Never downshifting out of stress

The stress response is designed to be a spike, not a setting. The thing most adults actually do is stay somewhere in the middle of it all day and call that coping. What the trial evidence supports is that deliberately doing something to come down is associated with feeling less stressed than not doing it: pooled across 12 randomised controlled trials in 785 adults, breathwork produced a small-to-medium effect on self-reported stress compared with non-breathwork controls, Hedges' g of -0.35 (95% CI -0.55 to -0.14). That held whether the comparison group was doing nothing (g -0.37) or doing something else active (g -0.26), which matters, because an active control is the harder test.

Fincham GW et al. 2023

It is a standardised effect size, not a unit of feeling. It says the average person in the breathwork groups scored about a third of a standard deviation lower on a stress questionnaire than the average person in the control groups. By convention 0.2 is small, 0.5 is medium and 0.8 is large, so this is small-to-medium - real, replicable across the set, and modest. It survived removing any single study (the effect moved between -0.27 and -0.39 and stayed significant), a test for publication bias came back clear, and it would take 69 further studies finding nothing to wash it out. The authors' own framing is worth repeating: breathwork may be effective, and they explicitly warn against a mismatch between the hype and the evidence.

Fincham GW et al. 2023

The pooled effect sat with slow breathing. In the 10 trials whose technique was slow-paced, the association held at g -0.35 (95% CI -0.61 to -0.10); the two fast-paced trials pointed the same way but did not reach significance, and the review is clear that the subgroups did not differ significantly from each other. The definition of slow comes from a separate systematic review: ten breaths a minute or fewer. Roughly six a minute is the pace that review's own authors and the anchor both point to, and it needs no equipment to hit - four seconds in, six seconds out is one arithmetically simple way there. What the evidence does not give you is a duration. The anchor tested three dose measures against the size of the effect - total practice time, number of sessions, and how often people practised - and none of them was statistically significant. Five minutes is a container we picked because it is short enough to survive a real day, not a threshold anyone measured.

Fincham GW et al. 2023 · Zaccaro A et al. 2018

Slowing the breath below ten a minute is consistently accompanied by higher heart-rate variability and larger respiratory sinus arrhythmia - the normal speeding and slowing of the heart across each breath - together with more EEG alpha power and less theta, and with people reporting more comfort, relaxation and alertness and less arousal, anxiety, anger and confusion. The review that documents this deliberately excluded anything mixing breathing with meditation, visualisation or yoga postures, so the changes are attributable to the breathing itself. It also, deliberately, pooled nothing: it is a map of directions, not a source of numbers, and it studied healthy adults in mostly single sessions. Bear in mind that heart-rate variability and brain-wave power are measurements of how the system is behaving in the moment, not health outcomes.

Zaccaro A et al. 2018

A separate review of 15 randomised trials in 1,097 participants reported that breathing exercises were associated with 7.06 mm Hg lower systolic and 3.43 mm Hg lower diastolic blood pressure than control groups, plus about 2.4 fewer beats a minute. That sounds dramatic, and it is the least reliable figure here. The trials disagreed with each other to an extreme degree - a standard measure of that disagreement reached 90% for systolic pressure - the review's own summary notes the studies are not free of bias, most participants already had diagnosed high blood pressure, and one trial used a commercial breathing machine rather than an unaided technique. Blood pressure is also a surrogate: it is a risk marker measured in a clinic, and nothing here followed anyone long enough to know whether an event was avoided.

Garg P et al. 2024

The autonomic changes - heart-rate variability, respiratory sinus arrhythmia, the shift in brain-wave power - were mostly observed within single sessions in healthy adults, so the in-the-moment shift is immediate. Whether repeating it changes your baseline stress over weeks is much less clear. The trials in the anchor ran for varying lengths, too few of them measured anything after the intervention ended for the review to pool a follow-up, and the analysis found no relationship between how much people practised and how much their stress score moved.

No. The anchor's secondary results on anxiety and depressive symptoms are questionnaire scores in largely non-clinical populations, and the one included trial in a diagnosed mental-health group was a single study with a non-significant result. Slow breathing is a lever you can pull between appointments, not an alternative to care. If stress, anxiety or low mood is affecting your daily functioning, that is a conversation with a clinician.

The review found no adverse events or lasting bad effects attributed to breathwork in the studies that reported on it. The one exception is worth knowing. In a trial of fast-paced breathwork in people with depression, a small subgroup reported hot flushes, shortness of breath and sweating, though none of them dropped out. That belongs to the fast, intense styles, not to slowing down. If breathing slowly makes you feel light-headed, you are probably breathing too deeply rather than too slowly - make the breaths gentler, not bigger.