The DSM-5-TR defines a panic attack as an abrupt surge of intense fear or discomfort that reaches a peak within minutes. The symptom list includes a pounding heart, shortness of breath, dizziness, tingling and a sense of unreality. It feels like it arrives from nowhere. One ambulatory study suggests it may not: Meuret et al. (2011) monitored 43 people with panic disorder in 24-hour sessions and found respiratory and autonomic instability as early as 47 minutes before attacks the patients reported as sudden. At onset itself, heart rate and breath volume rose and carbon dioxide fell. Only 13 attacks were captured, so read it as a hint, not a settled rule.
Teaching people to stop over-breathing is a popular intervention for panic; the Stanford trial cited below calls the evidence for the theory behind it tenuous. We found no trial of 4-7-8 during a panic attack. The protocol below is a structure for the minutes an attack takes to peak and pass, built on what the breathing research does support. Where that research contradicts the popular story, this page says so.
The in-the-moment protocol
The instant you recognize it — chest tight, breath fast, catastrophizing thought loop — switch to 4-7-8. Inhale through the nose for four counts. Hold for seven (this will feel impossible; it’s not, do it anyway). Exhale through the mouth for eight, slowly, against the urge to gasp the next breath in immediately.
Do not try to take a normal breath. The first inhale will be shaky and short, and the pull will be to over-breathe. Override it. Four seconds in, hold seven, eight seconds out, repeat.
Do four cycles. If the attack is still intense, do four more, and keep going until the wave passes. There is no trial-derived cycle count here: four, then four more, is a structure to hold on to, not a dose with a measured effect.
Why the long exhale specifically
The best direct evidence that the ratio matters, not just the pace, is a lab study. Van Diest and colleagues had 30 participants breathe at 6 or 12 breaths per minute, each with either a long exhale or a long inhale. The long-exhale pattern produced higher self-reported relaxation and stress reduction. Slowing the rate on its own raised only one score, positive energy. The long exhale also increased high-frequency heart rate variability, but only at the slow rate. That is a study of instructed breathing at rest, not of panic.
“Our results show that i/e ratio is an important modulator for the autonomic and subjective effects of instructed ventilatory patterns.”
4-7-8 is built on that ratio: the exhale is twice the inhale.
The popular explanation for the hold is carbon dioxide: over-breathing blows off CO2, low CO2 produces the tingling and dizziness, and slowing down lets it recover. The evidence is genuinely mixed. Reviewing it, Meuret and Ritz (2010) read it as suggesting an important role for low CO2 in panic disorder and reported beneficial effects from capnometry-guided training to raise it, while calling the evidence for both leading theories of why inconclusive. Kim, Wollburg and Roth then randomised 74 people with panic disorder to handheld-capnometer training that raised their CO2, training that lowered it, or a waitlist. Both active methods reduced panic severity (effect sizes 1.34 and 1.53 at one month), and the gains held at six months.
“Clinical improvement must have depended on elements common to both breathing therapies rather than on the effect of the therapies themselves on CO2 levels.”
The authors suggest the common elements may have been changed beliefs and expectancies, exposure to ominous bodily sensations, and attention to regular, slow breathing. So the CO2 explanation for the hold is, at best, unproven. Do the hold if you can; drop it if you can’t (below).
If 4-7-8 feels impossible mid-attack
The seven-second hold can feel unreachable when you’re already breathing fast and shallow. Two alternatives, in order of usefulness during a spike:
Extended exhale, no hold.Inhale four, exhale eight. Skip the hold entirely. You keep the long exhale; you skip the part that can feel like suffocation when you’re already panicky. Six to eight cycles. Switch to full 4-7-8 once the worst has passed.
Box breathing. Inhale four, hold four, exhale four, hold four. It gives up the long exhale in exchange for an even, four-part count. Use it if the count is what you can hold on to, typically when the catastrophizing thoughts are louder than the physical symptoms.
What not to do mid-attack
- “Just take a deep breath.”In Meuret’s recordings, panic onset came with bigger breaths and falling CO2. A fast, deep inhale adds more of the same. If you take a big breath, make the exhale the long part.
- Wim Hof or any cyclical hyperventilation. It is deliberate over-breathing followed by breath-holds. In Kox et al. (2014), healthy volunteers trained for ten days in the method (its breathing, plus meditation and cold exposure) then practised it in the lab: the result was intermittent respiratory alkalosis, the low-CO2 state, and profoundly increased adrenaline levels. Neither is what you want mid-panic. Save it for stable, controlled contexts.
Practice the protocol when you don’t need it
Rehearse 4-7-8 when you are calm. Once a day for a week. Get the rhythm familiar enough that you don’t have to think about it during a spike.
This is advice, not a finding; no study we know of measures it. The reasoning is simple. A count you have done a hundred times is one you can follow on autopilot. A count you are learning mid-attack is one more thing to hold.
Panic disorder vs occasional panic attacks
Isolated panic attacks are common. In the US National Comorbidity Survey Replication, Kessler et al. (2006) put the lifetime prevalence of panic attacks without panic disorder or agoraphobia at 22.7%, against 3.7% for panic disorder without agoraphobia. Panic disorder — recurrent, unexpected attacks plus persistent worry about the next one — is a clinical condition. Among psychological therapies, CBT is the most extensively studied and was often superior to the alternatives, though with small effect sizes, in a Cochrane network meta-analysis (Pompoli et al., 2016).
Breathwork is not a primary treatment for panic disorder, and a trial built to isolate it questions whether it adds anything to one. Schmidt and colleagues randomised 77 people with panic disorder to CBT with breathing retraining, CBT without it, or a waitlist. Some data suggested the breathing component made outcomes worse; overall the two CBT arms came out equivalent.
If your panic attacks are recurring weekly or more, or if the fear of having one is changing what you do or where you go, the highest-leverage move is a clinician, not a new breathing technique. Use the long exhale as something to do in the moment while you pursue the treatment that has the evidence.