This synthesis draws on DSM-5-TR diagnostic criteria,¹ the APA Practice Guideline for the Treatment of Patients with Panic Disorder (2nd edition with subsequent guideline watch),² NICE CG113 on generalised anxiety and panic disorder in adults,³ the Canadian clinical practice guidelines for the management of anxiety, posttraumatic stress and obsessive-compulsive disorders,⁴ the World Federation of Societies of Biological Psychiatry guidelines on anxiety disorders,⁵ Cochrane systematic reviews of pharmacotherapy and psychological therapies for panic disorder,^6-7 and the relevant chapters of Stahl's Essential Psychopharmacology and Kaplan and Sadock's Synopsis of Psychiatry. Search terms included "panic disorder", "agoraphobia", "SSRI", "SNRI", "benzodiazepine", "cognitive behavioral therapy", "interoceptive exposure", and "panic attack". Inclusion was weighted toward guidelines and high-quality systematic reviews; primary RCTs are cited where they anchor a specific clinical recommendation.

Twelve-month prevalence of panic disorder in U.S. adults is approximately 2-3%, with lifetime prevalence near 4-5%.⁸ Onset peaks in late adolescence and early adulthood, with a median age of onset in the early-to-mid 20s; first onset after age 45 is uncommon and should prompt a careful medical workup.^1,8 Women are affected roughly twice as often as men across most epidemiologic samples.⁸ Comorbidity is the rule rather than the exception. Lifetime co-occurrence with major depressive disorder approaches 50-60%; with another anxiety disorder, higher still. Agoraphobia accompanies panic disorder in roughly a third to half of cases depending on the sample. Substance use disorders, particularly alcohol use disorder, are common and often reflect self-medication of anticipatory anxiety. Suicidal ideation and attempts are elevated relative to the general population, and the elevation persists after controlling for comorbid depression.^8-9 Risk factors include female sex, family history of anxiety or mood disorders, childhood adversity (especially separation experiences and parental loss), respiratory illness including asthma, smoking, and recent stressful life events. Heritability estimates from twin studies cluster around 40-50%, consistent with a polygenic, moderately heritable phenotype.¹⁰

No single neurobiological model accounts for panic disorder, but a convergent picture has emerged across pharmacologic, imaging, and genetic work. The dominant framework is Gorman's revised neuroanatomical model, which locates the panic attack in a hyperexcitable "fear network" centered on the amygdala with reciprocal projections to the periaqueductal gray, locus coeruleus, hypothalamus, and brainstem autonomic nuclei, modulated top-down by medial prefrontal and hippocampal structures.¹¹ The model accommodates the rapid autonomic cascade of an attack (locus coeruleus and brainstem), the conditioned context fear that drives agoraphobia (hippocampus), and the impaired extinction learning that perpetuates the disorder (ventromedial prefrontal cortex). Neurochemically, three systems are most strongly implicated. Serotonergic tone is reduced or dysregulated, consistent with the therapeutic efficacy of SSRIs and SNRIs. GABAergic inhibition is decreased, with reduced benzodiazepine receptor binding reported on flumazenil PET in symptomatic patients, fitting both the rapid efficacy of benzodiazepines and the panicogenic effect of flumazenil challenge.¹² Noradrenergic hyperactivity, particularly in the locus coeruleus, is implicated in the autonomic symptom cluster and in yohimbine-induced panic. The interoceptive and respiratory phenotypes deserve specific mention. Patients with panic disorder show heightened sensitivity to CO2 inhalation (the Klein "suffocation false alarm" hypothesis), to sodium lactate infusion, and to cholecystokinin tetrapeptide — provocations that do not reliably trigger panic in healthy controls or in most other anxiety disorders.¹³ These findings underpin interoceptive exposure as a CBT component and link panic disorder to a candidate respiratory subtype. Structural and functional imaging shows amygdala hyperreactivity to threat-related stimuli, reduced anterior cingulate and ventromedial prefrontal volumes in some studies, and altered Default mode network connectivity, though findings are not specific to panic disorder.¹⁴ Genome-wide association studies have not yielded robust replicated loci to date; heritability estimates of around 40-50% from twin studies indicate substantial polygenic contribution.¹⁰ Cognitive models, principally Clark's catastrophic misinterpretation model, integrate cleanly with the neurobiology: ambiguous bodily sensations (a skipped beat, a wave of dizziness) are misinterpreted as imminent catastrophe ("I am dying," "I am losing my mind"), which amplifies the autonomic response, which confirms the catastrophic interpretation.¹⁵ This positive feedback loop is the explicit target of CBT.

