Skip to content

Red flags and safety concerns

TL;DR — Four safety domains matter in GAD, and they are asymmetric in how well they are quantified. (1) 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 in Canadian national data (Gilmour 2016, PMID 27849314). (2) Medication harms: antidepressants carry an AE-dropout NNTH of 17 (95% CI 13–112) in GAD (Kopcalic 2025, PMID 39880377); SSRIs raise fall risk in older adults (adjusted OR 2.02, 95% CI 1.85–2.20 — higher than benzodiazepines at 1.42) (Seppala 2018, PMID 29402652); antidepressant exposure raises hyponatraemia odds 3.16-fold (1.911–5.225), with SNRIs worse than SSRIs (OR 1.292) and mirtazapine best (0.607) (Gheysens 2024, PMID 38403888); SSRIs increase bleeding risk by ≥36% (pooled OR 1.41, 1.27–1.57 across 42 observational studies) (Laporte 2017, PMID 27521835); and antidepressants in paediatric patients carry a modestly increased suicidality risk (RR 1.95, 95% CI 1.28–2.98 across all indications; risk difference 0.02) with no completed suicides in 4,582 patients across 24 trials (Hammad 2006, PMID 16520440). (3) Dependence: benzodiazepine deprescribing is recommended for all adults ≥65 regardless of duration and for adults 18–64 after >4 weeks — although explicitly not for patients with untreated anxiety (Pottie 2018, PMID 29760253); gabapentinoid misuse is a documented and growing harm concentrated in people with opioid or multi-drug use disorder (Evoy 2021, PMID 33215352; Bonnet 2017, PMID 28988943). (4) Medical mimics: anxiety is a common symptom of thyrotoxicosis, which affects 0.2–1.4% of people worldwide (Lee 2023, PMID 37847271) — but over-testing for rarer mimics has its own cost: 10–22% of urine normetanephrine results exceed the reference limit in real-world pheochromocytoma screening, almost all of them false positives (Kline 2020, PMID 31978379).

1. Suicide risk

Finding Value Source
Past-year suicidal ideation, threshold GAD 32.0% (men), 21.2% (women) Gilmour 2016, PMID 27849314
Threshold and subthreshold GAD vs no GAD Both significantly associated with ideation after adjusting for sociodemographics and psychiatric comorbidity (attenuated but significant) Gilmour 2016, PMID 27849314
Non-excessive GAD (criteria-excluded cases) Comparable risk of temporally secondary comorbidity and suicidality to full GAD cases Ruscio 2024, PMID 39364896
Veterans Dose–response relation between anxiety symptom severity and suicidal thoughts/behaviours; probable GAD most strongly associated Macdonald-Gagnon 2024, PMID 38325107
Primary care Suicidality elevated in pure GAD, pure MDE and highest in comorbid GAD/MDE Wittchen 2002, PMID 12044105
Comorbid depression Lowers recovery and raises recurrence over 12 years Bruce 2005, PMID 15930067

The clinically load-bearing point is that the subthreshold group carries the signal too (Gilmour 2016, PMID 27849314; Ruscio 2024, PMID 39364896). A criteria-based triage that treats not-quite-GAD as not-at-risk is not supported by the data (the diagnostic boundary).

2. Medical conditions that present as generalized anxiety

Condition Why it matters Evidence
Thyrotoxicosis / hyperthyroidism Anxiety, insomnia, palpitations, weight loss, diarrhoea and heat intolerance are listed common symptoms; overt hyperthyroidism affects 0.2–1.4% and subclinical 0.7–1.4% worldwide; Graves disease prevalence 2% in women, 0.5% in men Lee 2023, PMID 37847271
Graves disease, post-treatment Depression and anxiety scores remained higher than matched controls even after 15 months of treatment (anxiety median 4.0 vs 3.0, p<0.05); prior psychiatric condition, mild eye symptoms and younger age predicted more residual anxiety Holmberg 2024, PMID 38215285
Bidirectional association In 349,993 UK Biobank participants over 13 years, baseline PHQ-4 depression/anxiety predicted incident hyperthyroidism (adjusted HR 1.19 mild → 1.84 severe) and hypothyroidism, in a linear dose–response Fan 2024, PMID 40226662
Caffeine In panic disorder, caffeine challenge produced panic attacks in 51.1% of 237 patients and none after placebo; patients 53.9% vs healthy controls 1.7% (log RR 3.47, 95% CI 2.06–4.87); subjective anxiety Hedges' g=1.02. Doses used were 400–750 mg (~5 cups) Klevebrant 2022, PMID 34871964 — panic disorder, not GAD
Pheochromocytoma A classic mimic, but the screening test performs badly at population scale: 10–22% of urine normetanephrine results exceeded the upper reference limit (highest at ages 40–60), while only 0.7% of 12,572 patients had results compatible with likely PPGL; 99% of high results were <3-fold the limit Kline 2020, PMID 31978379
Cardiopulmonary disease Among 180 veterans with COPD or heart failure meeting criteria for anxiety, depression or PTSD, only 39% had a documented mental-health diagnosis — the mimic problem runs in both directions Ratcliff 2017, PMID 28807139
Cardiovascular prognosis GAD in stable coronary heart disease: 62% higher cardiovascular event rate after full adjustment including major depression (HR 1.62, 1.11–2.37) Martens 2010, PMID 20603456

