Skip to content

Treatment-resistant GAD

TL;DR — About half of patients treated for GAD do not respond adequately to first-line treatment, and treatment-resistant GAD is usually defined as failure of at least one adequate antidepressant trial at adequate dose and duration (Ansara 2020, PMID 33224690; Generoso 2017, PMID 27643884). The evidence base is thin and old: the 2011 systematic review of refractory GAD found four placebo-controlled augmentation trials (pregabalin, olanzapine, quetiapine, risperidone) and four single-arm studies, with only the pregabalin trial (n=356, 8 weeks) adequately powered; pregabalin and risperidone augmentation reduced HAM-A significantly, olanzapine increased responder proportion, and quetiapine augmentation did not significantly reduce mean HAM-A (Samuel 2011, PMID 21088608). Fifteen years later the picture is broader but not deeper: an integrative systematic review across treatment-resistant anxiety disorders found 26 RCTs and 36 open-label studies, "mostly small sample sizes and several methodological limitations", and for TR-GAD specifically supports augmentation with quetiapine, risperidone, olanzapine or pregabalin (Schiele 2026, PMID 40946318). Neurostimulation is the most active frontier and the least trustworthy numbers: rTMS SMD −1.857 (95% CI −2.219 to −1.494) across 6 studies and 152 patients (Parikh 2022, PMID 34791241), and a network meta-analysis of 8 trials/405 participants reporting high-frequency rTMS response OR 291.40 (95% CI 13.08–6490.21) — an interval that is itself the finding (Duan 2025, PMID 40203547). The most important practical result is negative in an encouraging direction: the reference network meta-analysis concluded that failure of one drug is not a reason to abandon pharmacotherapy (Slee 2019, PMID 30712879).

Definition, and how unstable it is

Source Definition used
Ansara 2020 (PMID 33224690) Failure to respond to ≥1 trial of antidepressant therapy at adequate dose and duration
Samuel 2011 (PMID 21088608) "Failed to respond adequately to at least one earlier treatment for GAD"
Schiele 2026 (PMID 40946318) Reviewed across TR-GAD, TR-panic and TR-social anxiety without a single harmonised operational threshold
Gould 2021 (PMID 34542399) Treatment-resistant late-life GAD, operationalised for a feasibility trial in adults ≥65
Domschke 2024 (PMID 38214637) Trans-anxiety Delphi consensus: 36 international experts produced 14 recommendations covering pharmacological and psychotherapeutic resistance and a potential staging model

There is now a trans-anxiety consensus definition and proposed staging framework (Domschke 2024, PMID 38214637). It is not a GAD-only validated instrument, and older GAD trials predate it and use heterogeneous thresholds. Historical studies therefore still do not necessarily enrol comparable populations.

The magnitude of the problem is agreed even where its definition is not: as many as 50% of GAD patients have inadequate response to available agents (Generoso 2017, PMID 27643884), and CBT yields significant improvement in approximately 50% of patients (Nordahl 2018, PMID 30294448).

What a definition would need to contain

The 2024 Delphi consensus supplies a cross-disorder framework, but GAD-specific validation and consistent adoption remain outstanding. Any operational GAD definition still needs to specify at least:

Element Why it is needed Evidence
Which outcome defines non-response Response (≥50% HAM-A reduction) and remission (HAM-A ≤7) produce different drug rankings from the same trials — agomelatine leads on remission but not on mean change (Kong 2020, PMID 33343351; Slee 2019, PMID 30712879) —
Adequate dose Higher SSRI doses (but not SNRI doses) give greater benefit in pooled anxiety-disorder data, while escitalopram 20 mg was not superior to 10 mg in a GAD fixed-dose trial Jakubovski 2019, PMID 30479005; Baldwin 2006, PMID 16946363
Adequate duration SNRI improvement is logarithmic (front-loaded), SSRI improvement linear across 12 weeks — so the same "8-week trial" means different things by class Jakubovski 2019, PMID 30479005
How many failures, and of what Current definitions range from ≥1 antidepressant trial upward and do not specify whether a psychotherapy failure counts Ansara 2020, PMID 33224690; Schiele 2026, PMID 40946318
Whether prior benzodiazepine exposure is an exclusion or a stratifier Recent (<1 month) prior benzodiazepine use predicted higher attrition, more adverse events and the smallest drug–placebo difference on the next agent DeMartinis 2000, PMID 10732655
Comorbidity status Personality disorder cuts remission likelihood by 30%, independent of depression Massion 2002, PMID 11982447

Until these are applied consistently, "treatment-resistant GAD" still names heterogeneous historical populations, limiting pooled inference despite the newer consensus (Domschke 2024, PMID 38214637; Samuel 2011, PMID 21088608).

