Breathwork has a real but modest evidence base for anxiety. Fincham and colleagues pooled the randomised controlled trials: across the 20 with an anxiety outcome, breathwork was associated with lower self-reported anxiety than control conditions, a small-to-medium effect (g = −0.32). Most trials in their primary analysis were at moderate risk of bias, and the authors warn against “a miscalibration between hype and evidence.” Breathing is not a substitute for therapy or medication when those are indicated.
The two protocols below do different jobs, and they don’t have equal evidence behind them. Resonance breathing, in its biofeedback form, has a meta-analysis. 4-7-8 has a handful of trials in narrow clinical groups. Knowing which is which, and which to use when, is most of the practical knowledge worth having.
For an acute anxiety spike: 4-7-8
Inhale through the nose for four seconds. Hold for seven. Exhale through the mouth (lips slightly pursed) for eight. That is one cycle. Do four cycles.
The design idea is the long exhale. In a lab comparison, Van Diest et al. (2014) found that breathing with a long exhale produced higher self-reported relaxation and stress reduction than the same pace with a long inhale. That supports the ratio. It says nothing about 4-7-8 specifically.
The direct evidence for 4-7-8 is thin. The anxiety trials we found are in narrow clinical groups: patients before endoscopy, with tinnitus, after bariatric surgery. In the last, Aktaş and İlgin (2022) randomised 90 patients recovering from bariatric surgery to 4-7-8, deep breathing or routine care; the 4-7-8 group ended with lower state anxiety than both. In healthy young adults, Vierra et al. (2022) found that three sets of six cycles lowered heart rate and systolic blood pressure; anxiety wasn’t measured. We found no trial of it for everyday anxiety. Use 4-7-8 for a spike because the long exhale is the part with support, not because 4-7-8 has been shown to beat anything else.
The first attempts can feel light-headed. Sit down for the first few sessions; lie down if seated still feels intense.
For chronic anxiety: resonance breathing
For ongoing low-grade anxiety — the kind that stays in the background most of the day and surfaces in the evening — the intervention is daily resonance breathing. Three to ten minutes at six breaths per minute (five seconds in, five seconds out). No holds.
This is the breathing that heart rate variability biofeedback trains; the method is also called resonance frequency feedback (Lehrer & Gevirtz, 2014). HRV biofeedback is where the anxiety evidence is strongest. Goessl, Curtiss and Hofmann pooled 24 studies with 484 participants: the pre-post effect size was 0.81, and 0.83 against control conditions. The effect did not depend on the number of sessions or on whether participants had a diagnosed anxiety disorder. The authors still call for more well-controlled studies.
“HRV biofeedback training is associated with a large reduction in self-reported stress and anxiety.”
Two caveats. Those studies tested biofeedback, where a device shows the heart-rate signal as you breathe; paced breathing without the display is a close cousin, not the same thing. And nobody can give you a reliable timeline: in Goessl’s analysis, the effect didn’t vary with the number of sessions. On why it might work, Lehrer and Gevirtz review the candidates and rate a strengthening of the baroreflex as the best supported, with a vagal pathway to the frontal cortex as a newer proposal.
The stacking protocol
This is our suggestion, not a tested protocol; we know of no trial comparing the combination with either one alone. Use resonance breathing daily as the practice, and pull out 4-7-8 in the moment when anxiety surges.
For most adults with everyday anxiety: 5 minutes of resonance breathing in the morning, plus 4 cycles of 4-7-8 whenever the anxiety jumps mid-day.
What doesn’t work for anxiety
Two popular techniques push breathing the wrong way for an anxious state.
- Wim Hof breathing (rounds of fast deep breaths followed by retention). This is voluntary hyperventilation. In Kox et al. (2014), volunteers trained in the method and then practising it in the lab produced intermittent respiratory alkalosis and profoundly increased adrenaline levels: a lot of arousal, deliberately induced. Wim Hof has contexts where people choose that state; an anxiety spike is not one of them.
- Forced deep breathing.“Just take a deep breath,” done quickly and repeatedly, is over-breathing, and over-breathing lowers CO2. Reviewing the evidence, Meuret and Ritz (2010) read it as suggesting an important role for low CO2 in panic disorder. The protocols above emphasize the exhale, not the inhale, for that reason.
What slow breathing can’t fix
Slow breathing is a tool. It is not a substitute for clinical care for panic disorder, generalized anxiety disorder, PTSD, or any other clinically diagnosed condition. It can sit alongside therapy and (when indicated) medication; it does not replace either.
If your anxiety is impairing your ability to work, sleep, or maintain relationships, breathwork alone is the wrong primary intervention. Talk to a clinician. Use breathing as a daily adjunct alongside whatever the primary treatment is.
If you have a history of severe panic, dissociation, or trauma, proceed gently. Start with two-cycle sessions and stop if anything feels destabilizing.