Two cautions follow. First, investigation should be targeted to clinical features: the pheochromocytoma screening data show a high false-positive burden and do not support population-wide metanephrine testing for anxiety (Kline 2020, PMID 31978379). Second, anxiety can persist after treatment of thyroid disease (Holmberg 2024, PMID 38215285), so persistence alone neither proves nor disproves the original attribution.

3. Medication safety

Antidepressants

Harm Magnitude Source
AE-related dropout in GAD trials RR 2.18 (1.81–2.61) vs placebo; NNTH 17 (13–112); high certainty Kopcalic 2025, PMID 39880377
Falls, older adults (adjusted OR) Antidepressants 1.57 (1.43–1.74); SSRIs 2.02 (1.85–2.20); TCAs 1.41; antipsychotics 1.54; benzodiazepines 1.42; long-acting BZDs 1.81; short-acting 1.27 Seppala 2018, PMID 29402652
Hyponatraemia Antidepressant exposure OR 3.160 (1.911–5.225); event rates SNRIs 7.44% > SSRIs 5.59% > TCAs 2.66% > mirtazapine 1.02% > trazodone 0.89%; SNRIs vs SSRIs OR 1.292 (1.120–1.491); mirtazapine vs SSRIs 0.607 (0.385–0.957) Gheysens 2024, PMID 38403888 (39 studies, n=8,175,111)
Bleeding Pooled adjusted OR 1.41 (1.27–1.57) across 42 observational studies; 1.41 in case-control (1,255,073 patients), 1.36 in cohort studies (187,956) Laporte 2017, PMID 27521835
Discontinuation symptoms ≥1 symptom in 31% after stopping an antidepressant vs 17% after placebo (RCT difference 8%); severe 2.8% vs 0.6%; desvenlafaxine, venlafaxine, imipramine and escitalopram highest frequency Henssler 2024, PMID 38851198
Discontinuation, nature DESS at week 1 SMD 0.31 (≈1 extra symptom, below the clinical-syndrome threshold); dizziness OR 5.52 (3.81–8.01), nausea 3.16, vertigo 6.40, nervousness 3.15; mood worsening not associated — later depression indicates relapse Kalfas 2025, PMID 40632531
Paediatric suicidality RR 1.95 (1.28–2.98) across all indications; SSRIs in depression 1.66 (1.02–2.68); risk difference 0.02 (0.01–0.03); no completed suicides in 4,582 patients / 24 trials, of which 2 were GAD trials Hammad 2006, PMID 16520440
Paediatric adverse effects, management Reviewed Strawn 2023, PMID 36651686
Paediatric deprescribing Systematic review of discontinuation approaches, cross-titration and withdrawal Stimpfl 2025, PMID 39469761

The SSRI fall-risk figure (OR 2.02) exceeding the benzodiazepine figure (1.42) is counter-intuitive and worth stating explicitly (Seppala 2018, PMID 29402652); the authors note prescription bias could not be accounted for, and heterogeneity was substantial. It does not license substituting benzodiazepines for SSRIs in older adults, but it does undermine the assumption that SSRIs are the automatically safer choice on fall risk alone (special populations).

