What it is:A group of disorders where fear and worry stop being useful and start being disabling. This chapter covers three of them: generalized anxiety disorder (GAD), chronic diffuse worry; panic disorder (PD), sudden discrete attacks of terror; and PTSD, intrusive re-living of a specific trauma. Different triggers, overlapping biology, mostly overlapping drug list.
The core problem:An overactive amygdala and an underactive prefrontal cortex. The amygdala fires the threat alarm, the PFC is supposed to talk it down, and in these disorders that top-down brake doesn't work well enough.
What you do about it:SSRIs or SNRIs are first-line pharmacotherapy for all three. CBT (bottom line, exposure-based) is equally effective and more durable, but slower to start working. Benzodiazepines work fastest and best short-term, but you don't want to leave someone on one.
Think of it as a top-down vs bottom-upproblem. CBT is top-down. It strengthens the prefrontal cortex's ability to regulate the amygdala, which is why its benefits last for months after therapy ends. Pharmacotherapy is bottom-up. It directly dampens limbic and brainstem circuitry (serotonin, norepinephrine, GABA), which works faster but the benefit fades once you stop the drug. That's why relapse after stopping medication is so common and why guidelines push combination therapy when access allows.
Before diving into each one, get the shape of the differences straight. This is the fastest way to avoid mixing up diagnostic criteria on an exam.
Excessive worry about everything, most days, for 6+ months. No single trigger.
Recurrent, unexpected 10-minute attacks of terror, then fear of the next one.
Specific trauma exposure, then intrusion, avoidance, and hyperarousal for 1+ month.
| Feature | GAD | Panic Disorder | PTSD |
|---|---|---|---|
| Trigger | None specific, free-floating | None specific, "out of the blue" | Specific identifiable trauma |
| Time course | Chronic, most days ≥6 months | Attacks peak in 10 min, resolve in 20-30 min | Symptoms >1 month post-trauma |
| Core feature | Worry that's hard to control + ≥3 somatic symptoms | Fear of losing control or dying during the attack | Intrusive memories, avoidance, negative mood shift, hyperarousal |
| Demographics | Women 2x men, onset ~age 21 | Up to 50% develop agoraphobia | Co-occurs heavily with depression, substance use |
Exam writers love handing you a vignette with panic-attack symptoms and asking you to pick between PD and GAD with panic features. The tell: in PDthe attacks are the disease (recurrent, unexpected, followed by ≥1 month of worry about having another one). In GAD, anxiety is the constant baseline state, not discrete episodes. If the stem says "most days for 6 months," it's GAD even if panic-like symptoms show up during flares.
All three disorders trace back to the same fear circuit misfiring in different ways, which is why the same drug classes keep showing up across all of them.
| Structure/System | Role | What goes wrong |
|---|---|---|
| Amygdala | Bottom-up threat detection, fear generation | Hyperactive in all three; in PTSD it drives persistence of traumatic memory |
| Prefrontal cortex / ACC | Top-down regulation, cognitive control | Under-regulates the amygdala; CBT specifically strengthens this connection |
| Locus coeruleus (LC) | Fires norepinephrine, drives sympathetic/parasympathetic activation | Hypersensitive and overreactive. Down-regulates α2-receptors in GAD/PTSD; that receptor is hypersensitive instead in PD |
| GABA system | Main inhibitory brake on CNS excitability | Under-functioning; benzodiazepines are a positive allosteric modulator here, which is why they work in minutes |
| Serotonin (5-HT) | Modulates mood, facilitates avoidance behavior | Excess stimulatory 5-HT transmission may drive worry; paradoxically, boosting synaptic 5-HT with an SSRI is still the fix (down-regulates receptor sensitivity over weeks) |
| HPA axis / cortisol | Normally tempers the stress response | In PTSD specifically, patients hypersecrete corticotropin-releasing factor but have subnormal cortisol at the time of trauma, which may be a risk factor for PTSD developing in the first place |
Every drug class in this chapter maps onto that table. Benzodiazepinesboost GABA directly, so they work immediately but only patch the symptom. SSRIs/SNRIsraise synaptic serotonin (and norepinephrine for the SNRIs), which over 2-8 weeks down-regulates receptor sensitivity and quiets the whole circuit, which is why they're slower but disease-modifying rather than just symptom-masking. Buspironeis a partial agonist at the 5-HT1A autoreceptor, nudging serotonergic tone without sedation. Prazosinin PTSD targets α1-noradrenergic overactivity specifically implicated in nightmares.