Pharmacological augmentation

Strategy Best evidence Result
Pregabalin augmentation The single largest refractory-GAD trial: n=356, 8 weeks, placebo-controlled Significant HAM-A reduction (Samuel 2011, PMID 21088608); supported as augmentation in TR-GAD by RCT evidence (Schiele 2026, PMID 40946318)
Risperidone augmentation Placebo-controlled trial in the refractory review Significant HAM-A reduction (Samuel 2011, PMID 21088608)
Olanzapine augmentation Placebo-controlled Significantly higher responder proportion by HAM-A (Samuel 2011, PMID 21088608)
Quetiapine augmentation Placebo-controlled trial in partial/non-responders (Altamura 2011, PMID 21403524); and in the refractory review No significant mean HAM-A reduction in the pooled refractory review (Samuel 2011, PMID 21088608), yet supported by RCT evidence in the later integrative review (Schiele 2026, PMID 40946318) — an unresolved discrepancy
Aripiprazole augmentation Open-label in refractory GAD and panic (Hoge 2008, PMID 18567977); SSRI augmentation in depression and anxiety patients (Worthington 2005, PMID 15602109) Preliminary only (Schiele 2026, PMID 40946318)
Atypical antipsychotics generally Adjunctive use reviewed (Lorenz 2010, PMID 20795849); umbrella review of antipsychotics in anxiety disorders "Lack of high-quality studies … outside of the use of quetiapine in GAD"; 24 of 25 reviews rated low quality on AMSTAR-2 (Garakani 2024, PMID 38382649)
Ketamine Open-label evidence in treatment-resistant anxiety disorders Preliminary (Schiele 2026, PMID 40946318). Esketamine data exist in treatment-resistant depression with comorbid anxiety, not in GAD (Daly 2021, PMID 34293233)
Riluzole Systematic review of riluzole in disorders with anxiety or fear as primary symptoms Small and inconclusive (Kawashima 2023, PMID 37463744; Pittenger 2008, PMID 18698875)
Other open-label agents nefazodone, reboxetine, buspirone, ziprasidone, divalproex sodium, levetiracetam, zonisamide, flumazenil, cannabidiol, acamprosate All preliminary, open-label only (Schiele 2026, PMID 40946318)

Two structural cautions. First, the augmentation literature is dominated by second-generation antipsychotics, whose harms (weight gain, sedation, extrapyramidal effects) are established even where efficacy is not (Depping 2010, PMID 21154392). Second, the largest and most robust augmentation signal — pregabalin — comes from a single 8-week trial (Samuel 2011, PMID 21088608) in a disorder whose defining feature is chronicity (course, relapse and long-term outcome).

Sequencing and switching

  • Do not abandon pharmacotherapy after one failure. "The failure of initial pharmacological therapy might not be a reason to abandon a pharmacological treatment strategy" — the explicit conclusion of the 89-trial network (Slee 2019, PMID 30712879).
  • Psychotherapy after pharmacological failure works. CBT was effective in several RCTs in pharmacologically treatment-resistant anxiety disorders; after CBT non-response, first evidence supports ACT and mindfulness-based cognitive therapy (Schiele 2026, PMID 40946318).
  • Prior benzodiazepine exposure changes the next drug's performance. Patients who had stopped a benzodiazepine within the previous month had higher attrition, more adverse events and the smallest buspirone–placebo difference; those with no or remote prior use responded comparably to a benzodiazepine (DeMartinis 2000, PMID 10732655).
  • Older adults have a published sequencing algorithm. SSRIs first (sertraline or escitalopram preferred; buspirone if sexual side effects must be avoided) → different SSRI or venlafaxine/duloxetine → pregabalin/gabapentin, lavender oil or agomelatine → quetiapine; caution with benzodiazepines and hydroxyzine; low priority to vilazodone, vortioxetine, mirtazapine and cannabidiol (Chen 2025, PMID 39352792). This is the most explicit stepwise algorithm located for any GAD population and it is age-specific (special populations).
  • Sequential treatment as a framework. The sequential model — pharmacotherapy for acute response followed by psychotherapy for residual symptoms — was proposed for mood and anxiety disorders and remains under-tested in GAD specifically (Fava 2005, PMID 16420076).

