This article synthesizes current guidance from the APA, NICE Clinical Guideline 159, the Canadian Anxiety Guidelines Initiative (CANMAT), the British Association for Psychopharmacology (BAP) anxiety guidelines, and the Maudsley Prescribing Guidelines, supplemented by Cochrane reviews of pharmacotherapy and psychotherapy for social anxiety disorder, the network meta-analysis by Mayo-Wilson and colleagues in Lancet Psychiatry, and reference texts including Stahl's Essential Psychopharmacology and Kaplan and Sadock's Synopsis of Psychiatry. Search terms included "social anxiety disorder," "social phobia," "SSRI," "SNRI," "cognitive behavioral therapy," "exposure therapy," and "network meta-analysis," with priority given to guidelines and systematic reviews published since 2014. The literature pass for this draft was completed on 2026-05-09.

Social anxiety disorder is among the most common psychiatric disorders. The U.S. National Comorbidity Survey Replication estimated lifetime prevalence at approximately 12% and 12-month prevalence at approximately 7%, placing it third behind specific phobia and major depressive disorder among lifetime mental disorders.¹ Cross-national rates are lower in non-Western samples, with the World Mental Health Surveys reporting lifetime prevalence of roughly 4% across surveyed countries, suggesting both real cultural variation and methodological differences in case ascertainment.² Onset is characteristically early. The median age of onset is 13 years, with 75% of cases beginning between ages 8 and 15.^1,3 Onset after age 25 is uncommon and should prompt a search for medical, substance-related, or secondary causes. The female-to-male ratio in community samples is approximately 1.5 to 2:1, though clinical samples are closer to 1:1, reflecting referral and help-seeking patterns rather than true incidence.³ Comorbidity is the rule, not the exception. Approximately 70-80% of patients with social anxiety disorder meet criteria for at least one additional lifetime disorder, most commonly another Anxiety disorders, Major depressive disorder, and substance use disorders.^1,4 Social anxiety disorder typically precedes its comorbidities by 5-15 years, and longitudinal data suggest it is a robust prospective risk factor for the later development of major depression and alcohol use disorder.⁴ Suicidal ideation is elevated relative to the general population, particularly in the presence of depression, but completed suicide attributable specifically to social anxiety disorder is rare.⁵ Risk factors include behavioral inhibition in childhood (a temperamental construct identifying children who withdraw from novelty), parental overprotection or criticism, peer victimization, and a family history of anxiety disorders. The heritability estimate from twin studies is approximately 30-40%, with substantial overlap with neuroticism and the broader internalizing factor.⁶

Social anxiety disorder is best modeled as the convergence of a genetically influenced temperamental vulnerability, neurodevelopmental sensitization of fear circuitry, and learned cognitive-behavioral patterns that maintain avoidance. No single mechanism dominates. At the circuit level, functional neuroimaging consistently shows hyperactivation of the Amygdala and insula in response to social threat cues such as critical or fearful faces, with reduced top-down regulation from the medial prefrontal cortex and ventrolateral prefrontal cortex.⁷ This pattern parallels findings in other anxiety disorders but is most pronounced for socially evaluative stimuli. Resting-state studies suggest altered connectivity within the salience network and between the Default mode network and prefrontal regulatory regions, which has been proposed as a substrate for the persistent self-focused attention characteristic of the disorder.⁷ At the neurotransmitter level, serotonergic and dopaminergic dysregulation are both implicated, though neither is sufficient on its own. The therapeutic efficacy of SSRIs and SNRIs supports a serotonergic contribution, and PET studies have shown reduced striatal dopamine transporter and D2 receptor binding in some samples, consistent with the long-standing association between social anxiety and reward-system hypofunction.⁸ GABAergic and glutamatergic systems are likely involved in fear extinction and represent ongoing targets of investigation. Genetic studies consistently show familial aggregation. Twin heritability is moderate, around 30-40%, and genome-wide association studies have identified several candidate loci, none with large effect.⁶ Polygenic risk scores for neuroticism and depression overlap substantially with social anxiety risk, supporting a shared internalizing diathesis. Cognitive-behavioral models, formalized by Clark and Wells and by Rapee and Heimberg, remain the most clinically useful integrative frameworks. They posit that biased self-focused attention, distorted appraisals of social performance, anticipatory and post-event rumination, and safety behaviors (such as avoiding eye contact, scripting speech, or holding a glass tightly to mask tremor) prevent disconfirmation of feared outcomes and maintain the disorder over time.⁹ These cognitive constructs are the explicit targets of Cognitive behavioral therapy.