Two symptom buckets. Psychological/cognitive:excessive worry that's hard to control, feeling keyed up or on edge, trouble concentrating or mind going blank. Physical:restlessness, fatigue, muscle tension, sleep disturbance, irritability.
Excessive anxiety and worry most days for at least 6 months, with at least 3 of those physical symptoms, causing real distress or functional impairment, not explained by a substance or medical condition. Onset is gradual, average age 21, and the course is chronic with exacerbations and remissions, so this is a lifelong-management conversation, not a short course of treatment.
The GAD-7:a 7-item self-report screen, each item scored 0-3 over the past 2 weeks, total range 0-21. 0-4 minimal · 5-9 mild · 10-14 moderate · 15-21 severe.A score of ≥10is the standard cutoff for probable GAD (pooled sensitivity ~0.79, specificity ~0.89) and should prompt a full clinical interview, since the screen alone isn't diagnostic. Use it serially: a drop of 4 points is considered the minimal clinically important difference (MCID), and remission is generally a score of ≤5-7.
SSRIs/SNRIs for GAD have a Cohen's d around 0.33-0.44, statistically "small-to-moderate." Translated to the GAD-7, that's roughly a 2-4 point drop. CBT runs larger (SMD around -0.80, roughly a 7+ point drop, closer to remission) but takes 4-8 weeks to show it versus 2-4 weeks for medication. Neither number means the treatment doesn't work; it means set expectations with the patient about how much and how fast.
Goals:reduce severity, frequency, and duration of symptoms; restore functioning; long-term, aim for minimal-to-no anxiety, no functional impairment, and relapse prevention.
CBT is the most effective psychotherapy and the preferred first-line option alongside or instead of medication when access allows. Add stress management, psychoeducation, and exercise. Advise patients to cut caffeine, nicotine, stimulants, excessive alcohol, and diet pills, since all of them are anxiogenic.
SSRIs (escitalopram, paroxetine, sertraline) or SNRIs (venlafaxine XR, duloxetine) over an 8-12 week trialget response rates of 60-68% and remission around 30%. Venlafaxine, escitalopram, paroxetine, duloxetine, and quetiapine are the ones more likely to get a patient to full remission, but sertraline tends to be the best tolerated, so it's often the pragmatic first pick. Start low for the first week, since a subset of patients get transient jitteriness syndrome(paradoxical worsening of anxiety) before the drug kicks in. All antidepressants carry the black box warning for suicidality in patients under 25.
Fluoxetine, sertraline, or citalopram are the guideline-preferred agents in pregnancy. Avoid paroxetinespecifically, it carries a cardiovascular malformation risk in the fetus. Watch for neonatal jitteriness, myoclonus, and irritability regardless of which SSRI is used near term.
Only 7 benzodiazepines are FDA-approved for GAD, though all of them have anxiolytic activity off that indication. About 65-75% of patients get a marked-to-moderate response, mostly within the first 2 weeks, and they hit the somatic/autonomic symptoms harder than antidepressants do (which are better for the psychic worry itself). Course material puts their effect size around 0.50, medium, and notably rated equal or better than SSRIs across GAD, SAD, and panic disorder for raw effect size, but the tradeoff is dependence risk, so guidelines still push them to second-line, short-term (2-4 weeks) use. Always dose on a fixed schedule, never PRN, when treating an anxiety disorder specifically (PRN dosing reinforces the anxiety-relief association and drives escalation).
Diagnose → 4-6 week SSRI/SNRI trial at adequate dose → assess response. Full response:continue at least 1 year. Partial or no response:re-evaluate for comorbidity (depression, bipolar, another anxiety disorder, insomnia, substance use), then either switch within/across SSRI-SNRI-TCA-NaSSA classes, or augment with a benzodiazepine, buspirone, antihistamine, or an atypical antipsychotic. Comorbid insomnia gets its own branch (non-BZD hypnotic, trazodone, mirtazapine) rather than just piling on more anxiolytic.
