with the review by Uzun and colleagues in Medical Gas Research (2023), treats intermittent hypoxia as a dose-dependent intervention where the pattern determines whether the outcome is adaptive or harmful.
The variable is not oxygen. The variable is dosing.
Why a breath hold after the exhale lowers oxygen saturation
A breath hold on a full inhale is mostly a comfort test. Your lungs are loaded with a reserve of oxygen, so saturation holds fairly steady for a surprisingly long time while carbon dioxide climbs and your brainstem starts complaining.
A breath hold after a full exhale behaves differently.
Empty the lungs as far as is comfortable and you remove most of that reserve. Your blood keeps circulating and your tissues keep consuming oxygen, but there is very little left in the lungs to replenish it from. Saturation falls faster and further, on a much shorter hold. This is the mechanism behind the technique Dr Prakash Malshe described as nisshesha rechaka, breath holding at residual volume, in his 2011 paper in AYU (32(4):451-457), titled “Nisshesha rechaka pranayama offers benefits through brief intermittent hypoxia.”
Malshe's argument was counterintuitive and worth stating plainly: some of the effects attributed to pranayama may come not from taking in more oxygen but from briefly having less of it. That reframing is what makes retention the interesting part of the practice rather than a decorative pause. The physiology of oxygen delivery is covered in oxygen levels in the body and the science of breathing.
Altitude works on the same principle from the other direction, by lowering the oxygen available in the air rather than in your lungs. That comparison is explored in high altitude training.
Hormesis, mitochondria and cellular housekeeping
Hormesis is the term for a dose-response curve where a small, brief stressor produces an adaptive improvement while a large or sustained version of the same stressor produces damage. Exercise is the familiar case. A hard training session causes measurable muscle damage, and the body responds by rebuilding to a slightly higher specification. Train into the ground daily with no recovery and you get the damage without the rebuilding.
Two adaptive processes come up repeatedly in the intermittent hypoxia literature.
Mitochondrial biogenesis. Mitochondria convert fuel and oxygen into ATP, the energy currency your cells actually spend. Low-oxygen signalling is one of the recognised triggers for cells to build more of them and to improve how efficiently existing ones work. More capacity, and better capacity per unit of oxygen.
Autophagy. This is cellular quality control: the process by which a cell identifies damaged components, breaks them down and recycles the parts. Nutrient scarcity and metabolic stress are established triggers. The proposal here is that brief oxygen scarcity acts as another such trigger, which is why writers reach for the analogy of fasting. It is a useful analogy and it should be held loosely, because the two stressors are not identical and have not been compared head to head in humans under controlled conditions.
Both mechanisms are well characterised in cell biology generally. What remains unsettled is how much of each a few minutes of daily breath holding produces in a healthy adult. The mechanism is plausible. The effect size in real practitioners is not well quantified.
For a broader account of how these pieces fit together in this method, see the scientific explanation of SOMA Breath.
CO2 tolerance is the limiting factor, not oxygen
Here is the part that trips up almost every beginner.
The urge to breathe is driven mainly by rising carbon dioxide and falling blood pH, not by falling oxygen. Your chemoreceptors are far more sensitive to CO2 than to a modest drop in oxygen. So the discomfort you feel at 30 seconds into a hold is a CO2 signal, and it arrives long before your oxygen saturation has moved anywhere interesting.
This has two practical consequences.
First, CO2 tolerance is trainable. With regular practice the same level of carbon dioxide produces less alarm, which is what allows holds to lengthen over weeks. Buteyko's work built an entire method around this observation.
Second, and more importantly, a longer hold is not automatically a better one. Someone with high CO2 tolerance can hold long enough to drop saturation considerably before feeling much urgency, which is precisely the profile that makes breath holding in water so dangerous. Rising tolerance is a reason for more caution, not less.
Nobody is handing out medals in your living room.
How to practise intermittent hypoxia
The protocol is unglamorous. That is a feature.
Set up properly. Sit or lie down, on land, away from water, not in a car, not standing. Do this before you do anything else.
