Exercise with oxygen · what the search found, and what it did not
Plate 42
Oxygen fraction record
What a study breathed
29
Sources on the page
7
Goals in people
On this page. Counts, not a grade and not a ranking.
No goal on this page meets the rule for a stronger record. The rule is printed above the goals, and the count it read is printed under each one.
Read this first
Educational use only. This page is not medical advice and not a training plan. Oxygen fractions, flows and session lengths are what a study ran, not a setting for you. Oxygen equipment is a fire hazard. People who are low on oxygen at rest have a prescribed-oxygen question this page does not cover. Aaron Cuha is a coach, not a doctor.
Educational use only. This database summarizes published research and is not medical advice. Where a dose appears it is a record of what a trial or clinic ran, or what people describe doing, with its source attached. Nothing here is a recommendation for personal use.
Exercise with oxygen therapy, sold as EWOT, means exercising while breathing extra oxygen at normal air pressure, usually from a reservoir bag or a concentrator through a mask. A PubMed search for exercise with oxygen, EWOT and oxygen multistep therapy returned 21 records. 16 of them are by Manfred von Ardenne, most on his oxygen multistep therapy, the closest named method, and one is an unrelated software paper. None tests a modern EWOT system.
What has been studied under other names is hyperoxic exercise: breathing 30 to 100 percent oxygen during a test, a training block, or the rest between efforts. A 2018 meta-analysis of 51 studies split that into three questions with different answers: performance during the session, adaptation after weeks of training, and recovery between efforts.
It is not hyperbaric oxygen. At normal pressure, extra oxygen raises blood saturation only a little in a healthy person: in a sprint study, breathing 40 percent oxygen raised it by 2.7 percent. Hyperbaric oxygen dissolves far more oxygen in plasma because of the pressure, and has its own page.
The marketing story is that more oxygen in means more oxygen in tissue and better cell energy. The research story is narrower: during hard effort, extra oxygen eases a limit, and on and off it may act as a signal. This section says what the papers propose and where they are weak.
Easier effort, in the moment
A 2018 review describes better oxygen transport, lactate handling, power output and endurance while breathing hyperoxia, and attributes it to higher oxygen delivery and less central fatigue. It also says the few training studies conflict and potential harms have not been tested.
Oxygen swings as a signal
A proof-of-principle study in young healthy people paired light exercise with 30 or 100 percent oxygen and found an inflammation and oxidative stress response overlapping what low oxygen produces, which its authors call the normobaric oxygen paradox. It measured markers only.
The originator's claim
Manfred von Ardenne reported in 1984 that two 15-minute oxygen multistep sessions raised ergometer fitness two weeks later by a mean of 17 percent in 20 healthy men, in a double-blind study he co-authored. No independent replication appears in the search, most of whose records on the method are his.
Where the mechanism story is weak
In 16 trained cyclists, four weeks of interval training in hyperoxia raised 20 km time-trial power 2.1 percent against 4.9 percent for the same training in room air. The authors judged it not worthwhile for competitive endurance athletes.
After one hour at 100 percent oxygen, oxidation markers in healthy people peaked at 8 hours and inflammation markers rose after that, more than at 30 percent. The blood flow effect of hyperoxia during exercise is in the safety section.
In athletes training in low oxygen, breathing extra oxygen during recovery interfered with the metabolic changes the exercise produced, which the authors say may impair the adaptation the training was for.
Every number here is what a study ran, named beside it. None is a setting for a reader. The variable that matters most is the oxygen fraction actually reaching the lungs, which a mask and a flow rate do not guarantee.
Each row is a record from a paper. It is not a plan for you.