Benzodiazepines and gabapentinoids

Concern Evidence
Deprescribing guidance Taper slowly and offer deprescribing to all adults ≥65 regardless of duration of use, and to adults 18–64 who have used BZRAs >4 weeks. Explicitly stated not to apply to people with untreated anxiety, depression or other conditions causing the insomnia (Pottie 2018, PMID 29760253)
Guideline position in GAD Effective short-term; should not be given long-term because of addiction risk (Bandelow 2013, PMID 23671484); caution with benzodiazepines and hydroxyzine in older adults, with periodic benefit–harm review for legacy prescriptions (Chen 2025, PMID 39352792)
Countervailing trial evidence 56 studies, 7,556 participants: BZDs efficacious with a favourable safety profile; only diazepam differed from placebo on tolerability (RR 1.61, 1.32–1.96) (Fernandes 2025, PMID 40544830)
Effect on subsequent treatment Recent (<1 month) prior benzodiazepine use predicted higher attrition, more adverse events and the smallest buspirone–placebo difference (DeMartinis 2000, PMID 10732655)
Cognitive effects Reviewed (Stewart 2005, PMID 15762814)
Deprescribing interventions that work Masked taper with behavioural intervention (Fung 2024, PMID 39374004); patient self-guided direct-to-consumer education (YAWNS NB) (Gardner 2024, PMID 39292452)
Gabapentinoid misuse 55 studies across four continents: misuse to self-medicate, usually with other drugs; opioid use disorder is the greatest risk factor; increased hospital utilisation and opioid-related overdose mortality risk (Evoy 2021, PMID 33215352)
How addictive? 106 studies: no evidence of vigorous addictive power in people without prior substance use; pregabalin somewhat more addictive than gabapentin; principal at-risk population is people with current or past substance use disorders, mostly opioid and multi-drug users, who preferred pregabalin; pure overdose relatively safe but lethal in combination with opioids and sedatives — recommendation is to avoid gabapentinoids in patients with a SUD history or use them with strict monitoring (Bonnet 2017, PMID 28988943)

Second-generation antipsychotics

Quetiapine has efficacy evidence in GAD (Slee 2019, PMID 30712879; Depping 2010, PMID 21154392), but its regulatory status varies by jurisdiction. The Cochrane review documented more adverse-event dropout, weight gain, sedation and extrapyramidal effects than placebo, and more adverse-event dropout, weight gain and sedation than antidepressants (Depping 2010, PMID 21154392). An umbrella review of 25 systematic reviews rated 24 as low quality on AMSTAR-2 and found no high-quality antipsychotic evidence in anxiety outside quetiapine in GAD (Garakani 2024, PMID 38382649). Antipsychotics also carry fall risk (adjusted OR 1.54; Seppala 2018, PMID 29402652).

4. Trial-specific and emerging safety issues

  • Psychedelic trials: visual perceptual changes (illusion, pseudo-hallucination, visual hallucination) occurred in 46.2% of MM120-dosed participants (Robison 2025, PMID 40906494) — both an adverse event and a functional-unblinding problem (clinical trials landscape).
  • Higher doses cost tolerability: for both SSRIs and SNRIs, higher doses increase dropout due to side effects, while only SSRIs gain efficacy from dose escalation (Jakubovski 2019, PMID 30479005).
  • MBSR versus escitalopram adverse-event asymmetry: ≥1 study-related adverse event in 78.6% of escitalopram participants versus 15.4% of MBSR participants; 8% vs 0% dropped out for adverse events (Hoge 2023, PMID 36350591).

Practical red-flag list

Situation Why it is a flag
New-onset anxiety with weight loss, heat intolerance, palpitations or diarrhoea Consider thyrotoxicosis in the differential (Lee 2023, PMID 37847271)
Anxiety persisting after successful thyroid treatment Persistence has been reported; reassess rather than assuming either persistence or misattribution (Holmberg 2024, PMID 38215285)
High caffeine intake (≈5 cups/day and above) Caffeine challenge at 400–750 mg is anxiogenic and panicogenic in vulnerable people (Klevebrant 2022, PMID 34871964)
Subthreshold GAD Independently associated with suicidal ideation (Gilmour 2016, PMID 27849314)
Established coronary heart disease with GAD 62% higher cardiovascular event rate (Martens 2010, PMID 20603456)
History of opioid or multi-drug use disorder Avoid gabapentinoids or monitor strictly (Bonnet 2017, PMID 28988943; Evoy 2021, PMID 33215352)
Older adult on a legacy benzodiazepine Deprescribing recommended regardless of duration — but not while the underlying anxiety is untreated (Pottie 2018, PMID 29760253)
Older adult starting an SSRI Fall risk OR 2.02 and hyponatraemia risk; consider mirtazapine where hyponatraemia is the dominant concern (Seppala 2018, PMID 29402652; Gheysens 2024, PMID 38403888)
SSRI exposure Observational studies found pooled bleeding OR 1.41 (95% CI 1.27–1.57); this source does not isolate an NSAID or antiplatelet interaction (Laporte 2017, PMID 27521835)
Child or adolescent starting an antidepressant Modest suicidality signal, RR 1.95, risk difference 2% (Hammad 2006, PMID 16520440)
Abrupt antidepressant cessation 31% get ≥1 discontinuation symptom, dizziness most common; mood worsening indicates relapse, not withdrawal (Henssler 2024, PMID 38851198; Kalfas 2025, PMID 40632531)