Neurostimulation

Modality Evidence Effect
rTMS 6 studies, 152 patients (97 active, 55 sham), heterogeneity modest (I²=13.32) SMD −1.857 (95% CI −2.219 to −1.494, p<0.001), prediction interval −2.55 to −1.16 (Parikh 2022, PMID 34791241)
High-frequency rTMS (network) 8 RCTs, 20 arms, 405 participants; right DLPFC the commonest target Response OR 291.40 (95% CI 13.08–6490.21); remission OR 182.14 (8.72–3805.76) vs other active therapies (Duan 2025, PMID 40203547)
Continuous theta-burst stimulation Same network Post-treatment severity SMD −2.56 (−3.16 to −1.96) (Duan 2025, PMID 40203547)
Acceptability Same network No significant differences between strategies and sham (Duan 2025, PMID 40203547)
rTMS in treatment-resistant anxiety generally Integrative review "Only inconclusive support" (Schiele 2026, PMID 40946318)

The effect sizes here are implausibly large for a psychiatric intervention, and the confidence intervals say so: an odds ratio whose interval spans 13 to 6,490 is not an estimate, it is a signal that the network is sparse. The rTMS meta-analysis authors are explicit that "rigorously designed, randomized controlled trials of rTMS for GAD … are urgently needed" (Parikh 2022, PMID 34791241). This knowledge base treats neurostimulation in GAD as promising and unquantified.

The neurostimulation evidence in full

Four independent syntheses now exist, and they agree on direction while disagreeing on magnitude by an order of magnitude.

Source Scope Effect
Hyde 2022 (PMID 35365806) 208 RCTs across mental disorders, TMS and tDCS vs sham, uniform criteria TMS for GAD symptoms SMD −1.8 (95% CI −2.6 to −1.0), without significant heterogeneity — larger than TMS in OCD (−0.66) or unipolar depression (−0.60), both of which did show significant heterogeneity
Parikh 2022 (PMID 34791241) 6 studies, 152 GAD patients (97 active, 55 sham), I²=13.32 rTMS SMD −1.857 (−2.219 to −1.494), prediction interval −2.55 to −1.16
Qi 2024 (PMID 39208534) 7 RCTs, GAD-specific, PROSPERO CRD42023466285 Significant differences in HARS change, study-defined response and remission favouring stimulation; cTBS and rTMS outperformed tDCS, which showed no significant therapeutic effect; intervention groups had significantly more headaches
Duan 2025 (PMID 40203547) Network meta-analysis, 8 RCTs, 20 arms, 405 participants; right DLPFC the commonest target High-frequency rTMS response OR 291.40 (13.08–6490.21) and remission OR 182.14 (8.72–3805.76) vs other active therapies; cTBS post-treatment severity SMD −2.56 (−3.16 to −1.96); no acceptability differences vs sham

Three things are worth separating. (i) rTMS and cTBS estimates favour active stimulation across the syntheses, whereas tDCS did not separate from sham (Qi 2024, PMID 39208534). (ii) The very large estimates are imprecise and arise from sparse networks; they cannot be validly compared with drug or psychotherapy SMDs from different trial structures. (iii) The total randomised GAD evidence across all four reviews is a few hundred patients. Qi 2024 found significantly more headaches with active stimulation.