A round looks like this. A period of rhythmic breathing, then exhale fully and hold with empty lungs for as long as is genuinely comfortable, then release and breathe normally until you feel completely settled before starting the next round. Recovery between rounds is part of the dose, not an interruption of it.
Keep it short. Most protocols in this space, including the SOMA Breath approach, work in the range of a few minutes of accumulated retention per session, practised daily. Sessions of three to five minutes are a reasonable starting frame. Longer is not the goal.
Build gradually over weeks. Your first holds may be 15 to 20 seconds. That is fine and it is normal. Add a few seconds per week rather than chasing a number. Intensity should climb slowly enough that you never end a session feeling wrung out.
Stop the moment anything feels wrong. Dizziness that does not settle, chest discomfort, visual disturbance or a panic response means the session ends. Breathe normally and try again another day at a lower intensity.
Consider guidance for the retention phases. Music-paced sessions remove the counting and the temptation to compete with yourself, which is most of why people overdo this. Guided sessions built on this structure sit in the SOMA Breath app, and there is a free guided session if you would rather feel the pacing than read about it.
What the stem cell research actually says
This is the part of the topic where claims outrun evidence fastest, so it is worth going slowly.
Stem cells are undifferentiated cells that can develop into more specialised cell types. They sit in protected environments in the body, including regions of bone marrow where oxygen tension is markedly lower than in arterial blood. Low oxygen is their native condition rather than a threat to them, which is the observation the whole line of research starts from.
The research question is whether a brief drop in blood oxygen prompts some of those cells to leave the marrow and enter circulation. That process is called mobilisation.
What has been reported. Gharib and colleagues, publishing in Sleep in 2010 (33(11):1439-1446), reported that intermittent hypoxia mobilised stem cells from bone marrow into circulation in mice. Mice, explicitly. Serebrovskaya and colleagues reported on intermittent hypoxia and circulating stem cell responses in Experimental Oncology in 2011 (33(4):210-215). Malshe's 2011 AYU paper proposed stem cell mobilisation as part of the mechanism by which brief breath holding at residual volume might act.
About the 15-fold figure. A figure of up to a 15-fold increase in circulating stem cells is quoted widely in breathwork writing. It belongs to the hypoxia research literature, describes an upper bound observed under specific study conditions, and was not measured in people doing breathwork at home. Treat it as a ceiling reported somewhere in the research, not as a result you should expect. Anyone presenting it as a typical outcome of a breathing session is going beyond what the studies support.
What mobilisation is and is not. More stem cells appearing in the bloodstream is a measurable biological event. It is not, on its own, evidence that any tissue was repaired, that any symptom improved, or that any condition changed course. Those are separate claims requiring separate clinical evidence, and that evidence has not been produced for breath-hold practice.
If you want the shorter overview of this territory, see the power of intermittent hypoxia.
Where the evidence stops
A sceptical reader deserves the boundary drawn explicitly, so here it is.
Much of the stem cell work is animal research or small-sample human research. The Gharib 2010 findings were in mice. Rodent mobilisation results do not transfer automatically to humans, and plenty of promising rodent findings have failed to replicate in people. The human work in this area involves small samples, varied protocols and short follow-up.
Mobilisation is not demonstrated healing. Cells entering circulation is a biomarker. No study has shown that breath-hold practice repairs a damaged organ or alters the course of a diagnosed condition in humans, and this article makes no such claim. SOMA Breath is a breathing practice, not a treatment for anything.
Long-term human outcome data is limited. There is no large, long-duration randomised trial of daily breath-hold-induced intermittent hypoxia with hard clinical endpoints. What exists is mechanistic plausibility, animal work, small human studies and a dose-response literature that is still being characterised. Findings across this field are preliminary.
The dose question is unresolved even among researchers who study it. That is literally the title of the Serebrovska 2016 paper. If the people running the protocols are still asking how much is optimal, the honest answer for a home practitioner is conservative dosing.
