Panic attack

A panic attack is an abrupt surge of intense fear or discomfort that reaches its peak within minutes and then subsides. It is not a long bad mood arriving quickly. It has a shape: a steep climb, a short summit somewhere inside the first ten minutes, and a fall that happens whether or not anything is done about it.

What is happening physically is a complete emergency response — the one built for a predator, discharged in full — in a body with no emergency in front of it. Every symptom is a correct component of that response. The only thing wrong with a panic attack is its timing.

A first episode, or any episode with chest pain, belongs with a doctor before it belongs with a page like this one. That is not a way of saying nothing is wrong. It is the right order, and skipping it is how a thyroid problem or an arrhythmia gets called anxiety for two years.

The diagnostic list runs to thirteen, and four or more arriving together inside that fast-peaking surge is what makes an episode a panic attack rather than a bad spike of anxiety. Fewer than four is a recognised thing too, called a limited-symptom attack.

• Pounding, racing or noticeable heartbeat
• Sweating
• Trembling or shaking
• Shortness of breath, or air that will not arrive
• A feeling of choking
• Chest pain or discomfort
• Nausea or abdominal distress
• Dizziness, light-headedness, or feeling faint
• Chills or heat sensations
• Numbness or tingling
• Feeling detached from yourself, or that the world has stopped being real
• Fear of losing control or going mad
• Fear of dying

The last two items are the ones worth staring at, because they are on the list. The certainty of death in the middle of an attack is not a judgement someone made about their situation and got wrong. It is a symptom, in the same sense that the sweating is a symptom — a thing a maximally alarmed body reports. Almost everyone who has an attack has that thought, and almost nobody believes it is a symptom while it is happening.

The same goes for the eleventh item. Feeling unreal, or watching yourself from slightly outside, has clinical names — derealisation and depersonalisation — and is an expected feature of extreme arousal rather than a sign of anything coming apart. The sensation is dramatic. Its content is not evidence.

Sorting these sensations from medical ones, episode by episode, is another page's job: it is in was that a panic attack, or something else. This entry is about why the sensations exist.

The curve is the single most useful thing to know, and it is mechanism rather than comfort — which matters, because comfort stops working at four in the morning and mechanism does not.

An attack runs on a sympathetic discharge: adrenaline and noradrenaline released in quantity, heart rate and breathing driven up, blood shifted toward the large muscles. Those chemicals are cleared quickly, and a response built from something that clears cannot sustain its own peak. This is why attacks are short. Not because the person coped well, and not because the danger passed. Because the fuel is metabolised on a timescale of minutes, and what goes up with it comes down with it.

Three things follow from that, and they are worth having in advance.

It will not do the thing it promises. A heart beating fast is not a heart failing, and the surge driving it clears. Over-breathing lowers carbon dioxide; it does not reduce oxygen, so suffocation is not what is being approached. And an alarm state is not psychosis, which looks nothing like this and does not arrive in a ten-minute block and leave.

The aftermath is longer than the attack and is not a second attack. Trembling, exhaustion, a headache, an aching jaw, several hours of feeling scraped out. That is what a spent alarm response feels like, and it resolves on its own.

"Out of nowhere" is not quite right, and the correction is useful. Meuret and colleagues recorded people with panic disorder continuously and found significant instability across autonomic and respiratory measures for as long as forty-seven minutes before attacks the person experienced as sudden. Onset itself was marked by a rise in heart rate and in the volume of each breath, and a fall in carbon dioxide. The attacks were not spontaneous. They were invisible, which is a different claim — and the breathing was already involved before anything had been noticed.

Here is the part that explains the strangest items on the list, and it is the reason the most popular advice in the world is the wrong advice.

Fast breathing blows off carbon dioxide. Low carbon dioxide narrows the arteries supplying the brain, and shifts blood chemistry in a way that makes peripheral nerves fire on their own. That is where the dizziness, the unreality and the tingling in hands and around the mouth come from — not from the fear directly, but from how the body breathes while afraid. Which means the response to the fear is producing more of the sensations that caused it, and the loop closes.

So take a deep breath points the wrong way. A deep breath moves more air, and more air lowers carbon dioxide further. The instruction that matches the physiology is close to the opposite: fewer breaths, smaller rather than deeper, and a longer out-breath than in-breath. Less air, not more.