| Goal | Who | What the trial ran | Source |
|---|---|---|---|
| COPD, training | 29 patients not low on oxygen during exercise, average age 67 | 45 minutes of high-intensity cycling three times a week for 7 weeks, on 3 L/min oxygen or compressed air. | PMID 12869359 |
| COPD, training | 29 patients not low on oxygen, crossover | Two 6-week periods of interval cycling and strength work, 31-minute sessions three times a week, with oxygen or air at 10 L/min by nasal cannula. | PMID 27427325 |
| Sprint training, athletes | 18 college athletes | Six sessions over three weeks of four to six 30-second all-out sprints, breathing 60 percent oxygen or room air. | PMID 31359633 |
| Repeated sprints, one session | 13 team and racquet sport athletes, crossover | 6-second sprints with 24 seconds of rest until exhaustion, at 40 percent oxygen or room air. | PMID 35359505 |
| Oxygen multistep | 20 healthy working men | Two 15-minute oxygen multistep sessions, then an ergometer test two weeks later. | PMID 6711017 |
| Hypoxic-hyperoxic cycles | 29 people aged 52 to 76 | 15 sessions of four cycles: 5 minutes at 12 percent oxygen, 3 minutes at 33 percent. | PMID 35330183 |
Stronger records first. Nulls and marketing-ahead goals stay on the same page. Every card is people only; animal work is in its own section below. Nothing here is ranked against the others, and no clinic or device is recommended.
How the tier is decided, so you can check it against the counts printed under each one. Stronger: a clinical guideline or a medical society's indication list recommends it in a named setting, or a pooled analysis of randomised trials favours supplemental oxygen and the source rated the certainty moderate. Thin: the pooled result was null or went against supplemental oxygen for the goal as named, the source rated it very low, or nothing was pooled at all. Mixed: everything else, including a result that moved one measure and not another.
Mixed, local, or early
51 studies pooled; certainty not graded by the source; result favours supplemental oxygen
Who was studied
Mostly healthy and trained people tested on a cycle or a treadmill while breathing more oxygen. The 2018 meta-analysis pooled 51 studies, with 100 percent oxygen the most common fraction. A 2022 crossover of 13 athletes tested repeated sprints at 40 percent.
Time to exhaustion improved with a large effect (g = 0.89), time trials and graded tests with small to moderate effects (0.56 and 0.40). An FiO2 of 0.30 or more was enough. In the sprint study, athletes completed 14.5 percent more sprints and 13.6 percent more total work at 40 percent oxygen.
Setting in the evidence
Laboratory gas at a known fraction, 40 to 100 percent.
What this does not show
That anyone is fitter afterwards in room air. Performance with the mask on is not the same as adaptation, which is the next card.
Mixed, local, or early
studies pooled, count not printed in the abstract; certainty not graded by the source; result mixed
Who was studied
Athletes training for three to four weeks with or without extra oxygen, tested afterwards in room air: 16 trained cyclists, 16 male athletes, and 18 college athletes. The 2018 meta-analysis pooled the training studies without printing their count.
The cyclists gained 2.1 percent in time-trial power after hyperoxic training against 4.9 percent after normoxic training. In both Japanese trials maximal oxygen uptake rose in both groups; the lactate curve improved only with oxygen, while in the sprint trial anaerobic capacity and sprint power improved only with room air. The meta-analysis saw a trend towards benefit but could not conclude.
Setting in the evidence
Laboratory gas during interval training, 60 percent oxygen where stated.
What this does not show
A reliable training advantage. The best-trained group did worse with oxygen.
Mixed, local, or early
studies pooled, count not printed in the abstract; certainty not graded by the source; result mixed
Who was studied
Athletes breathing oxygen in the rest between hard efforts: 12 professional soccer players in a 1989 double-blind crossover and 12 elite swimmers in 2011. The 2024 study used blood from athletes training in low oxygen.
The soccer players saw no effect on lactate or on the second bout and could not tell which gas they had. The swimmers held more power in the third to fifth intervals with oxygen. In the 2024 study, oxygen during recovery interfered with the metabolic changes of the exercise. The meta-analysis saw a small trend for recovery.
Setting in the evidence
100 percent oxygen by mask during short rests.
What this does not show
Faster recovery from a training day, or anything beyond the next few minutes.
Mixed, local, or early
9 studies pooled, 703 people; certainty not graded by the source; result favours supplemental oxygen
Who was studied
People with COPD whose oxygen is adequate at rest: two double-blind trials of 29 patients each, a Cochrane review of four trials and 331 people, and a 2025 meta-analysis of nine controlled studies and 703 patients.
With oxygen during training, the 2003 trial's patients trained harder (62 against 52 watts) and gained more endurance (14.5 against 10.5 minutes). The 2016 crossover found peak work rate rose more than twice as much with oxygen, with no difference in strength. Cochrane found less breathlessness and fatigue in daily life; the meta-analysis found more gain in walking distance.