DSM-5-TR defines social anxiety disorder by marked, persistent fear or anxiety about one or more social situations in which the person is exposed to possible scrutiny by others. Examples include social interactions (conversations, meeting unfamiliar people), being observed (eating, drinking), and performing in front of others (giving a speech). The individual fears showing anxiety symptoms or behaving in a way that will be negatively evaluated, such as appearing awkward, leading to humiliation, embarrassment, rejection, or offense to others.³ The feared situations almost always provoke fear or anxiety and are actively avoided or endured with intense distress. The fear is out of proportion to the actual threat posed by the situation and to the sociocultural context. Symptoms persist for six months or longer, cause clinically significant distress or impairment, and are not better explained by the physiological effects of a substance or another medical condition, nor by another mental disorder. If a medical condition such as Parkinson disease, obesity, or a disfiguring injury is present, the social anxiety must be clearly unrelated or excessive.³ DSM-5-TR retains a single specifier: "performance only," indicating that the fear is restricted to speaking or performing in public. The performance-only subtype is clinically distinct, has later onset, less comorbidity, and tends to respond well to as-needed beta-blockade in addition to or instead of standard treatment.^3,10 The earlier DSM-IV "generalized" specifier was removed because the dimensional distinction between generalized and non-generalized forms proved unreliable. Three points warrant attention at the bedside. First, the diagnosis requires that the individual recognize the fear as excessive only insofar as is developmentally appropriate; in children, the criterion of insight is relaxed and crying, tantrums, freezing, clinging, or failure to speak (selective mutism, a related disorder) may substitute. Second, the six-month duration applies to all ages, distinguishing the diagnosis from transient situational anxiety. Third, the disorder must cause significant distress or impairment, which is what separates social anxiety disorder from clinically unimpairing shyness.

The phenotype is dominated by anticipatory anxiety, in-situation autonomic arousal, post-event processing, and avoidance. Patients describe rehearsing benign interactions for hours and replaying them for days, scrutinizing what they said, how they looked, and what others must have thought. The autonomic signature in feared situations is sympathetic surge: tachycardia, blushing, tremor, sweating, dry mouth, and the urgent wish to escape. Blushing is so characteristic that some clinicians treat it as a near-diagnostic feature when prominent, though specificity is imperfect.⁹ Avoidance is often subtle. Patients structure their lives to minimize exposure: choosing jobs without public-facing duties, avoiding the cafeteria, declining promotions that would require presentations, drinking before parties, and using text rather than voice calls. The clinician who screens only for overt avoidance will miss high-functioning patients who appear successful but are exhausted by the constant management of social threat. A 24-year-old graduate student presents with insomnia and intermittent palpitations before lab meetings. She denies depression and reports that she enjoys her research. On structured questioning she has avoided every elective requiring a presentation since high school, has never asked a question at a seminar, and turned down a teaching fellowship that would have funded her degree. The presentation looks like cardiac anxiety; the diagnosis is social anxiety disorder, performance-prominent, missed because the screening question was "do you feel anxious" rather than "what have you avoided." The course is typically chronic and unremitting without treatment. Spontaneous remission rates over 10-12 years are low, on the order of 35-40%, with most cases waxing and waning around a stable functional ceiling.¹¹ Comorbid depression and alcohol use commonly emerge in the third decade, often as the patient's avoidance fails to scale with adult role demands. Atypical presentations include selective mutism in children, prominent paruresis ("shy bladder"), and patients whose dominant complaint is a single feared symptom such as blushing, sweating, or tremor; these are variants, not separate disorders. Red flags arguing for an alternative or co-occurring diagnosis include onset after age 25, sudden onset, paranoid quality to social fears, prominent negative symptoms, fear of contamination rather than evaluation, intrusive autobiographical memories of trauma, and pervasive social discomfort dating from early childhood without a clear evaluative focus (suggesting autism spectrum disorder).