Breathing into a paper bag is dangerous and should not be used. It left emergency medicine in 1989, when Callaham reported three deaths among patients given bag rebreathing who turned out to be hypoxic or to be having cardiac ischaemia. A bag lowers inspired oxygen, which is survivable if the fast breathing is psychological and is not if it is physical — and the person for whom it is physical is precisely the person nobody can identify in the moment. The advice survives in circulation because it sounds like it addresses the carbon dioxide, and it does. That is not the reason to avoid it.

Why the sensations are read so loudly in the first place is a separate mechanism, and it has its own entry: interoception under alarm is a loud instrument and a badly calibrated one.

These are two very different facts about the world, and conflating them is the main way the first turns into the second.

In the US National Comorbidity Survey Replication, close to a quarter of adults — 22.7 per cent — reported at least one panic attack in their lifetime. Panic disorder was a fraction of that: 3.7 per cent without agoraphobia, 1.1 per cent with it. Across twenty-five countries in the World Mental Health surveys, the lifetime figure for panic disorder came out at 1.7 per cent. A single panic attack is an ordinary human event. A disorder is a much rarer pattern, and the gap between those two numbers is where most of the useful information lives.

The same cross-national work draws the line more precisely. Recurrent panic attacks predicted the later onset of a range of mental disorders, with an odds ratio of about two. Single attacks did not. Repetition is the signal; one episode is a bad afternoon.

What makes the difference is not the attacks. It is what forms around them, and it forms in the weeks afterwards rather than during. A hand goes to a wrist a few times a day. An unremarkable heartbeat on a staircase becomes legible as a warning. A bottle of water goes into a bag and stays there. Then the geography starts moving: a journey skipped, relief that is immediate and genuine, and immediate relief is the most effective teacher there is. The end state of that process has its own name and its own entry — agoraphobia, which does not mean fear of open spaces.

Two notes that belong here rather than in a footnote. Panic disorder's signature is the attack with no cue at all, including the one that wakes someone out of sleep; nocturnal panic is well documented rather than rare. And attacks that arrive reliably in front of other people and never when alone are usually the physical end of social anxiety instead, which matters because the help is different.

The overlap with physical causes is not marginal, which is why clinicians work through them first rather than as a courtesy. The short list that gets missed most often is not about the heart at all.

Caffeine, at doses people reach without noticing. Klevebrant and Frick's meta-analysis found that around 480 milligrams, roughly five cups of coffee, produced a panic attack in 51.1 per cent of people with panic disorder against 1.7 per cent of healthy controls. That is not a small effect, and it is the one variable on this list a person controls directly. Why the arousal is so hard to tell from anxiety from the inside is covered in caffeine and anxiety.

Alcohol withdrawal, on a slower clock — the morning after heavy drinking, or a few days into stopping, produces tremor, sweating, racing heart and dread in a combination that reads exactly like this.

Some medications, including certain inhalers and decongestants, and stimulants of various kinds. This is why a doctor asks about the last few weeks of medication early.

An overactive thyroid, which produces palpitations, tremor, breathlessness and insomnia in a combination that reads as anxiety, and whose first presentation is sometimes psychiatric-looking.

None of these are lesser explanations that somebody is being fobbed off with. They are different explanations, several with simpler answers than panic has. This is why the rule-out comes first, and why "have it checked" is a statement about sequence rather than about how likely you are to be ill. What a questionnaire is for, and what a normal result does and does not settle, is in anxiety screening.

The scope note, said plainly.

One attack needs information, not treatment, and the information is mostly on this page. Attacks that keep arriving, a standing fear of the next one that runs on quiet days, and a month arranged around preventing them is the recognised shape of panic disorder — and that is a clinician's territory rather than a reading problem.

The genuinely good news is narrow and real. Panic is among the most treatable things in mental health. The Cochrane network meta-analysis of psychological therapies for panic found the strongest evidence for behavioural and cognitive behavioural approaches, follow-up work finds the gains hold after treatment ends, and the guidelines put those approaches first. The active ingredient is approaching the sensations rather than avoiding them, which is a clinician's job to structure and not an article's to prescribe. Most people need weeks rather than years.

The most useful thing to bring to a first appointment is not a self-diagnosis. It is a short factual record: when the first one was, roughly how many since, whether any arrived with no trigger or out of sleep, and an honest week of caffeine and alcohol. Onset, frequency and cue are what an assessment is built from, and most people arrive without them.

The panic attacks check reads the surge, the fear of the next one, and what has already been given up as three separate things, because they call for different responses. It is a screen and it says so.

And if living around this has moved into not wanting to be here, that is not a next-appointment matter. That is today — a doctor, a crisis line in your country, or one person you trust. Please do not hold it until the appointment.