DSM-5-TR defines panic disorder by recurrent unexpected panic attacks. A panic attack is an abrupt surge of intense fear or discomfort that peaks within minutes and includes at least four of thirteen specified physical and cognitive symptoms: palpitations, sweating, trembling, shortness of breath, choking sensation, chest pain, nausea, dizziness or lightheadedness, chills or heat sensations, paresthesias, derealization or depersonalization, fear of losing control or going crazy, and fear of dying.¹ At least one attack must have been followed by a month or more of either persistent worry about additional attacks or their consequences, or a maladaptive change in behavior related to the attacks (such as avoiding exercise or unfamiliar places). The disturbance is not attributable to a substance or medical condition and is not better explained by another mental disorder.¹ Two diagnostic distinctions trip up trainees most often. First, panic attacks per se are a specifier that can be applied across DSM-5-TR diagnoses; the diagnosis of panic disorder requires the recurrent unexpected attacks plus the month of anticipatory worry or behavioral change. Second, agoraphobia is a separate DSM-5-TR diagnosis. In DSM-IV agoraphobia could only be diagnosed alongside panic disorder; DSM-5 unbundled them, and a patient may now carry both diagnoses, only panic disorder, or only agoraphobia.¹ Severity is typically captured by panic attack frequency, the magnitude of anticipatory anxiety, and the degree of avoidance. There are no DSM-5-TR severity specifiers within panic disorder itself, though the Panic Disorder Severity Scale (PDSS) is widely used clinically and in research.

The phenomenology of a panic attack is stereotyped enough that an experienced clinician can recognize it from the chief complaint. Onset is abrupt, often from a relaxed state or even from sleep (nocturnal panic affects roughly a quarter of patients). Symptoms peak within ten minutes — typically within two to five — and resolve over twenty to thirty minutes, leaving the patient drained, frightened, and convinced something terrible has happened or is about to. Cardiac symptoms (palpitations, chest tightness, dyspnea) and neurologic symptoms (dizziness, paresthesias, derealization) dominate the presentation; the patient's interpretation of these — heart attack, stroke, suffocation, going insane — drives the secondary fear. The course typically follows three phases. Initial unexpected attacks come without warning and are often attributed to a medical cause; this is the phase of repeated emergency department visits and negative cardiac workups. Anticipatory anxiety develops as the patient becomes hypervigilant for bodily sensations and begins to fear the next attack. Phobic avoidance follows, with situations that previously preceded an attack — driving, supermarkets, public transport, being alone — being progressively avoided. When avoidance broadens to two or more agoraphobic clusters and produces clinically significant impairment, comorbid agoraphobia is diagnosed.¹ Atypical presentations include patients whose attacks are predominantly cognitive (derealization, depersonalization, fear of losing control) with little autonomic flavor; patients who present with somatic preoccupation and repeated medical workups rather than overt anxiety; and the so-called non-fearful panic disorder seen in cardiology clinics, where patients deny anxiety but display all the autonomic features and respond to standard treatment. Red flags that should prompt re-examination of the diagnosis include first onset after age 45, attacks lasting longer than 30-60 minutes, attacks accompanied by focal neurologic findings or true syncope, attacks during sustained physical exertion, and any feature of a postictal state.

The differential splits cleanly into psychiatric and medical mimics, and both must be addressed before settling on the diagnosis. Among psychiatric conditions, the central question is whether the attacks are unexpected. In specific phobia, social anxiety disorder, and PTSD, attacks occur predictably on exposure to the feared cue. In generalized anxiety disorder, persistent worry dominates and discrete panic attacks are less prominent. In obsessive-compulsive disorder, anxiety surges follow obsessions or contamination triggers. Major depressive disorder commonly co-occurs with panic disorder and may present with panic-like episodes; the temporal relationship and the persistence of mood symptoms between attacks distinguish them. Substance-induced panic must be considered in every case. Stimulant intoxication (cocaine, amphetamines, methamphetamine), cannabis (especially high-THC products in young adults), caffeine in heavy users, and the synthetic cathinones can all produce panic that is indistinguishable phenomenologically from a primary attack. Withdrawal from alcohol, benzodiazepines, and opioids is a classic panicogenic state. A complete substance history with attention to over-the-counter sympathomimetics and energy drinks is non-negotiable. Medical mimics divide into the cardiopulmonary, endocrine, neurologic, and toxicologic. Supraventricular tachycardia and other paroxysmal arrhythmias produce abrupt palpitations with anxiety; the absence of typical cognitive panic symptoms and the persistence of tachycardia after the patient calms point toward a primary arrhythmia. Pheochromocytoma is rare but classically presents with paroxysmal hypertension, headache, and diaphoresis. Hyperthyroidism produces sustained autonomic symptoms with weight loss, heat intolerance, and tremor. Hypoglycemia and carcinoid syndrome are uncommon mimics worth keeping on the list. Pulmonary embolism and asthma exacerbation produce dyspnea-dominant attacks. Temporal lobe epilepsy can produce ictal fear that resembles panic but is typically briefer (seconds to a minute), often accompanied by automatisms or postictal confusion. Vestibular disorders produce dizziness-dominant attacks that are reproduced by positional maneuvers.