Reassess every 2 weeks initially for symptom trend, function, and adverse effects; the HAM-A or Sheehan Disability Scale can supplement the GAD-7. Call it treatment-resistant only after failing 2 antidepressants from different classesat adequate dose and duration, and rule out nonadherence or a missed comorbidity before escalating further.
Psychological:depersonalization (feeling detached from yourself), derealization (feeling detached from your surroundings), fear of dying or losing control. Physical:chest pain, palpitations, tachycardia, dizziness, choking sensation, shortness of breath, sweating, trembling, paresthesias, GI distress. Symptoms peak within 10 minutesand usually resolve within 20-30 minutes, which is a useful thing to tell a patient who thinks they're dying every time it happens.
Recurrent, unexpectedpanic attacks, with at least 4 symptoms during the attack, plus at least 1 month afterward of either persistent worry about another attack or a behavior change because of the attacks (avoiding the gym, avoiding driving, etc). Up to 50% develop agoraphobia, fear of situations where escape would be hard or help unavailable, and can become homebound in severe cases.
A first panic attack presentation overlaps heavily with cardiac, thyroid, and pulmonary emergencies. Diagnosis is one of exclusion after a reasonable medical evaluation, not a diagnosis of convenience in the ED.
Goals:complete resolution of attacks, marked drop in anticipatory anxiety, elimination of phobic avoidance, full functional recovery.
CBT gets 80-90% short-term improvement and 75% at 6 months; interoceptive exposure (deliberately provoking the physical sensations of panic in a controlled way to break the fear association) is the specific technique that sets PD-focused CBT apart from generic anxiety CBT. Cut caffeine, nicotine, alcohol, and stimulants; aerobic exercise helps.
SSRIs or venlafaxine XR are first-line, but benzodiazepines are used most often in practice because they work faster. In older adults and youth, favor the antidepressant route since benzodiazepines carry more disinhibition risk in those groups. SSRIs eliminate attacks in 60-80% of patients by about 4 weeks, though some need 8-12. Start low and titrate slowly, since the same stimulatory jitteriness that hits GAD patients can torpedo adherence here too.
Benzodiazepines are frequently started concurrently with an SSRI for the first 4-6 weeksspecifically to bridge the gap while the antidepressant ramps up, then tapered off once the SSRI is doing the work. Alprazolam and clonazepamare the preferred agents here (response in 1-2 weeks); use the extended-release forms to avoid interdose breakthrough panic. Clonazepam starts at 0.25 mg once or twice daily, titrating by 0.25-0.5 mg every 3 days up to 4 mg/day if needed. Avoid benzodiazepine monotherapy in anyone with current or past depression or substance use disorder.
Course-material effect sizes back this up: benzodiazepines carry an NNT of about 4 for panic disorder specifically, the best number in the chapter, but that has to be weighed against dependence liability with continued use.
If treatment is urgent and there's no substance use history: short course benzo (2-4 weeks) plus SSRI or venlafaxine XR for 12 weeks. If not urgent, or substance use history present: SSRI or venlafaxine XR alone for 12 weeks. No adequate response → switch to another SSRI or venlafaxine. Still no response → switch antidepressant class (imipramine is the TCA option) or add a benzodiazepine, gabapentin, or a second-generation antipsychotic on top. Responders continue therapy 12-24 months before a slow taper.
See patients every 1-2 weeks initially, then every 2 months once stable. The Panic Disorder Severity Scale (goal ≤3, with no or mild agoraphobic avoidance) and Sheehan Disability Scale (goal ≤1 per item) track response. Ramp up visit frequency again during any discontinuation attempt.