Popular claims run well ahead of all of the above. A good deal of online content links stem cell mobilisation to specific disease outcomes. Those links are not established, they are not claimed here, and you should discount any source that makes them.
None of this makes the topic uninteresting. Brief controlled hypoxia is one of the more genuinely intriguing areas in cellular physiology, and it costs nothing to practise carefully. It just is not a therapy, and the research does not currently support treating it as one.
Frequently asked questions
What is intermittent hypoxia?
Repeated short periods of reduced blood oxygen alternating with periods of normal oxygen. In breathwork it is produced by brief breath holds, usually after a full exhale, repeated for a few minutes. In clinical and sports settings it is sometimes produced with hypoxic gas mixtures or altitude simulation instead.
How do you do intermittent hypoxia?
Sit or lie down somewhere safe. Breathe rhythmically for a period, exhale fully, hold with empty lungs for as long as is comfortable, then release and breathe normally until you feel fully settled. Repeat for a small number of rounds. Total practice of a few minutes daily, built up gradually over weeks, is a typical structure.
What are the benefits of intermittent hypoxia?
Research points to adaptations rather than guaranteed outcomes. Proposed mechanisms include mitochondrial biogenesis, improved oxygen efficiency, autophagy and increased CO2 tolerance. Practitioners commonly report improved calm and focus, which is easier to attribute to the breathing pattern and the parasympathetic shift than to hypoxia specifically. The cellular claims are mechanistically plausible and not yet well quantified in humans.
Is intermittent hypoxia safe?
For healthy adults practising short, gentle holds in a safe position, it is generally well tolerated. It is not appropriate for everyone. Pregnancy, epilepsy and seizure disorders, cardiovascular disease, uncontrolled blood pressure, aneurysm, retinal conditions, severe asthma, psychiatric conditions and a history of panic or dissociation all warrant a conversation with a qualified health professional first. Never practise in or near water.
How is intermittent hypoxia different from sleep apnoea?
Dose, duration, control and recovery. Sleep apnoea produces hundreds of uncontrolled desaturation events per night, over years, with fragmented sleep and no recovery window, and it is associated with real cardiovascular and metabolic harm. A breathwork protocol involves a few short voluntary holds inside a session lasting minutes, with full normal breathing in between. Same broad stimulus, opposite dosing, different outcome. This is why the dose-response literature matters.
How long should breath holds be?
Long enough to feel a clear urge to breathe, short enough that you are never fighting it. Beginners often start at 15 to 20 seconds after a full exhale. Adding a few seconds every week or two is a sensible rate. Chasing personal records is where this practice stops being useful and starts being risky.
Where to go next
If the mechanism interests you, the useful next step is a short guided session rather than more reading, because the pacing is what keeps the dose sensible.
- Try a free guided SOMA Breath session to feel how retention is paced.
- Practise consistently with the SOMA Breath app.
- Read the scientific explanation of SOMA Breath for the wider physiology.
- Read the power of intermittent hypoxia for a shorter overview.
- Compare the altitude route in high altitude training.
Important safety and medical information
This article is educational only. It is not medical advice. SOMA Breath is not intended to diagnose, treat, cure, or prevent any disease, and it does not replace medication, therapy, or professional medical care. Nothing here should be read as a promise of any health outcome.
Breath retention and intense fast breathing carry real cautions. Speak with a qualified health professional before practising these if you are pregnant, or live with epilepsy or a seizure disorder, cardiovascular disease, uncontrolled high or low blood pressure, an aneurysm, a retinal condition, severe asthma, a psychiatric condition, or a history of panic attacks or dissociation. The same applies if you have obstructive sleep apnoea or any other condition affecting your oxygen levels.
Never practise breath retention or fast breathing in or near water, in a bath, while driving, or while standing without support. Breath holding in water can cause loss of consciousness and drowning, including in strong swimmers and including in shallow water.
Stop immediately if you experience dizziness that does not settle, chest discomfort, visual disturbance, numbness or a panic response. Breathe normally, rest, and reduce the intensity next time. If symptoms persist, seek medical attention.
