Skip to content

Course, relapse and long-term outcome

TL;DR — GAD's defining temporal feature is chronicity, and it is the feature its evidence base least addresses. In the Harvard/Brown Anxiety Research Program — the reference naturalistic cohort — the five-year probability of remission from GAD was 0.38, and reduced remission was predicted by low life satisfaction, poor spousal/family relationships, a cluster B or C personality disorder and low global functioning (Yonkers 2000, PMID 10974960). Over 12 years the same cohort showed an "insidious" chronic course with low recovery and high recurrence, worsened by comorbid major depression, alcohol and substance use disorders, and by GAD comorbid with panic-with-agoraphobia (Bruce 2005, PMID 15930067). Personality disorder cut the likelihood of GAD remission by 30% (avoidant PD 34%, dependent PD 14%), independently of major depression (Massion 2002, PMID 11982447). Against this, the prospective Zurich community cohort reports a much less grim picture: of 75 DSM-III GAD cases followed to their last assessment, 52% were symptom-free, 29% subthreshold and 16% still met criteria, with only 7% persistent (Angst 2009, PMID 18575915) — clinic cohorts and community cohorts describe different diseases. Treatment prevents relapse convincingly (escitalopram 19% vs 56% placebo; Allgulander 2006, PMID 16316482), but the psychological long-term evidence is weak: CBT's effect in GAD at ≥12 months is Hedges' g=0.22 across 10 studies, the lowest of the anxiety disorders (van Dis 2020, PMID 31758858). GAD also carries measurable somatic prognostic weight: a 62% higher rate of cardiovascular events in stable coronary heart disease after full adjustment (Martens 2010, PMID 20603456).

Remission and recovery rates

Cohort Design Follow-up Remission / recovery
HARP (Yonkers 2000, PMID 10974960) 167 GAD patients, clinic-recruited, assessed every 6–12 months 5 years Remission probability 0.38
HARP (Bruce 2005, PMID 15930067) Prospective, naturalistic, multicentre 12 years "Overall chronic course for the majority"; social phobia lowest recovery; recovered patients had a high probability of recurrence
HARP (Yonkers 2003, PMID 12768651) 558 patients, 6–12-month intervals 8 years Cumulative remission equivalent in men and women; remissions concentrated in the first 2 years; women with GAD continued remitting late and had fewer overall remission events by 8 years (difference not statistically significant)
Zurich cohort (Angst 2009, PMID 18575915) Community sample, 591 young adults, six interviews 1979–1999 20 years Of 75 DSM-III GAD cases with follow-up: 52% symptom-free, 29% subthreshold/symptoms, 16% re-diagnosed GAD, 7% persistent; annual incidence rose considerably between ages 20 and 40; mean symptom onset 15.6 years, 75% before age 20
NESDA (Bokma 2022, PMID 32524918) 887 patients with anxiety disorders, naturalistic cohort, 569 baseline predictors across five domains 2 years 54.6% recovered; machine-learning AUC 0.67 for recovery from anxiety disorders (sens 62.0%, spec 62.8%); clinical domain alone performed comparably to all domains combined

The gap between HARP's 0.38 five-year remission and the Zurich cohort's 52% symptom-free is not a contradiction to be averaged away. HARP recruited treatment-seeking clinic patients with current or past anxiety disorders; Zurich followed a community birth cohort. Which figure is "the" prognosis of GAD depends on where the patient is standing — the same sampling asymmetry that produces the pure-versus-comorbid dispute (the diagnostic boundary).

What predicts a worse course

Predictor Effect Source
Any personality disorder 30% lower likelihood of GAD remission over 5 years; not explained by major depression Massion 2002, PMID 11982447
Avoidant personality disorder 34% lower Massion 2002, PMID 11982447
Dependent personality disorder 14% lower Massion 2002, PMID 11982447
Comorbid major depression Lowers recovery, raises recurrence over 12 years Bruce 2005, PMID 15930067
Comorbid alcohol/substance use disorder Same direction Bruce 2005, PMID 15930067
GAD comorbid with panic disorder with agoraphobia Worsens overall clinical course Bruce 2005, PMID 15930067
Low life satisfaction; poor spousal/family relationships; low GAS score Reduced full or partial remission Yonkers 2000, PMID 10974960
Cluster B or C personality disorder Reduced remission Yonkers 2000, PMID 10974960
Longer episode duration (≥12 months) Most severe, most persistently symptomatic, most impaired subgroup across 17 countries Lee 2009, PMID 19091158
Lower-income country Later onset and more persistent clinical course Ruscio 2017, PMID 28297020

Across these cohorts, comorbidity, personality and social context repeatedly predicted course alongside clinical features. In one trans-anxiety prediction study, adding biological, lifestyle, sociodemographic and psychological domains did not improve the AUC beyond 0.67 compared with the clinical domain alone (Bokma 2022, PMID 32524918); that single model should not be treated as a biological ceiling.