Setting in the evidence
3 to 10 L/min during supervised training, by nasal cannula in the 10 L/min trial.
What this does not show
Anything about healthy people. These trials enrolled people with COPD under supervision, and the result does not transfer to healthy lungs.
Thin, null, or marketing ahead
nothing pooled; certainty not graded by the source; result mixed
Who was studied
20 healthy men in von Ardenne's 1984 study, 59 patients against 14 controls in his 1989 study, and 14 lung patients in an independent 1980 study. All are German-language papers with English abstracts.
The 1984 study reported a 17 percent rise in ergometer fitness two weeks after two sessions. The 1989 study combined the therapy with brain training and found short-term memory did not improve against controls. The independent 1980 study found no significant rise in blood oxygen after 56 hours of cumulative oxygen. A 1998 critical review found the effectiveness of oxygen multistep therapy unproven for any claimed indication.
Setting in the evidence
Oxygen breathing with light exercise, as the method's papers describe.
What this does not show
That the 1984 result has been repeated by anyone other than its author.
Thin, null, or marketing ahead
nothing pooled; certainty not graded by the source; result mixed
Who was studied
34 geriatric patients aged 64 to 92 in a 2017 randomised trial, 25 geriatric patients averaging 83 in a 2024 placebo-controlled trial, and 145 long COVID inpatients in a non-randomised 2024 pilot. One related group's authors declare ownership of hypoxia equipment and medical technology companies.
In the 2017 trial a dementia screening score rose 16.7 percent against -0.39 and walking distance 24.1 against 10.8 percent. The 2024 trial found no meaningful effect on BDNF, IL-6 or CRP. In the long COVID pilot, walking distance rose 91.7 against 32.6 metres, without randomisation.
Setting in the evidence
Device-delivered cycles of 10 to 12 percent and 30 to 35 percent oxygen.
What this does not show
What constant extra oxygen during exercise does. This is a different protocol that happens to use similar equipment.
Thin, null, or marketing ahead
nothing pooled; certainty not graded by the source; no pooled result
Who was studied
Nobody, for these goals. The closest studies are a proof-of-principle in young healthy people measuring inflammation and oxidative markers, and a one-hour exposure study of 14 people measuring the same.
Both found oxygen changes oxidative and inflammatory markers, rising for hours after a single hour at 100 percent. Neither measured lifespan, ageing, toxin clearance or disease. A 1998 critical review found alternative oxygen therapies unproven for any claimed indication.
Setting in the evidence
Laboratory gas at 30 and 100 percent.
What this does not show
Any health benefit. A shift in a marker after breathing oxygen is a response, not an outcome.
There is almost no animal research on exercising with extra oxygen, and none on EWOT itself. The two studies the search returned that touch it are here, apart from the people.
Rat Alzheimer model, hypoxic-hyperoxic cycles
In 41 rats with brain injury induced to mimic Alzheimer's, 15 daily sessions of alternating low and high oxygen restored spatial learning and memory and lowered amyloid beta and inflammation markers in the hippocampus.
Thoroughbred horses, one hard effort
In 8 horses running 2 minutes hard at 21, 26 or 16 percent oxygen, lactate, glycogen and muscle metabolites did not differ by oxygen, and fewer genes changed with exercise in hyperoxia than in normal air.
The name and the evidence do not match. EWOT is sold as a wellness method; the studies that exist are about athletic performance, lung disease rehabilitation, and a separate alternating-oxygen protocol. A seller citing the athlete meta-analysis is citing what happens with the mask on, not what happens to health.
In the moment and over weeks are different questions. Extra oxygen reliably lets people go further during a session. Whether training that way leaves them fitter is mixed, and in the best-trained athletes it left them less fit than training in room air.
The clearest positive training result is in COPD. That is a medical population training under supervision, and it does not transfer to a healthy person, in whom 40 percent oxygen moved blood saturation by 2.7 percent in the sprint study.
Von Ardenne's own studies are the main support for oxygen multistep therapy, and the one independent outcome study in the search was null. That is the shape of a record that has not yet been tested by anyone without a stake.
The blocks above are the studies. The block below is a judgment about what to do with them. It has to say where it goes past the measurements.
Aaron Cuha’s position
This is opinion, not a finding. Everything else on this page is the evidence.
