The Breathing Exercise That Beat Meditation
A randomised trial put three breathing patterns against mindfulness meditation for five minutes a day over a month. The breathing arms won, and the technique everybody has heard of was not the one that led. Here is what that does and does not license.
Key takeaways
- A randomised trial published in Cell Reports Medicine compared three five-minute daily breathing patterns against mindfulness meditation over one month.
- The breathing arms beat meditation on mood improvement and on reducing respiratory rate. Within them, cyclic sighing, the exhale-focused pattern, performed best.
- Box breathing, the technique in every app and on every wellbeing poster, improved people and did not come top. That is not a refutation, it is being out-performed once.
- A separate meta-analysis of twelve trials found a small-to-medium benefit for breathwork, judged most studies to be at moderate risk of bias, and warned against a miscalibration between hype and evidence.
- Cyclic hyperventilation with retention is deliberately not described here. Deliberate overbreathing with breath holds can cause fainting and has caused deaths in or near water.
Five minutes a day, for one month. That is the whole dose in the trial, short enough that the usual objection does not apply.
The trial is worth knowing about for a reason most coverage of it missed. It did not test breathing against nothing. It tested three breathing patterns against each other and against mindfulness meditation, which has the larger literature and the better reputation, and the breathing arms came out ahead.
Start with the exhale
Cell Reports Medicine published it in 2023, from a Stanford group led by Melis Yilmaz Balban. Participants were randomly assigned to one of four daily five-minute practices and kept it up for a month, entirely remotely. Three practices were breathing patterns. The fourth was meditation, matched for time.
The patterns differed only in where the emphasis fell.
| Pattern | What it emphasises | How it did here |
|---|---|---|
| Cyclic sighing | Prolonged exhalations | Best of the four on mood and on respiratory rate |
| Box breathing | Equal inhalation, hold, exhalation, hold | Improved, did not come top |
| Cyclic hyperventilation with retention | Longer inhalations, shorter exhalations, holds | Improved, and carries risks set out below |
| Mindfulness meditation | The comparison condition, matched for time | Improved by less on mood than the breathing arms |
Mood and anxiety were the primary outcomes, measured alongside respiratory rate, heart rate and heart rate variability through the month. Breathwork beat meditation on mood improvement and on reducing respiratory rate. Within the breathwork arms, cyclic sighing, the exhale-focused one, performed best.
The paper's own conclusion is narrower than the coverage: daily five-minute cyclic sighing has promise as an effective stress management exercise. Promise, and an exercise.
Check what the trial actually ranked
The comparison the design supports most strongly is breathwork against meditation. The ranking inside the three breathwork arms is the weakest comparison in the study, and it is the one every headline used.
So box breathing, the technique everybody has heard of, equal counts in and hold and out and hold, usually attached to a story about special forces training, did not fail here. It improved people's mood. It did not come top. Those are different findings, and a technique that fails to lead a four-arm trial has not been refuted. It has been out-performed once, which licenses a preference and not a verdict.
The mechanism behind the preference is not exotic. Slow breathing, under ten breaths a minute, has its own psychophysiological literature, reviewed by Andrea Zaccaro and colleagues in Frontiers in Human Neuroscience, where the effects are agreed and the mechanism is still being argued over. Lengthening the exhale is the part of it that turns up in these outcome data.
Treat one trial as one trial
A meta-analysis of breathwork trials appeared in Scientific Reports the same year, from Guy Fincham and colleagues at Sussex. Twelve randomised trials on self-reported stress, 785 adults between them, and a small-to-medium effect in breathwork's favour. Anxiety and depressive symptoms came out similar across larger sets of trials.
Two details in that paper matter more than the effect size.
Most of the included studies were judged to be at moderate risk of bias. And the authors close by urging caution and asking for low risk-of-bias designs, in order to avoid what they call a miscalibration between hype and evidence. That warning was published by people whose own result was positive.
One thing the meta-analysis cannot do is adjudicate the Stanford trial, because its literature search closed in February 2022 and the trial appeared in 2023. The two are separate pieces of evidence, and neither has yet checked the other.
Ask what the study cannot tell you
Participants were healthy adults who volunteered for a remote breathing study, a particular kind of person doing a particular kind of thing. Mood and anxiety were self-reported. A month is a month. The paper does not tell you, and this piece cannot tell you, how many people signed up and quietly stopped, which in any remote month-long trial is the number that decides how much the rest means.
One of the arms, cyclic hyperventilation with retention, is deliberately not described here. Deliberate overbreathing followed by breath holds can cause fainting, and has caused deaths when practised in or near water. It ran under a study protocol with supervision. It is not something to pick up from an article.
Weigh it honestly
The evidence is good enough to act on and not good enough to sell. Worth keeping: a five-minute daily practice with a longer exhale than inhale, because it led a randomised trial, because slow breathing has an independent literature behind it, and because the practice costs almost nothing.
Worth dropping: the assumption that the technique with the best story carries the best evidence. Box breathing has the story and finished behind.
Worth holding loosely: the size of any of this. The honest framing of the alternatives is not that breathing works and the rest does not. Meditation improved people too. If the exhale version turns out to be no better for you, you have spent five minutes on something that also helps.
And if what you are managing is a diagnosed condition instead of an ordinary difficult week, this belongs alongside treatment and not in place of it. The researchers say as much themselves.
Start tonight
Breathe out for longer than you breathe in, five minutes a day, and give it the month the trial gave it before deciding whether it did anything.
Sources
Yilmaz Balban, M., Neri, E., Kogon, M. M., Weed, L., Nouriani, B., Jo, B., Holl, G., Zeitzer, J.
Fincham, G. W., Strauss, C., Montero-Marín, J., & Cavanagh, K. (2023). 'Effect of breathwork on stress and mental health: A meta-analysis of randomised-controlled trials.' Scientific Reports, 13, 432.
Zaccaro, A., Piarulli, A., Laurino, M., Garbella, E., Menicucci, D., Neri, B., & Gemignani, A.
A note on what this piece does not report: the Yilmaz Balban abstract does not state the number of participants or the attrition rate, and this publication has not obtained the Both are material to a remote month-long trial and the article says so.