Novel mechanisms being tested in resistant or severe GAD

  • MM120 (lysergide D-tartrate). Phase 2b, 22 US sites, 198 randomised adults with HAM-A ≥20, single dose, blinded central raters: significant dose–response at 100 µg (LS mean difference −5.0, 95% CI −9.6 to −0.4) and 200 µg (−6.0, −9.8 to −2.0) versus placebo at week 4; 25 µg and 50 µg not significant. Visual perceptual changes in 46.2% of dosed participants. Stated MCID 2.5 HAM-A points (Robison 2025, PMID 40906494). This is the largest single-dose drug–placebo difference in the modern GAD literature and it comes from a psychedelic; see clinical trials landscape.
  • Investigational pipeline. Reviewed in Schanzer 2019 (PMID 31607187) and Fagan 2023 (PMID 37183813).
  • Acceptance and commitment therapy for treatment-resistant late-life GAD. The FACTOID programme developed and feasibility-tested an ACT intervention for adults ≥65 with treatment-resistant GAD; co-primary outcomes were acceptability and feasibility, not efficacy (Gould 2021, PMID 34542399; Gould 2021, PMID 33852722).

Open questions

  • Will the 2024 trans-anxiety Delphi definition and proposed staging model be validated and adopted in GAD trials? It postdates most of the heterogeneous evidence base (Domschke 2024, PMID 38214637; Ansara 2020, PMID 33224690).
  • Does quetiapine augmentation work? Two systematic reviews of overlapping literature reach opposite conclusions (Samuel 2011, PMID 21088608 vs Schiele 2026, PMID 40946318).
  • Is the rTMS effect real? SMD −1.86 from 152 patients (Parikh 2022, PMID 34791241) would make it the most effective GAD treatment ever measured, which is itself grounds for scepticism.
  • Should the sequence be drug→drug or drug→therapy? CBT works after pharmacological failure (Schiele 2026, PMID 40946318) and further pharmacotherapy can work after first-drug failure (Slee 2019, PMID 30712879). A PubMed search rerun on 2026-09-02 located no GAD trial directly randomising those sequences; this is a dated comparative-effectiveness gap.
  • Does psychedelic-assisted treatment hold at phase 3, and in whom? MM120's phase 2b enrolled moderate-to-severe rather than explicitly treatment-resistant GAD (Robison 2025, PMID 40906494).