Breathing extra oxygen during exercise makes the session itself easier and lets people with lung disease train harder. There is no evidence that it improves health or slows ageing in healthy people, and the training evidence in athletes points both ways.
The reasoning
Where this goes beyond the evidence
Saying it does nothing for health in healthy people is a judgment about absence: nobody has run the trial that would show a benefit or rule one out. The oxygen-swing mechanism is plausible and the markers do move. If a sham-controlled trial in healthy adults measured a real outcome and found a benefit, this position would be early. It also goes beyond the evidence in calling the harms small: the oxidative and blood flow effects are measured, but what they mean over months of use has not been studied.
Who this is not for
Anyone who smokes or is near a flame with oxygen equipment, the cause of 92 percent of home oxygen burns in the veterans' data; anyone low on oxygen at rest, whose oxygen is a prescription question this page does not cover; and anyone who wants this page to replace a clinician or a coach.
Aaron Cuha, who built this archive and reads the studies behind the pages. He is a coach, not a doctor, and no medical credential is claimed for him anywhere on this site. More about who writes this
Smokers, and anyone near flame
CDC counted 38 deaths in four states linked to smoking while on home oxygen. In the veterans' data, 92 percent of 123 home oxygen burns came from smoking.
People low on oxygen at rest
That is a medical condition with prescribed oxygen. The COPD trials here enrolled people who were not low on oxygen, and this page does not cover prescription.
Competitive endurance athletes chasing adaptation
The trained cyclists gained less from training with oxygen than without it.
Anyone wanting a personal protocol
Flows, fractions and session lengths here are records of what a study ran. They are not instructions for your body.
CDC counted 38 deaths from fires linked to smoking during home oxygen therapy in four states from 2000 to 2007, 37 of them at home. The veterans' health system recorded 123 flash burns on home oxygen, with 23 deaths.
One hour at 100 percent oxygen raised oxidation markers in healthy people for about 8 hours, followed by inflammation markers, more than one hour at 30 percent.
Hyperoxia cut forearm blood flow during exercise to 86.3 percent of normal, and to 67.1 percent in the most affected people.
Oxygen during recovery interfered with the metabolic response to exercise in low oxygen, and training with oxygen gave trained cyclists smaller gains.
This site teaches a blood panel, protein and training before gadgets. Extra oxygen does not move that order. The training itself is what the studies show works; the oxygen changes how a session feels.
Disclosed both ways: the owner of this archive lists oxygen therapy in his own protocol at /my-protocols. That is a personal choice recorded there, not a finding on this page.
Nothing on this page is for sale. No seller wrote it, and no device or brand is named or ranked.
Four setups sold under the same idea. No brand is named, and none is ranked.
A large bag filled from a concentrator, breathed through a mask during exercise. The setup most sold as EWOT.
Evidence match. None directly. No study in the search tested this setup.
Mismatch. The athlete studies used laboratory gas at a known fraction; a bag and mask do not guarantee one.
A home oxygen concentrator at a set flow, by mask or nasal cannula.
Evidence match. The COPD training trials used 3 L/min, and 10 L/min by nasal cannula.
Mismatch. Those trials were in COPD patients under supervision, not healthy people.
A known oxygen fraction from a tank or blender, used in the performance and training studies.
Evidence match. The setting of the 51-study meta-analysis and the athlete trials.
Mismatch. Rarely what a consumer has, and the performance result is with the mask on.
A device that alternates low and high oxygen in timed cycles at rest or with light exercise.
Evidence match. The setting of the geriatric, long COVID and cognitive impairment studies.
Mismatch. A different protocol from constant extra oxygen, with its own small and mixed record.
Floor habits and the neighbouring pages. None of these is a cart.
Oxygen
Oxygen under pressure. A larger record, and a list of accepted uses.
Floor
The training itself is what the studies show works.
Neighbour
Another wellness device, held to the same rule.
Floor
A marker, or a panel you still need. Not a diagnosis.
Disclosure
Where oxygen appears in a personal plan, labelled as one.
Order
Measure first. Floor habits before gadgets.
Map
Every pillar, this one among them, held to the same rule.
Knowledge stays free. No seller wrote this page. Every source below opens the record that was checked before the sentence was written, with what it found and who funded it.