Normative shyness lacks the impairment, persistence, and out-of-proportion quality required for the diagnosis. The duration criterion (six months) and explicit functional impairment requirement are the primary safeguards against overdiagnosis.³ Panic disorder produces unexpected panic attacks and fear of the attacks themselves; the patient with social anxiety disorder fears social evaluation and may have situationally bound panic attacks, but anxiety remits when alone. Agoraphobia is differentiated by fear of being unable to escape or get help, not fear of scrutiny; agoraphobic patients are often reassured by a trusted companion, while patients with social anxiety disorder are not. Generalized anxiety disorder centers on uncontrollable worry across multiple domains, including but not limited to social ones, and is accompanied by tonic somatic symptoms (muscle tension, restlessness, sleep disturbance). Overlap is substantial; both diagnoses are made when criteria for each are met independently. Major depressive disorder may produce social withdrawal, but the withdrawal is driven by anhedonia, fatigue, and hopelessness rather than fear of evaluation, and remits as depression improves. Persistent social withdrawal that predates and outlasts depressive episodes argues for an independent social anxiety diagnosis. Autism spectrum disorder shares social discomfort but reflects deficits in social communication, restricted interests, and repetitive behaviors with onset in early development; the patient with autism may want connection but lacks the implicit social skills, while the patient with social anxiety has the skills and is paralyzed by their fear of being judged on them. The two can co-occur, and recognition of social anxiety in autistic patients is clinically important and often delayed.¹² Body dysmorphic disorder presents as fear of being judged for a perceived physical defect; the focus is on the imagined flaw, not generalized social evaluation. Avoidant personality disorder shares the phenomenology of social anxiety disorder so closely that many investigators consider it a severity variant rather than a separate entity; in DSM-5-TR they remain distinct, with avoidant personality disorder requiring pervasive feelings of inadequacy and hypersensitivity to negative evaluation across nearly all interpersonal contexts.³ Schizophrenia and schizotypal personality disorder may present with social withdrawal, but social fears in those conditions tend to have a paranoid or referential quality rather than fear of negative evaluation, and are accompanied by other psychotic-spectrum features. Medical and substance-induced mimics warrant attention, particularly when onset is late or atypical. Hyperthyroidism, pheochromocytoma, hypoglycemia, cardiac arrhythmia, vestibular disorders, and Parkinson disease can produce autonomic symptoms that the patient subsequently interprets as social anxiety. Stimulants (caffeine, amphetamines, cocaine), bronchodilators, sympathomimetics, and withdrawal from alcohol or benzodiazepines can mimic or exacerbate the disorder. A targeted medical workup is required when the history suggests it; a routine "anxiety panel" in every patient is not.

The interview is the primary diagnostic tool. Open with what the patient avoids rather than how they feel, since avoidance is more reliably reported than fear. Ask specifically about the feared situations on the DSM-5-TR exemplar list: speaking in meetings, eating in public, using public restrooms, being introduced, returning items to a store, dating. Quantify avoidance and impairment in concrete terms: classes dropped, jobs declined, friendships not pursued. Mandatory history elements include age of onset, course, prior treatments and responses, alcohol and substance use (including "liquid courage" before social events), depression and suicidal ideation, trauma history, family history of anxiety and mood disorders, and the patient's own treatment goals. A medication review should specifically capture stimulants, sympathomimetics, and recent changes in dose of any psychotropic. Physical examination is generally unremarkable. Thyroid examination, resting heart rate and rhythm, and observation for tremor are reasonable. Routine labs in a well-appearing patient with classic presentation are low yield; in late-onset, sudden-onset, or atypical presentations, consider TSH, basic metabolic panel, urine toxicology, and an EKG if cardiac symptoms are prominent. Neuroimaging is not indicated absent focal neurological signs. Validated rating scales are useful for severity grading, treatment monitoring, and research. The clinician-rated Liebowitz Social Anxiety Scale (LSAS) is the most widely used outcome measure in trials and clinically discriminates social anxiety disorder from controls at thresholds around 30, with severe disease typically scoring above 60.¹³ The patient-rated Social Phobia Inventory (SPIN) and its three-item screen, the Mini-SPIN, are practical for primary care; a Mini-SPIN score of 6 or above has reasonable sensitivity for case identification.¹⁴ The PHQ-9 and GAD-7 should be administered routinely to capture comorbid depression and generalized anxiety. Do not order "anxiety panels," cortisol curves, or QEEG. They have no role in routine diagnosis, generate false positives, and delay evidence-based care.