Requires exposure (direct, witnessed, or learned about happening to someone close) to actual or threatened death, serious injury, or sexual violence. Symptoms fall into four clusters, and diagnosis needs at least one from intrusion, one from avoidance, two from negative mood/cognition, and two from hyperarousal, all lasting more than 1 month.
| Cluster | Examples |
|---|---|
| Intrusion | Recurrent distressing memories, nightmares, dissociative flashbacks, physiologic reaction to trauma reminders |
| Avoidance | Avoiding thoughts, feelings, people, places, or activities tied to the trauma |
| Negative mood/cognition | Can't recall parts of the trauma, anhedonia, feeling estranged, distorted self-blame, persistent negative mood |
| Hyperarousal | Poor concentration, easily startled, hypervigilance, insomnia, irritability or anger outbursts, self-destructive behavior |
PTSD heavily co-occurs with mood disorders, other anxiety disorders, and substance use, and it raises lifetime suicide attempt risk, so a suicide screen belongs in every visit, not just the first.
Goals:reduce core symptoms and disability, treat comorbidities, restore quality of life.
Trauma-focused CBT (TF-CBT) and EMDR (eye movement desensitization and reprocessing) both beat stress management or group therapy for reducing PTSD symptoms. Current guidance, including the VA's, actually puts trauma-focused psychotherapy ahead of medicationas the preferred initial approach when it's available, a shift from older algorithms that led with drugs. Brief TF-CBT started close to the trauma can help prevent PTSD from developing in the first place.
SSRIs and venlafaxine are first-line.Only sertraline and paroxetine are FDA-approvedfor acute PTSD, and sertraline alone carries the long-term maintenance approval. Start low and titrate slowly toward antidepressant-range doses; the acute phase runs 8-12 weeks. Long-term use (9-12 months) meaningfully cuts relapse. One head-to-head: venlafaxine XR was better at the avoidance/numbing and hyperarousal clusters, while sertraline moved the needle across all clusters more evenly, which is a reasonable tiebreaker if one symptom cluster clearly dominates the picture.
Bupropion is not recommendedfor PTSD. Mirtazapine, amitriptyline, and imipramine are second-line; phenelzine and nefazodone are third-line and mostly of historical interest given tolerability. Second-generation antipsychotics (risperidone, quetiapine) can augment intrusive symptoms in partial responders, but the metabolic tradeoff has to be part of that conversation.
After an adequate 8-12 week SSRI/SNRI trial (4-6 weeks at max tolerated dose), augmentation is chosen by which residual symptom is worst, not generically: sleep-onset trouble→ trazodone, a TCA, mirtazapine, or hydroxyzine. Nightmares/mid-sleep waking→ prazosin(SOR B). Hyperarousal→ clonidine, guanfacine, or propranolol (quetiapine as an alternative). Avoidance or re-experiencing symptoms→ lamotrigine, with aripiprazole, risperidone, or quetiapine as alternatives. Only augment on top of an SSRI, SNRI, or mirtazapine, not on top of a TCA or MAOI. If there's no adequate SSRI/SNRI trial yet, the move is switching agents or extending the trial, not jumping straight to augmentation.
Prazosindeserves its own mention: 1-25 mg/day targeting nightmares and sleep disruption specifically through α1-blockade. It has real support in older algorithms and course material, but recent VA guidelines have actually walked back a routine recommendation for it after larger trials underwhelmed, so treat it as an individualized trial, not an automatic add.
See patients frequently for the first 3 months, then monthly through month 6, then every 2 months through month 12. Responders continue treatment at least 12 months, then taper slowly over 1+ month to limit relapse. The Clinician-Administered PTSD Scale (CAPS) is the gold-standard severity measure. Track suicidal ideation, disability, and adherence at every visit, not just symptom counts.