Review. Oxygen transport, lactate metabolism, power output and endurance improve while breathing hyperoxia, attributed to higher oxygen delivery and reduced central fatigue. The few training studies conflict, especially in well-trained endurance athletes, and potential harms have not been tested. No conflict statement in the PubMed record.
Proof-of-principle study in young healthy people: 5 weeks of low-intensity exercise breathing 30 or 100 percent oxygen or light hypoxia, with inflammation and oxidative stress markers measured. Intermittent hyperoxia produced a normobaric oxygen paradox response overlapping hypoxia. Markers only, no health outcome. No conflicts declared.
Two-centre double-blind study, 20 healthy working men. Ergometer fitness two weeks after two 15-minute oxygen multistep sessions rose by a mean of 17 percent, more in those who started lower. Co-authored by Manfred von Ardenne. German-language paper with an English abstract; no funding or conflict statement in the PubMed record.
Single-blind randomised trial, 16 well-trained cyclists, four weeks of interval training twice a week in hyperoxia or normoxia. 20 km time-trial power rose 2.1 percent after hyperoxic training and 4.9 percent after normoxic training. Authors: not worthwhile for competitive endurance athletes. No conflict statement in the PubMed record.
Randomised, 14 healthy non-smokers, one hour at 30 or 100 percent oxygen. Oxidation markers rose from 30 minutes and peaked at 8 hours, then inflammation markers took over, more with 100 percent than 30 percent. No conflicts declared.
Blood metabolome and iron measures from athletes. Oxygen during recovery periods interfered with the metabolic changes that follow exercise in low oxygen, which the authors say may impair adaptation. No competing interests declared.
Double-blind trial, 29 COPD patients not low on oxygen during exercise, average age 67. Seven weeks of high-intensity cycling three times a week on 3 L/min oxygen or compressed air. Training work rate in the last week 62 against 52 watts; endurance rose 14.5 against 10.5 minutes. PubMed tags it as non-government funded; no conflict statement in the record.
Randomised double-blind crossover, 29 COPD patients not low on oxygen, two 6-week training periods with oxygen or air at 10 L/min by nasal cannula. Peak work rate rose 0.16 against 0.07 W/kg, more than double; strength gains did not differ. No conflict statement in the PubMed record.
18 college athletes randomised to six sprint interval sessions over three weeks breathing 60 percent oxygen or room air. Both improved maximal oxygen uptake; the lactate curve improved only with hyperoxia, while anaerobic capacity and sprint power improved only with room air. No conflicts declared.
Randomised single-blind crossover, 13 team and racquet sport athletes, repeated 6-second sprints at 40 percent oxygen or room air. Oxygen saturation rose 2.7 percent, sprints completed 14.5 percent, total work 13.6 percent and session load 19.4 percent. No conflicts declared.
29 people aged 52 to 76, healthy or with mild cognitive impairment, in four groups, 15 sessions of four cycles of 5 minutes at 12 percent and 3 minutes at 33 percent oxygen. Some inflammatory markers rose and cognitive evoked-potential latency improved. One author owns a hypoxia equipment company and another co-founded a medical technology company, as the paper declares; funded by the National Research Foundation of Ukraine.
51 studies in people. During a single bout, hyperoxia had a large effect on time to exhaustion (g = 0.89) and small to moderate effects on time trials (0.56), graded tests (0.40) and dynamic muscle function (0.28); an FiO2 of 0.30 or more was enough. Training and recovery with hyperoxia trended towards benefit, but small samples and varied protocols prevented a conclusion. No conflict statement in the PubMed record.
Narrative review of 89 studies from 1975 to 2016. Acute hyperoxia may enhance endurance and sprint interval performance; long-term adaptation to training in hyperoxia remains inconclusive. No conflict statement in the PubMed record.
16 male athletes randomised to three weeks of interval training in hyperoxia or normoxia, twice a week. Maximal oxygen uptake rose in both groups; the blood lactate curve and mean power in a 90-second test improved only after hyperoxic training. PubMed tags it as non-government funded; no conflict statement in the record.
Randomised double-blind crossover, 12 professional soccer players, two exhausting bouts with 5 minutes of rest breathing 100 percent oxygen or room air. No effect on plasma lactate or on the second bout; players could not tell which gas they had. PubMed tags it as non-government funded; no conflict statement in the record.
