Proco sells supplements. None of them treats panic disorder. None of them treats any anxiety disorder. That is not a disclaimer parked at the bottom of this article — it is the reason the article exists. Panic disorder has a treatment evidence base assembled from tens of thousands of randomised participants, and nothing in it is a capsule you can buy from us or from anyone else. If what you have is the disorder, the useful step is a GP or a therapist, not a purchase.
A panic attack and panic disorder are not the same diagnosis
DSM-5 defines a panic attack as "an abrupt surge of intense fear or intense discomfort that reaches a peak within minutes," during which "four or more" of thirteen listed symptoms occur.
Panic disorder is a different thing. It requires recurrent and unexpected attacks, plus one further criterion: at least one attack must have been followed by "1 month or more" of either "persistent concern about additional attacks or their consequences" or "a significant maladaptive change in behavior related to the attacks." The attacks must also not be attributable to a substance or a general medical condition, and not be better accounted for by another mental disorder.
This is the same distinction that separates anxiety the feeling from anxiety the disorder. The diagnostic line is not drawn at severity. It is drawn at what happens in the weeks afterwards. The disorder is the anticipation and the rearranged life, not the surge.
The epidemiology shows how wide that gap is. The World Mental Health Surveys pooled 142,949 respondents across 25 countries: lifetime prevalence of panic attacks was 13.2% (s.e. 0.1%), against 1.7% (s.e. 0.0%) for panic disorder. Twelve-month figures were 4.9% and 1.0%. Of people who had ever had a panic attack, 12.8% also met criteria for the disorder — the large majority never did. Among those with attacks but no disorder, 66.5% had recurrent attacks, so recurrence by itself does not make the diagnosis either.
Why the physical symptoms are so convincingly cardiac
First, the symptom list is mostly cardiorespiratory. Cochrane's reviews describe the thirteen characteristic symptoms as "including racing heart, chest pain, sweating, shaking, dizziness, flushing, stomach churning, faintness and breathlessness." Four are enough to meet the threshold — so it is possible to qualify on a cluster that reads, symptom for symptom, like an acute coronary presentation.
Second, the tempo matches. "Reaches a peak within minutes" is also the tempo of a cardiac event. There is no slow onset to reassure you.
Third, there is a feedback loop. The cognitive model CBT for panic is built to target is described in the literature as "catastrophic misinterpretation of bodily sensations": the sensation is read as evidence of catastrophe, which drives more arousal, which produces more sensation. The symptoms are not imagined. The interpretation is what therapy works on.
How often this actually happens in an emergency department is less settled than the commonly repeated figures suggest. Older literature reports panic disorder in anywhere from 14% to 50% of patients attending with noncardiac chest pain — a spread that wide is a sign of heterogeneous methods, not a finding. A prospective Australian study of 338 consecutive adults presenting for investigation of possible acute coronary syndrome, diagnosed using a structured psychiatric interview, found panic disorder in just 5.6% (95% CI 3.4% to 8.6%). It also found that the symptoms routinely recorded in a chest-pain workup could not separate the two groups: only "pain felt heavy" differed at all (73.7% versus 48.9%, p=0.04), and every other symptom was similar. Take the honest reading — the overlap is real and common enough to matter, the headline percentages are unreliable, and the presenting symptoms will not tell you which it is.
First-time chest pain is not something to self-diagnose
The NHS lists "anxiety, depression or panic attack" as one possible cause of chest pain, and on the same page says plainly: "Do not self-diagnose – see a GP if you're worried." It also sets out when the answer is not a GP at all: call 999 for "sudden pain or discomfort in your chest that does not go away," "pain that spreads to your left or right arm, or your neck, jaw, stomach or back," or chest pain with feeling "sweaty, sick, light headed or short of breath."
Thyroid disease overlaps too. The NHS symptom list for an overactive thyroid includes "nervousness, anxiety and irritability," "an irregular and/or unusually fast heart rate," "warm skin and excessive sweating," and "twitching or trembling." It notes that while those symptoms have several possible causes, "a blood test can often help to determine whether they're caused by a thyroid problem." We have written separately on the thyroid test that explains a lot of otherwise unexplained symptoms.
Of established panic attacks, the NHS says "although panic attacks are frightening, they're not dangerous" and an attack "will not cause you any physical harm." Both things hold at once: the attacks are not injurious, and a first presentation of chest pain still needs someone who can order the test.
What the treatment evidence actually shows
Psychological therapy
NICE guideline CG113 (published January 2011, last updated June 2020) states that for panic disorder "CBT should be used," and that "CBT in the optimal range of duration (7 to 14 hours in total) should be offered," typically weekly sessions of one to two hours completed within a maximum of four months.
Be precise about what that is. CG113 attaches no letter grades; it uses directive wording with the year of the underlying evidence. A firmly worded recommendation records what a committee concluded should happen — not that the evidence behind it is certain. Those are two different axes, and the Cochrane work makes the gap visible.
The Cochrane network meta-analysis of psychological therapies for panic disorder (published April 2016; 60 studies, 54 analysed numerically, 3,021 participants) found talking therapies generally better than no treatment, but noted the advantage over waiting list was inflated by small-study effects. Its finding on choosing between therapies was negative: "there was no strong evidence to support one talking therapy over the others." Overall quality of evidence was rated low to very low.
Which components are doing the work
A 2018 component network meta-analysis took 72 studies and 4,064 participants and priced each ingredient of CBT separately, as incremental odds ratios with 95% credible intervals. Read carefully, it is largely a map of uncertainty.