References

  1. Samuel M, et al. Efficacy and safety of treatments for refractory generalized anxiety disorder: a systematic review. Int Clin Psychopharmacol. 2011;26:63-8. PMID 21088608
  2. Schiele MA, et al. Integrative Systematic Review on Pharmacological, Psychotherapeutic, and Neurostimulatory Treatment Options in Treatment-Resistant Anxiety Disorders. Psychother Psychosom. 2026;95:214-247. PMID 40946318
  3. Ansara ED. Management of treatment-resistant generalized anxiety disorder. Ment Health Clin. 2020;10:326-334. PMID 33224690
  4. Pollack MH. Refractory generalized anxiety disorder. J Clin Psychiatry. 2009;70 Suppl 2:32-8. PMID 19371505
  5. Generoso MB, et al. Pregabalin for generalized anxiety disorder: an updated systematic review and meta-analysis. Int Clin Psychopharmacol. 2017;32:49-55. PMID 27643884
  6. Slee A, et al. Pharmacological treatments for generalised anxiety disorder: a systematic review and network meta-analysis. Lancet. 2019;393:768-777. PMID 30712879
  7. Altamura AC, et al. Augmentative quetiapine in partial/nonresponders with generalized anxiety disorder: a randomized, placebo-controlled study. Int Clin Psychopharmacol. 2011;26:201-5. PMID 21403524
  8. Hoge EA, et al. Aripiprazole as augmentation treatment of refractory generalized anxiety disorder and panic disorder. CNS Spectr. 2008;13:522-7. PMID 18567977
  9. Worthington JJ, et al. Aripiprazole as an augmentor of selective serotonin reuptake inhibitors in depression and anxiety disorder patients. Int Clin Psychopharmacol. 2005;20:9-11. PMID 15602109
  10. Lorenz RA, et al. Adjunctive use of atypical antipsychotics for treatment-resistant generalized anxiety disorder. Pharmacotherapy. 2010;30:942-51. PMID 20795849
  11. Garakani A, et al. Antipsychotic agents in anxiety disorders: An umbrella review. Acta Psychiatr Scand. 2024;149:295-312. PMID 38382649
  12. Depping AM, et al. Second-generation antipsychotics for anxiety disorders. Cochrane Database Syst Rev. 2010:CD008120. PMID 21154392
  13. Daly EJ, et al. The effect of esketamine in patients with treatment-resistant depression with and without comorbid anxiety symptoms or disorder. Depress Anxiety. 2021;38:1120-1130. PMID 34293233
  14. Kawashima Y, et al. Effects of riluzole on psychiatric disorders with anxiety or fear as primary symptoms: A systematic review. Neuropsychopharmacol Rep. 2023;43:320-327. PMID 37463744
  15. Pittenger C, et al. Riluzole in the treatment of mood and anxiety disorders. CNS Drugs. 2008;22:761-86. PMID 18698875
  16. DeMartinis N, et al. Prior benzodiazepine use and buspirone response in the treatment of generalized anxiety disorder. J Clin Psychiatry. 2000;61:91-4. PMID 10732655
  17. Chen A, et al. A Proposed Algorithm for the Pharmacological Treatment of Generalized Anxiety Disorder in the Older Patient. J Geriatr Psychiatry Neurol. 2025;38:155-171. PMID 39352792
  18. Fava GA, et al. Sequential treatment of mood and anxiety disorders. J Clin Psychiatry. 2005;66:1392-400. PMID 16420076
  19. Parikh TK, et al. Repetitive Transcranial Magnetic Stimulation for Generalized Anxiety Disorder: A Systematic Literature Review and Meta-Analysis. Int J Neuropsychopharmacol. 2022;25:144-146. PMID 34791241
  20. Duan N, et al. Evaluating the efficacy and acceptability of non-invasive brain stimulation for generalized anxiety disorder: a systematic review and network meta-analysis. Psychiatry Res Neuroimaging. 2025;349:111989. PMID 40203547
  21. Robison R, et al. Single Treatment With MM120 (Lysergide) in Generalized Anxiety Disorder: A Randomized Clinical Trial. JAMA. 2025;334:1358-1372. PMID 40906494
  22. Schanzer B, et al. Novel investigational therapeutics for generalized anxiety disorder (GAD). Expert Opin Investig Drugs. 2019;28:1003-1012. PMID 31607187
  23. Fagan HA, Baldwin DS. Pharmacological Treatment of Generalised Anxiety Disorder: Current Practice and Future Directions. Expert Rev Neurother. 2023;23:535-548. PMID 37183813
  24. Gould RL, et al. Acceptance and commitment therapy for older people with treatment-resistant generalised anxiety disorder: the FACTOID feasibility study. Health Technol Assess. 2021;25:1-150. PMID 34542399
  25. Gould RL, et al. Acceptance and commitment therapy for late-life treatment-resistant generalised anxiety disorder: a feasibility study. Age Ageing. 2021;50:1751-1761. PMID 33852722
  26. Nordahl HM, et al. Metacognitive therapy versus cognitive-behavioural therapy in adults with generalised anxiety disorder. BJPsych Open. 2018;4:393-400. PMID 30294448
  27. Hyde J, et al. Efficacy of neurostimulation across mental disorders: systematic review and meta-analysis of 208 randomized controlled trials. Mol Psychiatry. 2022;27:2709-2719. PMID 35365806
  28. Qi L, et al. Non-invasive brain stimulation in the treatment of generalized anxiety disorder: A systematic review and meta-analysis. J Psychiatr Res. 2024;178:378-387. PMID 39208534
  29. Kong W, et al. Comparative Remission Rates and Tolerability of Drugs for Generalised Anxiety Disorder: A Systematic Review and Network Meta-analysis of Double-Blind Randomized Controlled Trials. Front Pharmacol. 2020;11:580858. PMID 33343351
  30. Jakubovski E, et al. Systematic review and meta-analysis: Dose-response curve of SSRIs and SNRIs in anxiety disorders. Depress Anxiety. 2019;36:198-212. PMID 30479005
  31. Baldwin DS, et al. Escitalopram and paroxetine in the treatment of generalised anxiety disorder: randomised, placebo-controlled, double-blind study. Br J Psychiatry. 2006;189:264-72. PMID 16946363
  32. 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
  33. Domschke K, et al. The definition of treatment resistance in anxiety disorders: a Delphi method-based consensus guideline. World Psychiatry. 2024;23:113-123. PMID 38214637