Doses grouped by class. Where a drug appears in more than one disorder, the dose ranges can differ, so match the row to the indication you're treating.
| Class / Drug | Starting Dose | Target / Usual Range | Used in |
|---|---|---|---|
| SSRIs | |||
| Escitalopram | 5-10 mg/day | 10-20 mg/day | GAD, PD |
| Sertraline | 25-50 mg/day | 50-200 mg/day | GAD, PTSD (FDA-approved) |
| Paroxetine | 10-20 mg/day | 20-50 mg/day (avoid in pregnancy) | GAD, PD, PTSD (FDA-approved) |
| Citalopram | 10 mg/day | 20-40 mg/day (QT risk) | PD |
| Fluoxetine | 5-10 mg/day | 10-40 mg/day | PD, PTSD |
| Fluvoxamine | 25 mg/day | 100-300 mg/day | PD |
| SNRIs | |||
| Venlafaxine XR | 37.5-75 mg/day | 75-225 mg/day | GAD, PD, PTSD |
| Duloxetine | 30-60 mg/day | 60-120 mg/day | GAD |
| Benzodiazepines | |||
| Alprazolam | 0.25 mg TID | 0.75-4 mg/day (GAD); 4-10 mg/day (PD); avoid abrupt stop, interdose rebound | GAD, PD |
| Clonazepam | 0.25 mg QD-BID | 1-4 mg/day | GAD, PD |
| Lorazepam | 0.5-1 mg TID | 0.5-10 mg/day; preferred in older adults | GAD, PD |
| Diazepam | 2-5 mg TID | 2-40 mg/day | GAD, PD |
| Other anxiolytics | |||
| Buspirone | 15 mg/day (divided) | 15-60 mg/day, titrate q2-3 days by 5 mg | GAD (2nd line) |
| Hydroxyzine | 100-200 mg/day | 200-400 mg/day | GAD (2nd line) |
| Pregabalin | 150 mg/day | 150-600 mg/day (renal adjust) | GAD (2nd line) |
| Quetiapine XR | 50 mg/day | 150-300 mg/day (not FDA-approved) | GAD (alternative) |
| TCAs / MAOIs / augmentation | |||
| Imipramine | 10-50 mg/day | 75-250 mg/day | GAD (alt), PD (2nd line) |
| Phenelzine | 15 mg/day | 45-90 mg/day | PD, PTSD (3rd line) |
| Prazosin | 1 mg at bedtime | 1-25 mg/day | PTSD (nightmares) |
Benzodiazepines show up as an option across all three disorders, so the pharmacology and the discontinuation mechanics are worth knowing cold rather than re-deriving each time.
Diazepam and clorazepateare highly lipophilic: rapid absorption, fast CNS penetration, fast anxiolytic effect, but a shorter duration after a single dose than the half-life would predict, because they redistribute quickly to peripheral fat. Lorazepam and oxazepamare less lipophilic, slower onset, longer duration; not the choice for immediate relief, but steadier.
Several agents (diazepam included) get converted to desmethyldiazepam, a long-half-life active metabolite that accumulates with repeated dosing, especially dangerous in older adults or hepatic impairment. Lorazepam, oxazepam, and temazepam lack active metabolites, have a more predictable, shorter path to steady state, and are the preferred choices in older adults and liver disease.
Avoid IM diazepam and IM chlordiazepoxide, absorption is erratic by that route. IM lorazepam, by contrast, is rapid and reliable. Never combine any benzodiazepine with alcohol or another CNS depressant; that combination can be fatal via respiratory depression.
Interactions:CYP3A4 inhibitors (ketoconazole, ritonavir, nefazodone) raise alprazolam and diazepam levels; CYP3A4 inducers (carbamazepine, St. John's wort) lower them. Fluoxetine, fluvoxamine, and omeprazole inhibit CYP2C19 and can alter diazepam/clonazepam metabolism through that pathway.
Dependence can start developing in as little as 2-4 weeksof regular use per the Ashton Manual, though other sources put the more reliable threshold closer to ≥8 weeks. Either way, "just a few weeks" is not automatically safe, and risk scales with dose, potency, and duration.
Three distinct things can happen after stopping abruptly, and mixing them up is a common error: reboundis the original symptoms returning immediately but more intensely than baseline; recurrence/relapseis the original symptoms returning at the same intensity as before treatment; withdrawalis genuinely new symptoms plus worsening of old ones (anxiety, insomnia, agitation, tremor, nausea, diaphoresis, hyperreflexia, tinnitus, and in severe cases delusions, hallucinations, or seizures).