The interview begins with a detailed account of a representative attack: setting, onset, peak time, symptom inventory, cognitive content, duration, resolution, and what the patient did during and after. The key diagnostic determinations follow from this account: was the attack unexpected, did it meet the four-symptom threshold, and is there a month of anticipatory worry or avoidance. A timeline of attack frequency and triggers, the presence of agoraphobic avoidance, prior treatment trials, family history of anxiety and mood disorders, substance use including caffeine and over-the-counter stimulants, and a medication review are mandatory. Validated rating scales support diagnosis and treatment monitoring. The Panic Disorder Severity Scale (PDSS) is the most widely used clinician-administered measure, with seven items rated 0-4; a self-report version (PDSS-SR) performs comparably.¹⁷ The Panic and Agoraphobia Scale (PAS) and the Anxiety Sensitivity Index (ASI) — the latter measuring fear of bodily sensations — round out the standard battery. The Hamilton Anxiety Rating Scale (HAM-A) is broader and useful when generalized anxiety symptoms coexist. Depression screening with the PHQ-9 at baseline and follow-up is essential given comorbidity rates and shared treatment implications. Physical examination should document vital signs (resting tachycardia or hypertension warrants attention), thyroid examination, cardiopulmonary exam, and a focused neurologic exam. Laboratory workup is targeted, not reflexive: TSH in all patients, basic metabolic panel and CBC if not recent, and a urine toxicology screen when substance use is plausible. ECG is reasonable in patients with palpitations or chest pain prior to attributing them to panic, and is mandatory before starting medications with QT effects (such as citalopram at higher doses). Echocardiography, Holter monitoring, neuroimaging, and EEG are not routinely indicated and should be reserved for atypical presentations or specific findings on history and examination.

First-line treatment of panic disorder is either an SSRI or an SNRI, cognitive behavioral therapy, or both in combination, with the choice driven by patient preference, comorbidity, and access to a trained therapist.^2-5 Strong evidence supports both modalities as superior to placebo and waitlist; meta-analyses generally show comparable efficacy at endpoint, with some advantage for combined treatment over monotherapy in the acute phase.^6-7 Treatment goals extend beyond reduction of panic frequency to elimination of attacks, resolution of anticipatory anxiety, reversal of phobic avoidance, and full functional recovery.

Pharmacotherapy

SSRIs are the recommended first-line agents. Sertraline, paroxetine, fluoxetine, and escitalopram have the most extensive evidence base; class effects appear larger than between-agent differences, so selection is typically driven by side-effect profile, drug interactions, and prior response.^2-5 Initiate at a low dose to minimize early activation — for example, sertraline 25 mg, escitalopram 5 mg, or paroxetine 10 mg — and titrate upward over one to two weeks toward standard antidepressant target doses (sertraline 50-200 mg, escitalopram 10-20 mg, paroxetine 20-50 mg).¹⁸ Onset of anxiolytic effect lags the antidepressant effect somewhat; meaningful improvement typically requires four to six weeks at therapeutic dose, and a full trial is eight to twelve weeks. SNRIs, principally venlafaxine extended-release, are an established alternative with efficacy comparable to SSRIs.^2-5 Venlafaxine XR is FDA-approved for panic disorder; typical target is 75-225 mg daily, started at 37.5 mg. Duloxetine is a reasonable off-label choice when comorbid pain or depression supports it. Tricyclic antidepressants — clomipramine and imipramine — are effective but second-line because of anticholinergic burden, cardiac conduction effects, and overdose toxicity.^2,5 Reserve for patients who have failed two SSRI/SNRI trials, with baseline ECG and gradual titration. Monoamine oxidase inhibitors (phenelzine) have evidence for refractory cases but the dietary and drug interaction burden limits routine use. Benzodiazepines — alprazolam, clonazepam, and lorazepam — produce rapid relief of panic and are FDA-approved for the indication. Their role is contested. Major guidelines vary in their positioning, with NICE explicitly recommending against benzodiazepine use for panic disorder beyond short courses, while APA and CANMAT permit them as second-line or as short-term adjuncts during SSRI initiation.^2-4 Clinically defensible uses include a short bridge (two to four weeks) during SSRI titration to mitigate early activation and improve adherence, and longer use in patients with severe symptoms who have failed SSRI/SNRI/TCA trials and CBT, with documented informed consent regarding tolerance, dependence, cognitive effects, falls, and complex withdrawal. Avoid in patients with substance use histories.