Relapse after treatment stops

Trial Drug Continuation Relapse on active Relapse on placebo
Allgulander 2006 (PMID 16316482) Escitalopram 20 mg 24–76 weeks 19% 56% (risk 4.04× higher)
Davidson 2008 (PMID 18559291) Duloxetine 60–120 mg 26 weeks 13.7% 41.8%
Stein 2012 (PMID 22901350) Agomelatine 25–50 mg 26 weeks 19.5% 30.7%

Psychological treatment's long-term record is much less well characterised:

  • CBT at ≥12 months in GAD: Hedges' g=0.22 (k=10) — significant but the smallest durable effect among anxiety-related disorders in that analysis (social anxiety 0.42, PTSD 0.84); at shorter follow-ups the GAD range is 0.07–0.40 (van Dis 2020, PMID 31758858).
  • Relapse is barely reported. Relapse rates of 0–14% at 3–12 months appeared in only 6 of 69 RCTs, predominantly panic-disorder trials (van Dis 2020, PMID 31758858).
  • Withdrawal can complicate relapse ascertainment. Across antidepressant studies, discontinuation increased symptom scores modestly at week 1, with dizziness, nausea, vertigo and nervousness prominent; the review was dominated by depression rather than GAD trials (Kalfas 2025, PMID 40632531). GAD continuation studies therefore need explicit withdrawal adjudication.
  • The 2007 Cochrane review found no study at all conducting longer-term assessment of CBT against treatment as usual or waitlist in GAD (Hunot 2007, PMID 17253466).
  • The exception. Metacognitive therapy versus CBT at 9 years: recovery 57% vs 38%; sustained recovery from post-treatment 43% vs 13%; GAD re-diagnosis 9.5% vs 23.1% (Solem 2021, PMID 34520637). Small sample (39 of 60 original patients), single centre.
  • Guided internet CBT at 1 and 3 years: results improved or were maintained (Paxling 2011, PMID 21770848).

Mortality

The largest quantification of excess mortality in anxiety and stress-related disorders was published in 2026: 165 studies, 7,395,722 people with any anxiety or stress-related disorder and 135,059,023 controls, from 27 countries across all continents (Wagner 2026, PMID 42136520).

Outcome GAD-specific risk ratio Comparison
All-cause mortality RR 1.48 (95% CI 1.23–1.78, p<0.001; n=9 studies) Any anxiety/stress-related disorder RR 1.54 (1.14–2.08); PTSD and other stress-related disorders 1.39 (1.15–1.67); not elevated for panic disorder, phobias or mixed anxiety/stress disorders
Suicide mortality RR 1.93 (1.17–3.17, p<0.01; n=3) Any anxiety/stress-related disorder 2.88 (2.13–3.89); panic disorder 3.58 (1.39–9.25); PTSD and other stress-related 3.13 (1.85–5.28); mixed 2.77 (1.89–4.07)
Natural-cause mortality RR 1.55 (1.19–2.02, p=0.001; n=5) Any anxiety/stress-related 1.25 (1.09–1.44); mixed 1.26 (1.02–1.56); PTSD and other stress-related 1.17 (1.03–1.33); not elevated in panic disorder
Suicide attempt Elevated across all anxiety/stress-related disorders, from RR 6.33 (4.08–9.82) in panic disorder to 2.74 (1.72–4.35) in phobias —

Three features of this table matter. First, GAD estimates were elevated for both natural-cause mortality (RR 1.55) and all-cause mortality (1.48), but overlapping outcome categories and confidence intervals do not justify attributing the excess to physical illness or comparing contributions directly. Second, among the individual disorder groupings, GAD and the PTSD/stress-related group showed significantly elevated all-cause mortality in this synthesis. Third, the GAD-specific estimates rest on 3–9 studies each, so the intervals are wide and the point estimates provisional.

Cognition over time

Question Finding Source
Is executive functioning impaired in GAD? Across 32 studies and 13,084 participants, GAD was associated with poorer cognitive flexibility and working memory but not inhibitory control; effects were similar in youths and young/middle-aged adults but absent in older adults; stronger in clinical than non-clinical samples and stronger for self-report than performance-based indices. GAD was linked to slower reaction times and poorer accuracy on working-memory and flexibility tasks — which challenges attentional control theory's prediction that anxiety degrades efficiency rather than accuracy Nguyen 2025, PMID 40513710
Does GAD cause cognitive decline? Across 13 longitudinal studies and RCTs, no consistent evidence of sustained cognitive decline: no significant differences on dementia-rating-scale memory or attention subscores or total error scores. CBT was associated with greater improvement in clinical symptoms than non-cognitive interventions (MD −4.39, p=0.0004) Li 2026, PMID 42191132
Late-life GAD and cognition Cognitive impairment documented in late-life GAD in a comparison study Mantella 2007, PMID 17426260

The pattern — measurable cross-sectional executive deficits, no demonstrated longitudinal decline — is the same shape as in the imaging literature: reliable group differences without a demonstrated trajectory (mechanism and models).