Open questions

  • Why do SSRIs carry a higher adjusted fall-risk OR than benzodiazepines in older adults (Seppala 2018, PMID 29402652)? Prescription bias is the authors' own candidate explanation and has not been resolved.
  • Is the benzodiazepine dependence risk in GAD quantified anywhere at trial scale? The within-class network measures dropout, not dependence (Fernandes 2025, PMID 40544830); the guideline position rests on a different evidence base (guidelines).
  • What is the yield of targeted medical investigation in new-onset GAD? A PubMed search rerun on 2026-09-02 located no cohort estimating thyroid or catecholamine abnormalities specifically among people presenting with apparent GAD; this is a dated evidence gap, not support for indiscriminate testing.
  • Does treating GAD reduce cardiovascular events? The observational signal is consistent in both directions (Martens 2010, PMID 20603456; Wu 2025, PMID 40466339) but untested in a trial.
  • Are the paediatric suicidality data from GAD trials specifically separable? Only 2 of the 24 pooled trials were GAD trials (Hammad 2006, PMID 16520440).

References

  1. Gilmour H. Threshold and subthreshold Generalized Anxiety Disorder (GAD) and suicide ideation. Health Rep. 2016;27:13-21. PMID 27849314
  2. Ruscio AM, et al. The case for eliminating excessive worry as a requirement for generalized anxiety disorder. Psychol Med. 2024;54:3447-3458. PMID 39364896
  3. Macdonald-Gagnon G, et al. Generalized anxiety and mild anxiety symptoms in U.S. military veterans. J Psychiatr Res. 2024;171:263-270. PMID 38325107
  4. Wittchen HU, et al. Generalized anxiety and depression in primary care. J Clin Psychiatry. 2002;63 Suppl 8:24-34. PMID 12044105
  5. Bruce SE, et al. Influence of psychiatric comorbidity on recovery and recurrence in generalized anxiety disorder, social phobia, and panic disorder. Am J Psychiatry. 2005;162:1179-87. PMID 15930067
  6. Lee SY, Pearce EN. Hyperthyroidism: A Review. JAMA. 2023;330:1472-1483. PMID 37847271
  7. Holmberg M, et al. Psychiatric complications in Graves' disease. Eur Thyroid J. 2024;13. PMID 38215285
  8. Fan T, et al. The Association between Depression, Anxiety, and Thyroid Disease: A UK Biobank Prospective Cohort Study. Depress Anxiety. 2024;2024:8000359. PMID 40226662
  9. Klevebrant L, Frick A. Effects of caffeine on anxiety and panic attacks in patients with panic disorder: A systematic review and meta-analysis. Gen Hosp Psychiatry. 2022;74:22-31. PMID 34871964
  10. Kline GA, et al. Very high rate of false positive biochemical results when screening for pheochromocytoma in a large, undifferentiated population. Clin Biochem. 2020;77:26-31. PMID 31978379
  11. Ratcliff CG, et al. Recognition of anxiety, depression, and PTSD in patients with COPD and CHF: Who gets missed? Gen Hosp Psychiatry. 2017;47:61-67. PMID 28807139
  12. Martens EJ, et al. Scared to death? Generalized anxiety disorder and cardiovascular events in patients with stable coronary heart disease. Arch Gen Psychiatry. 2010;67:750-8. PMID 20603456
  13. Wu JY, et al. Association between SSRI use and cardiovascular outcomes in patients with coronary artery disease and generalized anxiety disorder. Atherosclerosis. 2025;407:120390. PMID 40466339
  14. Kopcalic K, et al. Antidepressants versus placebo for generalised anxiety disorder (GAD). Cochrane Database Syst Rev. 2025;1:CD012942. PMID 39880377