Psychotherapy

Cognitive behavioral therapy is the best-evidenced psychotherapy for panic disorder, with strong evidence supporting efficacy versus waitlist and active controls and durable effects after treatment ends.⁷ A standard course is 12-16 sessions delivered individually or in groups. Core components include psychoeducation about the panic cycle, cognitive restructuring of catastrophic misinterpretations, breathing retraining (used judiciously, since over-reliance can become a safety behavior), interoceptive exposure (deliberate induction of feared bodily sensations such as hyperventilation, spinning, or running in place to extinguish the conditioned fear), and in vivo exposure to avoided situations when agoraphobia is present. Internet-delivered CBT (iCBT) with therapist support has accumulated robust evidence and is a reasonable alternative when access to in-person therapy is limited.¹⁹ Panic-focused psychodynamic psychotherapy has more limited but positive RCT evidence and may be considered when CBT is unavailable or unacceptable.²⁰ Mindfulness-based interventions and acceptance and commitment therapy show promise but have a smaller evidence base in panic disorder specifically.

Neuromodulation

Neuromodulation has no established role in routine panic disorder treatment. Repetitive transcranial magnetic stimulation has been studied in small trials with inconsistent results and is not guideline-recommended.⁵ Electroconvulsive therapy is not indicated for panic disorder absent comorbid severe depression with treatment-resistant features.

Adjunctive

Caffeine reduction — and, in heavy users, complete elimination during the early treatment phase — is a low-cost intervention that often produces rapid symptomatic benefit. Smoking cessation is similarly worthwhile given the association between nicotine use and panic. Aerobic exercise has limited but positive evidence as a stand-alone and adjunctive treatment.²¹ Sleep optimization and reduction of alcohol use should be addressed routinely. Self-help materials and peer support can supplement formal treatment but should not replace evidence-based therapy in moderate-to-severe disease.

Pharmacotherapy harms are predictable and largely class-based. SSRIs cause early activation in roughly 10-20% of panic patients during the first one to two weeks — a paradoxical worsening of anxiety, jitteriness, and insomnia that is the most common reason for early discontinuation and the principal rationale for low starting doses.¹⁸ Sexual dysfunction occurs in 30-60% of patients on sustained SSRI treatment and is dose-related; weight gain is more variable. Citalopram carries a dose-dependent QT prolongation warning above 40 mg daily (20 mg in patients over 60 or with relevant interactions); paroxetine has the most pronounced anticholinergic burden and the most difficult discontinuation syndrome among SSRIs. SSRI/SNRI use in late pregnancy carries a small risk of neonatal adaptation syndrome and a smaller absolute risk of persistent pulmonary hypertension of the newborn.²² SNRI venlafaxine produces dose-dependent diastolic hypertension above 150 mg daily and a particularly difficult discontinuation syndrome owing to its short half-life. TCAs carry the harms familiar from depression treatment: anticholinergic burden, orthostasis, sedation, weight gain, conduction effects (QRS widening, QT prolongation), and lethality in overdose with as little as a 1-2 week supply. MAOIs require tyramine dietary restriction and carry serotonin syndrome risk with co-administered serotonergic agents; a 14-day washout from SSRI to MAOI (5 weeks for fluoxetine) is mandatory. Benzodiazepines carry the harms that drive guideline ambivalence: tolerance to therapeutic effect, physiological dependence, withdrawal that can include seizures and protracted symptoms, falls and hip fractures (particularly in older adults), motor vehicle crash risk, cognitive blunting that may persist after discontinuation, and a substantial overdose hazard when combined with opioids or alcohol.²³ The FDA boxed warning regarding co-prescription with opioids reflects this risk. CBT harms are modest but real. Interoceptive and in vivo exposure produce transient distress, and a minority of patients drop out during the exposure phase. Therapist quality varies widely; outcomes from non-CBT-trained therapists labeling treatment as CBT are weaker than the controlled-trial literature implies. Limitations of the evidence base merit explicit mention. Most pharmacotherapy RCTs are 8-12 weeks in duration, with relatively few studies extending beyond a year, and long-term comparative effectiveness data are sparse. Trial populations have historically underrepresented older adults, pregnant patients, and patients with active substance use disorders or significant medical comorbidity. Publication bias favors positive trials, and effect sizes in pragmatic real-world settings tend to be smaller than in industry-sponsored efficacy trials.