Somatic outcomes

  • Cardiovascular events. In 1,015 outpatients with stable coronary heart disease followed a mean 5.6 years (Heart and Soul Study), age-adjusted annual cardiovascular event rate was 9.6% with GAD versus 6.6% without (p=0.03); after adjustment for demographics, comorbid conditions including major depressive disorder, cardiac disease severity and medication, GAD remained associated with a 62% higher event rate (HR 1.62, 95% CI 1.11–2.37, p=0.01), and adding behavioural and biological mediators did not attenuate it (HR 1.74, 1.13–2.67) (Martens 2010, PMID 20603456).
  • Treatment and cardiovascular risk. In a propensity-matched real-world cohort of 54,526 SSRI users and 54,526 non-users with both coronary artery disease and GAD, SSRI use was associated with lower one-year MACE (HR 0.77, 95% CI 0.74–0.81, p<0.001), consistent across subgroups and landmark analyses, with null negative-control analyses (Wu 2025, PMID 40466339). Observational; residual confounding by indication cannot be excluded.
  • Suicidality. Threshold and subthreshold GAD are independently associated with 12-month suicidal ideation after adjusting for sociodemographics and psychiatric comorbidity; 32.0% of men and 21.2% of women with threshold GAD reported past-year ideation (Gilmour 2016, PMID 27849314). See red flags and safety concerns.

The structural mismatch

Feature of the disorder Feature of its evidence base
Median course measured in years to decades (Bruce 2005, PMID 15930067) Longest double-blind acute phase in the Cochrane antidepressant review: 28 weeks (Kopcalic 2025, PMID 39880377)
Recurrence after recovery is high (Bruce 2005, PMID 15930067) Relapse reported in 6 of 69 psychotherapy RCTs (van Dis 2020, PMID 31758858)
Comorbidity and personality drive prognosis (Massion 2002, PMID 11982447) Trials exclude serious comorbidity by design (Kopcalic 2025, PMID 39880377)
Onset often in adolescence, incidence rises to age 40 (Angst 2009, PMID 18575915) Trials enrol adults with a mean age around 42 (Papola 2024, PMID 37851421)

This mismatch is the substance of open question OQ-4 (OPEN-QUESTIONS.md).

Open questions

  • Which prognosis is right for a given patient — HARP's 0.38 five-year remission (Yonkers 2000, PMID 10974960) or Zurich's 52% symptom-free (Angst 2009, PMID 18575915)? No study has calibrated clinic against community prognosis in the same population.
  • Do the drug relapse-prevention results (Allgulander 2006, PMID 16316482) support indefinite treatment, and for whom? A PubMed search rerun on 2026-09-02 located no GAD trial comparing fixed-duration with indefinite continuation under a specified taper; this remains a dated evidence gap.
  • Why is CBT's durability in GAD so much weaker than in PTSD or social anxiety (van Dis 2020, PMID 31758858)?
  • Is the metacognitive-therapy durability advantage at 9 years (Solem 2021, PMID 34520637) real, and if so does it come from the treatment or from the sample?
  • Does treating GAD change cardiovascular outcomes causally? The observational signal is strong and consistent (Martens 2010, PMID 20603456; Wu 2025, PMID 40466339); no RCT has tested it.
  • What drives GAD's excess natural-cause mortality (RR 1.55; Wagner 2026, PMID 42136520)? Cardiovascular events are one documented pathway (Martens 2010, PMID 20603456), but the meta-analysis rests on five GAD-specific studies and no mediation analysis exists.
  • Can course be predicted at the individual level? AUC 0.67 with 569 candidate predictors in one trans-anxiety cohort shows limited discrimination and needs external replication and improvement (Bokma 2022, PMID 32524918).