Once the diagnosis is settled, the harder question is what to do about it. First-line options for adults are an SSRI or SNRI, individual cognitive behavioral therapy with an exposure component, or both in combination. Major guidelines (NICE, CANMAT, BAP) converge on this menu, with NICE giving slight preference to individual CBT as initial monotherapy where access permits.^15-17 Choice between modalities is driven by patient preference, severity, comorbidity, prior response, and access.

Pharmacotherapy

Strong evidence supports SSRIs as first-line pharmacotherapy. Paroxetine, sertraline, escitalopram, and fluvoxamine are FDA-approved or have regulatory approval in other jurisdictions for social anxiety disorder, and the class effect is consistent across SSRIs.^15-17 Venlafaxine extended-release is the SNRI with the strongest evidence and is also FDA-approved.¹⁵ Effect sizes versus placebo in meta-analyses are moderate, with response rates of approximately 50-65% versus 30-35% for placebo and number needed to treat in the range of 4-6.¹⁸ Starting doses should be lower than those used for depression, because patients with social anxiety disorder are characteristically sensitive to early activation effects and prone to attribute somatic side effects to anxiety, leading to early discontinuation. Reasonable starting doses are sertraline 25 mg daily, escitalopram 5 mg daily, paroxetine 10 mg daily, fluvoxamine 50 mg daily, and venlafaxine XR 37.5 mg daily, titrated to standard target ranges over two to four weeks.¹⁹ Onset of clinical response is typically observed at 4-6 weeks, with continued gains through 12 weeks and beyond. An adequate trial is at least 12 weeks at a therapeutic dose before declaring nonresponse.¹⁵ Continuation for at least 6-12 months after response is recommended to consolidate gains and reduce relapse, with longer maintenance for patients who have had multiple episodes or severe functional impairment.^15,17 Discontinuation should be gradual; abrupt cessation, particularly of paroxetine and venlafaxine, produces a discontinuation syndrome with dizziness, paresthesias, flu-like symptoms, and rebound anxiety. Second-line options with reasonable evidence include other SSRIs (fluoxetine, citalopram), duloxetine (less data than venlafaxine), and pregabalin, which has shown efficacy in randomized trials at doses of 450-600 mg/day, though it is approved for this indication in Europe but not the United States.²⁰ Phenelzine, an irreversible MAOI, has historical evidence of high efficacy and is reserved for refractory cases given its dietary and drug-interaction burden.²¹ Moclobemide, a reversible MAOI available outside the United States, has more modest and less consistent effects. Beta-blockers (propranolol 10-40 mg taken 30-60 minutes before a feared performance) are commonly used for performance-only social anxiety, particularly in musicians, public speakers, and trainees facing oral examinations. Evidence is limited and largely from small trials and clinical experience rather than rigorous RCTs; benefits target peripheral autonomic symptoms (tremor, tachycardia) without altering subjective fear. Some experts recommend propranolol for performance-only presentations, though high-quality evidence is lacking.^10,15 Beta-blockers have no established role in generalized presentations. Benzodiazepines, principally clonazepam, can produce rapid relief but are not recommended as first-line treatment because of dependence, cognitive effects, falls risk in older adults, interaction with alcohol (a high-prevalence comorbidity in this population), and interference with the extinction learning that drives durable response to exposure therapy.^15,17 Their role is largely limited to short-term bridging in selected severe cases without substance use comorbidity.

Psychotherapy

Strong evidence supports cognitive behavioral therapy with an explicit exposure component as first-line psychological treatment, with effect sizes comparable to or modestly larger than pharmacotherapy at end of treatment and superior durability after treatment ends.^22-23 Cochrane meta-analysis of psychological therapies for social anxiety disorder found CBT, particularly in individual format, to be more effective than waitlist and treatment as usual, with moderate certainty of evidence.²² The two best-validated protocols are the cognitive therapy program developed by Clark and Wells, which targets self-focused attention and post-event processing through video feedback and behavioral experiments, and the cognitive behavioral group therapy of Heimberg, which combines in-session exposure with cognitive restructuring.^9,23 A typical course is 12-16 weekly sessions. Internet-delivered CBT (iCBT) with therapist support has demonstrated efficacy approaching that of face-to-face CBT in head-to-head trials and is recommended by NICE as a stepped-care option.^15,24 Exposure must target the patient's actual feared situations and must include dropping safety behaviors, since safety behaviors block the corrective learning that exposure is meant to produce. Pure exposure without cognitive restructuring is effective; cognitive restructuring without exposure is not. Mindfulness-based interventions and acceptance and commitment therapy show promise but with smaller and less consistent effects than CBT.²⁵ Psychodynamic psychotherapy has been studied in a smaller evidence base and may be considered when CBT is unavailable, declined, or has failed; manualized short-term psychodynamic therapy has shown efficacy in at least one large multicenter trial.²⁶