Pediatric panic disorder is uncommon before puberty and uncommon to diagnose before adolescence. When it occurs, CBT is first-line; SSRI use parallels adult practice with attention to the FDA boxed warning regarding suicidality in patients under 25 and with informed consent and close follow-up during initiation.²⁴ Sertraline and fluoxetine have the most pediatric data across anxiety disorders. Geriatric onset is uncommon and should prompt a thorough medical workup before settling on a primary diagnosis. When the diagnosis is established in older adults, SSRIs at reduced starting doses are first-line, with careful attention to hyponatremia (SIADH), bleeding risk with concurrent antiplatelet or anticoagulant therapy, falls, and drug-drug interactions. Benzodiazepines carry an unfavorable risk-benefit profile in this population (Beers Criteria) and should be avoided where possible. Perinatal management requires shared decision-making weighing untreated illness against medication exposure. Untreated panic disorder in pregnancy is associated with preterm birth, low birth weight, and impaired maternal-infant bonding. SSRIs (sertraline, fluoxetine, citalopram, escitalopram) are generally compatible with pregnancy and lactation; paroxetine has been associated with a small absolute increase in cardiac malformations in some studies and is typically avoided when alternatives exist.²² CBT is the preferred non-pharmacologic option in pregnancy and the postpartum. Comorbid medical illness — particularly cardiovascular disease, asthma, and thyroid disease — alters the calculus. In ischemic heart disease the cardiotoxic profile of TCAs limits their use; SSRIs are safer and may have small cardiovascular benefits. In asthma, beta-blockers used for autonomic symptom management should be avoided. In hyperthyroidism, panic frequently remits with euthyroid restoration; pursue endocrine treatment first. Comorbid substance use disorders, particularly alcohol use disorder, must be addressed concurrently. Treating panic alone in an actively drinking patient produces poor outcomes. Benzodiazepines are relatively contraindicated; SSRIs and CBT are the workhorses, and integrated dual-diagnosis treatment is preferred when available. Cultural considerations include culturally specific panic-like presentations: ataque de nervios in some Caribbean and Latin American populations, kyol goeu in Cambodian populations, and trung gio attacks in Vietnamese populations. These overlap but do not map cleanly onto DSM-5-TR panic attacks, and clinicians should elicit the patient's own explanatory model rather than impose the diagnostic label first.

Panic disorder is typically chronic and relapsing without treatment. Naturalistic follow-up studies show that without sustained intervention, roughly 30-40% of patients are symptom-free in long-term follow-up, 50% have continued mild symptoms, and 10-20% remain moderately to severely impaired.⁸ With evidence-based treatment outcomes are substantially better: response rates (typically defined as 50% reduction in panic frequency or PDSS) of 60-80% are typical in 8-12 week pharmacotherapy and CBT trials, and panic-free status at endpoint is achieved by roughly half of treated patients.^6-7 Relapse following pharmacotherapy discontinuation is common. Rates of 25-50% within 6-12 months of stopping SSRI/SNRI treatment are reported, which underpins the guideline recommendation to continue successful pharmacotherapy for at least 12 months after remission before considering taper.^2-4 CBT-treated patients have lower relapse rates than pharmacotherapy-treated patients in head-to-head follow-up, plausibly reflecting durable acquisition of coping skills. Functional outcome correlates more strongly with the presence and extent of agoraphobia than with panic frequency per se. Patients with severe agoraphobia have more occupational impairment, more medical service utilization, and lower quality of life. Suicide risk is elevated in panic disorder both with and without comorbid depression; the elevation is real and persistent, and routine suicide screening is warranted at every visit.⁹