- Interoceptive exposure — deliberately provoking the feared bodily sensations — had the highest point estimate for remission, 1.49, but the credible interval ran 0.94 to 2.36, crossing the line of no effect. Same for response: 1.43 (0.94 to 2.16).
- Face-to-face delivery: 1.27 (0.57 to 2.66) for remission, 1.79 (0.93 to 3.53) for response. Both cross 1.0.
- Cognitive restructuring: 1.11 (0.73 to 1.73) for remission. Crosses 1.0.
- The only components whose intervals excluded 1.0 were the two that came out worse: progressive or applied muscle relaxation, 0.59 (0.40 to 0.90) for remission and 0.64 (0.46 to 0.88) for response; and virtual-reality exposure, 0.41 (0.19 to 0.84) for response.
The paper's summary describes interoceptive exposure and face-to-face delivery as "associated with better treatment efficacy." Those intervals are why that is a ranking rather than a demonstrated difference. The authors also record that "the risk of bias of many included studies was judged as unclear or high in various important domains." The one large interval-excluding result was best-against-worst plausible CBT package — both deliverable as "evidence-based CBT" — at an odds ratio for remission of 7.69 (95% credible interval 1.75 to 33.33).
Medication
CG113 states that "unless otherwise indicated, an SSRI licensed for panic disorder should be offered," and that if an SSRI is unsuitable or there is no improvement after a 12-week course, "imipramine or clomipramine may be considered" — both off-label here. It also recommends against two classes outright: "benzodiazepines are associated with a less good outcome in the long term" and "should not be prescribed"; "sedating antihistamines or antipsychotics should not be prescribed for the treatment of panic disorder."
The Cochrane review of antidepressants versus placebo (41 RCTs, 9,377 participants, 8,252 in placebo-controlled arms) found what it graded as low-quality evidence of benefit on failure to respond: risk ratio 0.72 (95% CI 0.66 to 0.79), 6,500 participants across 30 studies, number needed to treat 7 (95% CI 6 to 9). The nulls matter. On dropouts for any cause the overall risk ratio was 0.88 (0.81 to 0.97), but by class, "TCAs showed a benefit over placebo, while for SSRIs and serotonin-norepinephrine reuptake inhibitor (SNRIs) we observed no difference." On dropouts due to adverse effects, antidepressants as a group were less well tolerated than placebo, with TCAs and SSRIs specifically producing more. The authors flagged pharmaceutical-company funding of some trials and short-term outcomes only, concluding that "the choice of whether antidepressants should be prescribed in clinical practice cannot be made on the basis of this review."
A 2023 Cochrane network meta-analysis of 70 drug trials sharpens this. Most drugs beat placebo on response. Between classes: "we found no difference between classes." Tricyclics ranked most effective on average and SSRIs fifth — a ranking that coexists with no demonstrated difference between them, which is how rankings mislead. Benzodiazepines ranked well on short-term response and were the only class with a lower dropout rate than placebo; NICE still recommends against them on the long-term grounds quoted above, while this analysis covers acute treatment only. Quality was moderate for antidepressants versus placebo and low for benzodiazepines, with trials "all having unclear or high risk of bias across multiple domains."
Why "just breathe" is weaker than its reputation
Breathing retraining is the most widely offered self-management advice for panic, and it has the least support for doing anything.
In the component analysis above, breathing retraining's incremental odds ratio for remission was 0.84, credible interval 0.54 to 1.26. The point estimate sits below 1.0 and the interval crosses it: an absence of demonstrated benefit, not evidence of harm. Its dropout estimate, 0.71 (0.45 to 1.14), also crosses the line of no effect, so the suggestion that it at least improves tolerability is not a statistically supported finding either.
The dedicated trial points the same way. A randomised dismantling study assigned 77 patients with panic disorder to CBT with breathing retraining, CBT without it, or delayed treatment. The authors reported that "some data suggested that the addition of BR yielded a poorer outcome," but that "findings were generally more consistent with treatment equivalence, questioning whether BR produces any incremental benefits." Equivalence — not harm, and not help.
There is also a mechanistic hint. When researchers trained 35 patients to raise their carbon dioxide levels using a capnometry device, what predicted lower panicogenic cognitions was not the carbon dioxide change but the breathlessness felt during the exercise — the association held after controlling for anxiety and for carbon dioxide. That study had no control group and measured a mechanism rather than an outcome, so it cannot establish efficacy. It is a hint, not a finding: what may matter is exposure to the frightening sensation, close to the opposite of what breathing exercises are usually offered to achieve.
That matters clinically, because a technique performed to prevent catastrophe can operate as avoidance — one more thing you must do or something terrible will happen. None of this makes a slow breath a bad idea in the moment. It makes it not the treatment, and presenting it as one leaves people believing they have tried the treatment when they have not.
What to do with this
If you have had recurrent unexpected attacks plus a month or more of dread about the next one, or you have started avoiding places, that is the disorder — and it has treatments with real trial evidence behind them, none of which you buy. The NHS advises: "see a GP if you've been experiencing symptoms of panic disorder." Talking therapies based on CBT can be accessed by self-referral as well as through a GP.
No dietary supplement appears anywhere in NICE's recommended treatments for panic disorder. Where CG113 addresses over-the-counter preparations, it tells clinicians to explain "the potential for interactions with other prescribed and over-the-counter medications and the lack of evidence to support their safe use." We are a supplement company telling you the honest answer here is a referral, and that we have nothing to sell you for this.
This article is general information, not medical advice. If you are worried about chest pain, breathlessness or a change in your heart rhythm, get it assessed. Written 18 October 2026.