Neuromodulation

Neuromodulation has no established first-line role. Repetitive transcranial magnetic stimulation has been studied in small trials with mixed results and is not currently recommended in major guidelines for social anxiety disorder.^15,27 Electroconvulsive therapy is not indicated.

Adjunctive

Adjunctive strategies include treatment of comorbid major depression, substance use disorder, and other anxiety disorders, which often improves social anxiety outcomes substantially. Sleep optimization, exercise, and reduction of caffeine and alcohol are reasonable. D-cycloserine, a partial NMDA agonist hypothesized to enhance extinction learning when administered shortly before exposure sessions, was promising in early trials but has not shown consistent benefit in larger studies and is not part of routine practice.²⁸ MDMA-assisted psychotherapy and psilocybin have been studied for related indications but not for social anxiety disorder specifically and remain investigational. Patients who fail to respond to a 12-week trial of an adequately dosed SSRI or SNRI plus CBT should be reassessed for diagnostic accuracy, comorbidity, treatment adherence, and substance use before being labeled treatment-resistant. Switching within or across antidepressant classes, augmenting with pregabalin, or trying phenelzine in carefully selected patients are reasonable next steps in consultation with a specialist.^15,17

SSRI and SNRI adverse effects are dose-related and time-limited for most patients. Early-treatment effects (first one to two weeks) include nausea, headache, restlessness, transient increases in anxiety, insomnia, and rarely jitteriness severe enough to limit titration. Persistent effects include sexual dysfunction (orgasmic delay, decreased libido, anorgasmia) in up to 30-50% of treated patients, weight changes, sweating, and bruxism.¹⁹ Hyponatremia, particularly in older adults, and bleeding risk (especially with concomitant NSAIDs or anticoagulants) deserve mention. Discontinuation syndromes are most pronounced with paroxetine and venlafaxine and warrant cross-titration or slow taper. Suicidal ideation in young people on SSRIs is the subject of an FDA black-box warning covering patients up to age 24. The absolute risk increase is small, the underlying disorder itself elevates suicide risk, and most meta-analytic and observational data support a net benefit for treated youth, but informed consent and close monitoring during the first weeks of treatment are required.²⁹ CBT harms are modest. Transient symptom worsening during early exposure exercises is expected and clinically managed; rare patients drop out for this reason. Time and access burdens are real, particularly outside metropolitan areas; iCBT addresses access but has higher dropout in unguided formats. Pregabalin causes sedation, weight gain, peripheral edema, and has misuse potential, particularly in patients with substance use histories. Phenelzine carries the dietary tyramine restriction (avoid aged cheeses, cured meats, fermented soy products, draft beer, certain wines), risk of hypertensive crisis with sympathomimetics, and serotonin syndrome risk if combined with serotonergic agents; a two-week washout is required when switching from or to most antidepressants.²¹ Limitations of the evidence base include short trial durations (most pharmacotherapy RCTs are 12 weeks), narrow demographic sampling (heavy overrepresentation of white, middle-class, treatment-seeking patients), exclusion of substantial substance use comorbidity from many trials, sparse pediatric data outside small RCTs and open-label extensions, and possible publication bias favoring positive trials. Comparative effectiveness data between specific SSRIs are limited and most evidence supports a class effect rather than within-class superiority.¹⁸