Most panic disorder is managed in outpatient settings. Indications for inpatient admission are uncommon and overlap substantially with general psychiatric admission criteria: active suicidal ideation with intent or plan, severe self-neglect, comorbid major depression with safety concerns, or comorbid substance use requiring detoxification. Panic disorder itself, however severe, is rarely the sole indication for admission. In the emergency department, the patient who presents during or immediately after a panic attack should receive a focused medical evaluation proportionate to the presentation, calm reassurance, and a slow-breathing intervention. Single-dose oral lorazepam 0.5-1 mg is reasonable when symptoms are severe and the patient is not taking other CNS depressants. The more important emergency-department task is to make the diagnosis explicit, communicate it clearly to the patient, and arrange outpatient psychiatric follow-up; repeated negative workups without diagnostic closure perpetuate health anxiety and drive further utilization. Suicide risk warrants explicit attention. Panic disorder confers a roughly two- to threefold increase in lifetime suicide attempts relative to the general population, with risk concentrated among patients with comorbid depression, substance use, or borderline personality traits.⁹ Routine assessment with a structured tool (such as the Columbia Suicide Severity Rating Scale) and direct inquiry about ideation and access to means are warranted at every visit during acute treatment. Agitation in the panic patient is rare; if present and severe, it should prompt re-examination of the diagnosis (consider stimulant intoxication, hyperthyroidism, akathisia from a recently started SSRI). De-escalation, environmental control, and oral benzodiazepine are first-line if behavioral intervention is needed.

Several areas remain unsettled. The role of benzodiazepines in routine panic disorder treatment is the most prominent guideline disagreement: NICE recommends against, APA permits with caveats, CANMAT and WFSBP allow short-term or refractory use.^2-5 The disagreement reflects a genuine tension between rapid efficacy and long-term harms, and clinicians should be prepared to defend their position with explicit reference to the literature and to individualized risk-benefit reasoning. The respiratory subtype of panic disorder — patients with prominent dyspnea, choking sensations, and CO2 hypersensitivity — has accumulated supporting evidence but is not a DSM-5-TR specifier and its treatment implications remain debated.¹³ Some authors suggest such patients respond preferentially to TCAs, but the evidence is suggestive rather than definitive. The optimal duration of pharmacotherapy after remission is uncertain. Guidelines converge on at least 12 months but acknowledge limited high-quality long-term comparative data. Whether some patients require indefinite treatment, and how to identify them, remains unsettled. Emerging treatments include ketamine (limited data in anxiety disorders, mostly extrapolated from depression literature), psilocybin-assisted therapy (early-stage research with no panic-specific RCTs), cannabidiol (small studies in social anxiety, none specific to panic disorder), and second-generation digital therapeutics. None has guideline support for panic disorder at present, and clinicians should be cautious about extrapolating from the depression and PTSD literatures.

• Panic disorder requires recurrent unexpected panic attacks plus at least one month of anticipatory worry or maladaptive behavioral change.
• A panic attack peaks within minutes and requires four or more of thirteen DSM-5-TR symptoms; attacks alone do not equal panic disorder.
• DSM-5 unbundled agoraphobia from panic disorder; both can now be diagnosed independently or together.
• Twelve-month prevalence is approximately 2-3%; women are affected roughly twice as often as men; median onset is in the early-to-mid 20s.
• First onset after age 45 should prompt a careful medical workup before settling on a primary psychiatric diagnosis.
• SSRIs and SNRIs are first-line pharmacotherapy; start low (e.g., sertraline 25 mg, escitalopram 5 mg) to mitigate early activation.
• CBT with interoceptive and in vivo exposure is first-line psychotherapy; standard course is 12-16 sessions with durable post-treatment effects.
• Continue successful pharmacotherapy for at least 12 months after remission before considering taper.
• Benzodiazepines provide rapid relief but are positioned variably across guidelines; NICE recommends against routine use, APA and CANMAT permit short-term or refractory use.
• Citalopram has a dose-dependent QT warning above 40 mg daily (20 mg in older adults); paroxetine has the most difficult discontinuation syndrome among SSRIs.
• Lifetime comorbidity with major depressive disorder approaches 50-60%; panic disorder independently elevates suicide risk.
• The CO2 inhalation, sodium lactate, and CCK-4 challenges preferentially provoke panic in panic disorder patients, supporting the suffocation false alarm hypothesis.
• Medical mimics include hyperthyroidism, pheochromocytoma, supraventricular tachycardia, temporal lobe epilepsy, and stimulant intoxication.
• Buspirone is not effective for panic disorder despite its efficacy in generalized anxiety disorder.
• The Panic Disorder Severity Scale (PDSS) is the standard severity measure for diagnosis and treatment monitoring.

No external funding. No conflicts of interest declared. Drafted with Claude (Anthropic) and human-edited by the Psychpedia editorial team. Peer-review status: pending.