It is nine degrees Celsius on the dock this morning, gray water, a wind out of the north that will have whitecaps on the sound by afternoon. Before anyone goes down the ladder I have them stand still and breathe slow, in through the nose, longer on the way out, six rounds, maybe eight. You can watch the shoulders drop. What I will not tell them, because it would be a lie, is that this is remaking their nervous system. It is doing something real, just much smaller than the story they have been sold.
The best evidence on slow breathwork is a 2023 meta-analysis in Scientific Reports, Fincham and colleagues, pooling the randomized trials to date. Boiled down, it found a real but modest benefit, not a transformation. For stress the effect was small-to-medium, real but nowhere near dramatic (minus 0.35 on the standardized scale researchers use to compare studies, pooled from twelve trials and 785 people). Breathwork helped anxiety about the same amount (minus 0.32, from a larger set of twenty trials) and depression a little more (minus 0.40, from eighteen). These are not nothing. They are also not what you would guess from how breathwork gets sold, as a reset button or a biology hack.
Read the fine print and it gets more sober. Most studies sat at what the authors call moderate risk of bias, and three of the twelve stress trials were scored high risk outright. Almost every outcome was self-reported and the follow-up short: people told a questionnaire they felt calmer, which matters but is not a measurement. To their credit, the publication-bias test (Egger's) came back clean, not just the flattering studies surviving. It is a genuine, small effect on a soft outcome.
The physiology underneath is real, and worth naming without letting it overclaim. Breathe slowly, around six times a minute, and you fall into step with the body's own loop for holding blood pressure steady. Match that rhythm and your heart rate starts to swing more with each breath. That widening swing is the basis of heart-rate-variability biofeedback, which Lehrer and Gevirtz trace mainly to a strengthened version of that same blood-pressure loop. The best rate is individual, somewhere between 4.5 and 6.5 breaths a minute. Six is a rule of thumb, not a magic number.
Even where it works, the effect is modest. A 2023 Stanford trial in Cell Reports Medicine had people do five minutes a day of cyclic sighing, a long exhale after a double inhale; it beat mindfulness meditation on mood and slowed resting breathing, but moved neither heart rate nor variability. The good feeling was real; the autonomic fireworks were not there to measure.
The adherence problem is what the influencers skip. A 2026 trial in Stress and Health taught paramedicine students a specific branded routine, the A52 Breath Method, ten minutes twice a day for twelve weeks. It worked: stress climbed in the control group as exams neared and stayed flat among the breathers. But only 24 of the 49 assigned to breathwork finished, against 42 of 49 controls asked to do nothing. That lopsided dropout is the study's biggest weakness: the flat-stress result leans on those who kept it up, and there was no benefit for sleep or wellbeing, which its authors blamed partly on who quit. Cheap, low-risk, effective enough, and people cannot keep it up when life gets heavy, exactly when they need it.
One more thing, because it is my beat and my old mistake. Slow breathing and the fashionable hard-breathing methods are not the same animal. The Wim Hof style of rapid hyperventilation with breath holds does not calm you down. As Kox and colleagues showed in 2014, it floods you with adrenaline, a deliberate surge on the sympathetic, fight-or-flight side. That can be worth doing. But it strips out carbon dioxide, and near water the delayed urge to breathe lets oxygen fall until you gray out without warning. That is shallow-water blackout, a documented cause of drowning even in strong swimmers. Hof's own method gives the same rule, never in or near water. One January I caught a woman by the wrist as her knees buckled on the ladder, her feet already in the water. That is not a wellness tip. It is a safety rule.
So where does that leave the practice I still lead every morning in the cold. Roughly where the data leaves it. Slow breathing is one of the few levers you carry everywhere, it costs nothing, it will not hurt you, and it will take the top off a bad hour. Ask the breath to lower your stress a modest amount and it delivers. Ask it to fix your life and you will quit in six weeks, like half those students did, and decide the breath failed you. It did not. It is a small tool for a bad hour, something you can use on a crowded train or at a desk, no ladder and no cold sea required. Keep it that size, and it keeps working.