References

  1. Yonkers KA, et al. Factors predicting the clinical course of generalised anxiety disorder. Br J Psychiatry. 2000;176:544-9. PMID 10974960
  2. Bruce SE, et al. Influence of psychiatric comorbidity on recovery and recurrence in generalized anxiety disorder, social phobia, and panic disorder: a 12-year prospective study. Am J Psychiatry. 2005;162:1179-87. PMID 15930067
  3. Yonkers KA, et al. Chronicity, relapse, and illness-course of panic disorder, social phobia, and generalized anxiety disorder: findings in men and women from 8 years of follow-up. Depress Anxiety. 2003;17:173-9. PMID 12768651
  4. Massion AO, et al. Personality disorders and time to remission in generalized anxiety disorder, social phobia, and panic disorder. Arch Gen Psychiatry. 2002;59:434-40. PMID 11982447
  5. Angst J, et al. The generalized anxiety spectrum: prevalence, onset, course and outcome. Eur Arch Psychiatry Clin Neurosci. 2009;259:37-45. PMID 18575915
  6. Bokma WA, et al. Predicting the naturalistic course in anxiety disorders using clinical and biological markers: a machine learning approach. Psychol Med. 2022;52:57-67. PMID 32524918
  7. Lee S, et al. Implications of modifying the duration requirement of generalized anxiety disorder in developed and developing countries. Psychol Med. 2009;39:1163-76. PMID 19091158
  8. Ruscio AM, et al. Cross-sectional Comparison of the Epidemiology of DSM-5 Generalized Anxiety Disorder Across the Globe. JAMA Psychiatry. 2017;74:465-475. PMID 28297020
  9. Allgulander C, et al. Prevention of relapse in generalized anxiety disorder by escitalopram treatment. Int J Neuropsychopharmacol. 2006;9:495-505. PMID 16316482
  10. Davidson JR, et al. Duloxetine treatment for relapse prevention in adults with generalized anxiety disorder: a double-blind placebo-controlled trial. Eur Neuropsychopharmacol. 2008;18:673-81. PMID 18559291
  11. Stein DJ, et al. Agomelatine prevents relapse in generalized anxiety disorder. J Clin Psychiatry. 2012;73:1002-8. PMID 22901350
  12. van Dis EAM, et al. Long-term Outcomes of Cognitive Behavioral Therapy for Anxiety-Related Disorders: A Systematic Review and Meta-analysis. JAMA Psychiatry. 2020;77:265-273. PMID 31758858
  13. Hunot V, et al. Psychological therapies for generalised anxiety disorder. Cochrane Database Syst Rev. 2007;2007:CD001848. PMID 17253466
  14. Solem S, et al. Metacognitive therapy versus cognitive-behavioral therapy in adults with generalized anxiety disorder: A 9-year follow-up study. Brain Behav. 2021;11:e2358. PMID 34520637
  15. Paxling B, et al. Guided internet-delivered cognitive behavior therapy for generalized anxiety disorder: a randomized controlled trial. Cogn Behav Ther. 2011;40:159-73. PMID 21770848
  16. Martens EJ, et al. Scared to death? Generalized anxiety disorder and cardiovascular events in patients with stable coronary heart disease: The Heart and Soul Study. Arch Gen Psychiatry. 2010;67:750-8. PMID 20603456
  17. Wu JY, et al. Association between SSRI use and cardiovascular outcomes in patients with coronary artery disease and generalized anxiety disorder: A real-world cohort study. Atherosclerosis. 2025;407:120390. PMID 40466339
  18. Gilmour H. Threshold and subthreshold Generalized Anxiety Disorder (GAD) and suicide ideation. Health Rep. 2016;27:13-21. PMID 27849314
  19. Kopcalic K, et al. Antidepressants versus placebo for generalised anxiety disorder (GAD). Cochrane Database Syst Rev. 2025;1:CD012942. PMID 39880377
  20. Papola D, et al. Psychotherapies for Generalized Anxiety Disorder in Adults: A Systematic Review and Network Meta-Analysis of Randomized Clinical Trials. JAMA Psychiatry. 2024;81:250-259. PMID 37851421
  21. Wagner E, et al. Risk of all-cause and cause-specific mortality, and suicide attempt in people with anxiety and stress-related disorders: a systematic review, meta-analysis and meta-regression analysis of 165 studies. World Psychiatry. 2026;25:307-320. PMID 42136520
  22. Nguyen L, et al. Executive functioning in individuals with generalized anxiety disorder: A systematic review and meta-analysis. J Affect Disord. 2025;389:119683. PMID 40513710
  23. Li J, et al. Longitudinal Association Between Generalized Anxiety Disorder and Cognitive Change: A Meta-Analysis. Psychiatry Investig. 2026;23:588-596. PMID 42191132
  24. Mantella RC, et al. Cognitive impairment in late-life generalized anxiety disorder. Am J Geriatr Psychiatry. 2007;15:673-9. PMID 17426260
  25. Kalfas M, et al. Incidence and Nature of Antidepressant Discontinuation Symptoms: A Systematic Review and Meta-Analysis. JAMA Psychiatry. 2025;82:896-904. PMID 40632531