Onset of withdrawal: 24-48 hoursafter stopping a short-half-life agent, 3-8 daysafter stopping a long-half-life agent, because the long-acting drug is essentially auto-tapering itself via its own elimination curve.
CIWA-Ar(the same scale used for alcohol withdrawal) is used clinically for benzodiazepine withdrawal too, since the cross-tolerant GABA mechanism overlaps. A dedicated CIWA-Bexists (22 symptoms) but is mostly a research tool, not routine practice. The Severity of Dependence Scale (SDS), a 5-item self-report, uses a score above 6 as the threshold flagging problematic use, with reported sensitivity 97.9% and specificity 94.2%.
Real-world signal:a 2023 comparative-effectiveness study (not an RCT, so treat as hypothesis-generating) found abrupt discontinuation in patients on stable long-term benzodiazepine therapy was associated with increased nonfatal overdose, suicide attempts/ideation, ED visits, and even mortality. The takeaway isn't "never stop a benzodiazepine," it's "don't stop one abruptly without a plan."
General shape from the guideline algorithm: reduce by 25% per weekuntil you reach 50% of the original dose, then slow to one-eighth reductions every 4-7 days. Total taper length scales with how long the patient has been on it: 2-3 weeksif therapy was under 8 weeks, 4-8 weeksif it's been about 6 months, and 2-4 monthsfor longer-term use. Course material frames it slightly differently but consistently: cut by 5-25% every 2-4 weeks, stretching to every 6-8 weeks as you go, since the last 10-20% is reliably the hardest part, especially with high-potency agents.
Adjuncts that help:pregabalin can blunt withdrawal symptoms during the taper. Prefer switching a short-half-life, high-potency agent (alprazolam) to a longer-acting one (diazepam, clonazepam) before starting the taper, since the smoother pharmacokinetic decline is gentler on the receptor.
In older adults specifically:if a benzodiazepine truly can't be avoided, weigh active metabolites, onset, duration, and lipophilicity together. Oxazepam and temazepam are generally considered the safest picks in this population precisely because they lack active metabolites and have a slower, gentler onset.
5-HT1A partial agonist. No antiseizure, muscle relaxant, sedative-hypnotic, or dependence-producing activity, which makes it the go-to when a patient has a substance use history that rules out benzodiazepines. The tradeoff is a real one: onset takes 2+ weeks, max benefit at 4-6 weeks, no PRN utility at all, and it tends to work less well in patients who've already been exposed to benzodiazepines (possibly because they're anchored to what fast relief feels like).
Mean half-life ~2.5 hours, so it's dosed 2-3 times daily, titrated in 5 mg increments every 2-3 days. Watch interactions: verapamil, itraconazole, and fluvoxamine raise levels via CYP3A4 inhibition; rifampin can cut levels roughly 10-fold. It can also raise blood pressure in a patient on an MAOI.
| Parameter | When | Watching for |
|---|---|---|
| GAD-7 / HAM-A / CAPS | Baseline, then every visit; every 2 weeks initially | Symptom trend, ≥4 pt drop on GAD-7 = clinically meaningful |
| Suicidality | Weekly for first few weeks if age <25 or depression comorbid, then every visit | All antidepressants carry the black-box warning |
| Jitteriness / activation | First 1-2 weeks of any SSRI/SNRI | Transient worsening of anxiety before benefit kicks in |
| Blood pressure | Baseline and with SNRI titration | SNRIs (especially venlafaxine) can raise BP dose-dependently |
| ECG / QT | Before citalopram if cardiac risk factors present | QT prolongation |
| Sodium (BMP) | Baseline and periodically, especially in adults >65 | SSRI-associated hyponatremia (SIADH) |
| Weight / metabolic panel | Periodically on paroxetine or quetiapine | Weight gain, metabolic syndrome |
| Sedation, falls, cognition | Every visit on a benzodiazepine, especially older adults | Psychomotor impairment, anterograde amnesia, fall risk |
| Refill pattern / PDMP | Ongoing for anyone on a benzodiazepine | Early refills, dose escalation, use disorder |
| Discontinuation visits | Increase frequency during any taper | Rebound, relapse, or withdrawal, and tell them apart |