Pediatric patients show a similar evidence base with smaller numbers. Sertraline and fluoxetine have the strongest pediatric anxiety data, including the landmark Child/Adolescent Anxiety Multimodal Study (CAMS), which compared sertraline, CBT, their combination, and placebo across separation anxiety, generalized anxiety, and social anxiety in youth aged 7-17. Combination treatment was superior to either monotherapy, and both monotherapies were superior to placebo, with response rates of approximately 81% for combination, 60% for CBT, 55% for sertraline, and 24% for placebo at 12 weeks.³⁰ CBT is the preferred first-line modality in children where access permits; medication is added for moderate-to-severe disease, partial response, or when CBT is unavailable. The FDA black-box warning applies. Geriatric patients are underrepresented in trials. Late-onset cases (after age 60) are uncommon and warrant a careful workup for medical contributors, medication effects, and neurocognitive disorder. SSRI doses should be lower at start, drug interactions reviewed (particularly QT-prolonging combinations with citalopram, anticoagulant interactions, and falls risk), and hyponatremia monitored. Benzodiazepines should generally be avoided in older adults given falls, fractures, and cognitive risks (Beers Criteria). Perinatal patients require an explicit risk-benefit conversation. Untreated maternal anxiety carries its own risks (preterm delivery, low birth weight, postpartum depression, impaired bonding). Sertraline and escitalopram are commonly chosen in pregnancy and lactation given their safety profiles; paroxetine has been associated with a small increase in cardiac malformations in some studies and is generally avoided when alternatives are acceptable.³¹ Psychotherapy is preferred when feasible. Decisions should be individualized and documented. Comorbid medical illness modifies choice. Cardiac disease and arrhythmia argue against high-dose citalopram and against tricyclic antidepressants. Hepatic impairment requires dose reduction for several SSRIs. Patients with bleeding diatheses, recent gastrointestinal bleeding, or concurrent NSAID/anticoagulant use require risk-benefit reassessment. Asthma and reactive airway disease are relative contraindications to non-selective beta-blockers; cardio-selective agents (e.g. atenolol) are alternatives, though propranolol remains the agent with the most performance-anxiety data. Comorbid substance use, particularly alcohol use disorder, is high-prevalence and changes the calculus. Treat both disorders concurrently; sequential treatment of alcohol use disorder followed by social anxiety disorder rarely produces durable remission of either. Avoid benzodiazepines. SSRIs and CBT remain first-line; gabapentin and naltrexone are reasonable adjuncts for alcohol use disorder when indicated. Cultural considerations matter. The phenotypic emphasis on fear of offending others (taijin kyofusho) in Japanese and Korean clinical traditions overlaps substantially with social anxiety disorder but foregrounds the patient's perceived effect on others rather than personal humiliation. DSM-5-TR retains this as a cultural concept. Across cultures, fear of negative evaluation is a stable construct; its content varies.³

Without treatment, social anxiety disorder follows a chronic, often lifelong course with limited spontaneous remission. Naturalistic follow-up data from the Harvard Anxiety Research Project showed full recovery in approximately 35-40% over 12 years, with most patients oscillating around a partial-remission baseline.¹¹ Earlier age of onset, more severe baseline impairment, and avoidant personality features predict worse course; later onset, performance-restricted presentations, higher educational attainment, and absence of comorbid depression predict better course. With treatment, response rates of 50-65% to first-line SSRI/SNRI or CBT translate to remission rates of approximately 30-40% at 12 weeks, with continued gains over the next year on maintenance therapy. Combination of pharmacotherapy and CBT is commonly recommended for severe disease or partial response and probably outperforms either monotherapy in adolescents (CAMS); in adults the additive benefit is real but modest.^15,30 Relapse after medication discontinuation is common, with rates of approximately 30-50% in the year following taper, supporting maintenance for at least 6-12 months and considerably longer for severe or recurrent cases.¹⁵ CBT-treated patients show better durability, with maintained gains at 1-year and 5-year follow-up in cohort studies, though the literature has limitations of attrition and selection.²³ Functional outcomes are the meaningful endpoint. Successful treatment is reflected not only in symptom scores but in the patient's resumption of declined opportunities: classes attended, jobs accepted, relationships pursued. Social anxiety disorder is associated with reduced educational attainment, lower lifetime earnings, lower marriage rates, and higher unemployment, much of which is recoverable when the disorder is treated.^1,4 Mortality data specific to social anxiety disorder are limited. Excess mortality, where observed, is mediated largely through comorbid depression and alcohol use disorder rather than the disorder itself.⁵

Social anxiety disorder rarely produces psychiatric emergencies on its own. Hospitalization is uncommon and almost always driven by comorbid major depression with active suicidal ideation, substance intoxication or withdrawal, or rare panic-related presentations. The clinician's job is to recognize the common comorbid drivers of acute risk rather than to manage social anxiety acutely. Suicide risk should be assessed at every visit using the same approach as in any patient with a chronic anxiety disorder, with particular attention to comorbid depression, alcohol use, and recent worsening of social or occupational function. Adolescents, young adults on newly initiated SSRIs, and patients with prior attempts deserve closer follow-up during the first weeks of treatment.²⁹ Agitation specific to social anxiety disorder is uncommon. When acute autonomic distress in a feared situation is severe enough to require intervention, removing the stimulus and providing reassurance are usually sufficient. Acute pharmacologic management with a benzodiazepine should be a rare and specifically justified intervention given the harms outlined above.