  15. Seppala LJ, et al. Fall-Risk-Increasing Drugs: A Systematic Review and Meta-Analysis: II. Psychotropics. J Am Med Dir Assoc. 2018;19:371.e11-371.e17. PMID 29402652
  16. Gheysens T, et al. The risk of antidepressant-induced hyponatremia: A meta-analysis of antidepressant classes and compounds. Eur Psychiatry. 2024;67:e20. PMID 38403888
  17. Laporte S, et al. Bleeding risk under selective serotonin reuptake inhibitor (SSRI) antidepressants: A meta-analysis of observational studies. Pharmacol Res. 2017;118:19-32. PMID 27521835
  18. Henssler J, et al. Incidence of antidepressant discontinuation symptoms: a systematic review and meta-analysis. Lancet Psychiatry. 2024;11:526-535. PMID 38851198
  19. Kalfas M, et al. Incidence and Nature of Antidepressant Discontinuation Symptoms. JAMA Psychiatry. 2025;82:896-904. PMID 40632531
  20. Hammad TA, et al. Suicidality in pediatric patients treated with antidepressant drugs. Arch Gen Psychiatry. 2006;63:332-9. PMID 16520440
  21. Strawn JR, et al. Adverse Effects of Antidepressant Medications and their Management in Children and Adolescents. Pharmacotherapy. 2023;43:675-690. PMID 36651686
  22. Stimpfl JN, et al. Deprescribing Antidepressants in Children and Adolescents. J Child Adolesc Psychopharmacol. 2025;35:3-22. PMID 39469761
  23. Pottie K, et al. Deprescribing benzodiazepine receptor agonists: Evidence-based clinical practice guideline. Can Fam Physician. 2018;64:339-351. PMID 29760253
  24. Bandelow B, et al. The diagnosis and treatment of generalized anxiety disorder. Dtsch Arztebl Int. 2013;110:300-9. PMID 23671484
  25. 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
  26. Fernandes H, et al. Comparative Efficacy and Safety of Benzodiazepines in the Treatment of Patients with Generalized Anxiety Disorder. Psychother Psychosom. 2025;94:373-388. PMID 40544830
  27. 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
  28. Stewart SA. The effects of benzodiazepines on cognition. J Clin Psychiatry. 2005;66 Suppl 2:9-13. PMID 15762814
  29. Fung CH, et al. Masked Taper With Behavioral Intervention for Discontinuation of Benzodiazepine Receptor Agonists: A Randomized Clinical Trial. JAMA Intern Med. 2024;184:1448-1456. PMID 39374004
  30. Gardner DM, et al. Patient Self-Guided Interventions to Reduce Sedative Use and Improve Sleep: The YAWNS NB Randomized Clinical Trial. JAMA Psychiatry. 2024;81:1187-1197. PMID 39292452
  31. Evoy KE, et al. Abuse and Misuse of Pregabalin and Gabapentin: A Systematic Review Update. Drugs. 2021;81:125-156. PMID 33215352
  32. Bonnet U, Scherbaum N. How addictive are gabapentin and pregabalin? A systematic review. Eur Neuropsychopharmacol. 2017;27:1185-1215. PMID 28988943
  33. Depping AM, et al. Second-generation antipsychotics for anxiety disorders. Cochrane Database Syst Rev. 2010:CD008120. PMID 21154392
  34. Garakani A, et al. Antipsychotic agents in anxiety disorders: An umbrella review. Acta Psychiatr Scand. 2024;149:295-312. PMID 38382649
  35. Slee A, et al. Pharmacological treatments for generalised anxiety disorder. Lancet. 2019;393:768-777. PMID 30712879
  36. Robison R, et al. Single Treatment With MM120 (Lysergide) in Generalized Anxiety Disorder. JAMA. 2025;334:1358-1372. PMID 40906494
  37. 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
  38. Hoge EA, et al. Mindfulness-Based Stress Reduction vs Escitalopram for the Treatment of Adults With Anxiety Disorders. JAMA Psychiatry. 2023;80:13-21. PMID 36350591