12 elite male swimmers, five high-intensity bench-swimming intervals with 6-minute recoveries breathing 100 percent oxygen or room air. Peak and mean power were higher with oxygen in the third to fifth intervals; lactate, pH and reactive oxygen species did not differ. No conflict statement in the PubMed record.
Four trials, 331 participants with COPD not low on oxygen at rest, oxygen during daily activity and exercise. Two trials found less breathlessness after exercise; quality of life domains for breathlessness and fatigue improved. The authors declare no known conflicts.
Nine controlled studies, 703 COPD patients, long-term oxygen plus exercise rehabilitation against rehabilitation alone. 6-minute walk distance, lung function and quality of life improved more; cardiac benefits unclear. No conflicts declared.
59 patients before and after oxygen multistep therapy combined with brain training, against 14 volunteer controls. Short-term memory capacity rose 19 to 23 percent, but short-term memory itself did not improve against the controls. The therapy was combined with brain training, so the two cannot be separated. By von Ardenne; no funding or conflict statement in the PubMed record.
14 patients aged 55 to 72 with chronic lung disease. After 56 hours of cumulative oxygen multistep application, mean arterial oxygen pressure of 68.1 Torr did not improve significantly. German-language paper; no funding or conflict statement in the PubMed record.
German critical review. Conventional oxygen treats diseases of low blood oxygen. The effectiveness of alternative oxygen therapies, oxygen multistep therapy among them, has not been proved for any of their numerous claimed indications. No funding or conflict statement in the PubMed record.
Randomised trial, 34 geriatric patients aged 64 to 92, multimodal training for 5 to 7 weeks with intermittent hypoxic-hyperoxic breathing or room air. Dementia screening score rose 16.7 against -0.39 percent and 6-minute walk 24.1 against 10.8 percent. No conflict statement in the PubMed record.
Single-blind placebo-controlled trial, 25 geriatric patients averaging 83, 30 minutes of intermittent hypoxic-hyperoxic exposure before 20 minutes of cycling, 17 sessions. No meaningful effect on BDNF, IL-6 or CRP. No conflicts declared.
Non-randomised controlled pilot, 145 long COVID inpatients, rehabilitation with or without intermittent hypoxic (10 to 12 percent) and hyperoxic (30 to 35 percent) breathing three times a week. 6-minute walk rose 91.7 against 32.6 metres. Not randomised, so who received it may explain the gap. The lead author declares consulting and speaker fees from several drug and device companies.
41 rats, Alzheimer-like pathology induced by injecting streptozotocin into the brain. 15 daily sessions of intermittent hypoxia-hyperoxia restored spatial learning and memory and reduced amyloid beta and inflammation markers in the hippocampus. No competing interests declared.
8 Thoroughbred horses, 2 minutes of hard exercise at 21, 26 or 16 percent inspired oxygen. Lactate, glycogen and muscle metabolites did not differ by oxygen; fewer genes changed with exercise in hyperoxia than in normoxia. Funded by the Japan Society for the Promotion of Science and Kanagawa University; no competing interests declared.
CDC report. 38 deaths in four states from 2000 to 2007 linked to smoking during home oxygen therapy, 37 in private homes, median age 67. A government surveillance report, not a funded study.
US Veterans Health Administration root cause reports, 2009 to 2015. 123 flash burn events on home oxygen: 100 injuries and 23 deaths. 92 percent of burns came from smoking, though 89 percent of veterans had said they quit. No conflict statement in the PubMed record.
Randomised, 12 young healthy adults, forearm handgrip exercise in normoxia and hyperoxia. Hyperoxia cut forearm blood flow to 86.3 percent of normoxic values, and to 67.1 percent in the most affected subgroup; intra-arterial vitamin C reversed that in the subgroup. Funded by the US NIH.
Educational use only. This page is not medical advice and not a training plan. Oxygen fractions, flows and session lengths are what a study ran, not a setting for you. Oxygen equipment is a fire hazard. People who are low on oxygen at rest have a prescribed-oxygen question this page does not cover. Aaron Cuha is a coach, not a doctor.
Educational use only. This database summarizes published research and is not medical advice. Where a dose appears it is a record of what a trial or clinic ran, or what people describe doing, with its source attached. Nothing here is a recommendation for personal use.