The boundary between social anxiety disorder and avoidant personality disorder remains contested. Empirical work suggests they lie on a single severity continuum, with avoidant personality disorder representing the more impaired end; DSM-5-TR retains the categorical distinction for clinical and historical reasons.¹² Whether early identification and treatment of childhood behavioral inhibition prevents later social anxiety disorder is an active research question. Small trials of parent-mediated interventions have shown promise in reducing incidence at follow-up, but population-level prevention programs are not established practice.³² The role of cannabis is unsettled. Patients commonly self-medicate, citing acute symptom relief; epidemiological data link heavy cannabis use with worse anxiety outcomes over time, and cannabis-induced anxiety is a recognized differential consideration.³³ Pharmaceutical-grade cannabidiol has been studied in small trials with positive but preliminary results, and is not part of mainstream practice. Psychedelic-assisted therapy for social anxiety disorder is investigational. A small open-label trial of MDMA-assisted psychotherapy in autistic adults with social anxiety reported sustained symptom reduction; replication is awaited and regulatory approval has not occurred.³⁴ Digital and AI-driven exposure tools, including virtual reality exposure therapy for public speaking and conversation simulation, show promise in pilot data and have entered some specialty programs. The evidence base is moderate for VR-based exposure approaching the efficacy of in vivo exposure for circumscribed performance fears, with expanding but not yet definitive support for broader presentations.³⁵ Finally, screening for social anxiety disorder in primary care is recommended by some authorities and not others; over-medicalization of normative shyness remains a credible concern, balanced against the clear underdiagnosis of impairing disease.

• Lifetime prevalence in U.S. samples is approximately 12%; median age of onset is 13 years; female-to-male ratio in the community is roughly 1.5-2:1.
• DSM-5-TR requires duration of six months or more, fear out of proportion to actual threat, and clinically significant distress or impairment, with a single "performance only" specifier.
• First-line pharmacotherapy: SSRIs (sertraline, escitalopram, paroxetine, fluvoxamine) and the SNRI venlafaxine XR. Class effect is consistent; start low to avoid early activation.
• First-line psychotherapy: individual CBT with exposure (Clark-Wells, Heimberg). Internet-delivered, therapist-supported CBT is an evidence-based stepped-care option.
• Comorbidity is the rule: depression, alcohol use disorder, and other anxiety disorders co-occur in 70-80% of cases. Social anxiety disorder typically precedes its comorbidities by 5-15 years.
• Performance-only social anxiety responds to as-needed propranolol; this is not appropriate for generalized presentations.
• Avoid benzodiazepines as monotherapy and avoid them in patients with alcohol or sedative use disorders.
• Phenelzine (irreversible MAOI) has historical evidence of efficacy and is reserved for refractory disease; requires tyramine-restricted diet and a two-week washout from serotonergic agents.
• Pediatric trial benchmark: CAMS demonstrated combination CBT plus sertraline (~81% response) > CBT or sertraline alone (~55-60%) > placebo (~24%).
• Behavioral inhibition in childhood is the strongest temperamental predictor of later social anxiety disorder.
• Functional neuroimaging shows amygdala-insula hyperactivation to social threat with reduced prefrontal regulation.
• Differential: agoraphobia (fear of escape), panic disorder (fear of attacks themselves), GAD (worry across domains), autism spectrum disorder (social communication deficit), body dysmorphic disorder (perceived defect), avoidant personality disorder (severity continuum).
• Liebowitz Social Anxiety Scale (LSAS) is the standard outcome measure; Mini-SPIN is a practical screen.
• Maintenance pharmacotherapy is recommended for at least 6-12 months after response; relapse rates after discontinuation reach 30-50% in the first year.
• The disorder is chronic and underdiagnosed; spontaneous remission rates are low (~35-40% over 12 years untreated).

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