Skip to content

Suicide risk and mortality

TL;DR — MDD is strongly associated with suicidal ideation, attempt, and death, but most people with MDD do not die by suicide and prediction at the individual level remains weak. Meta-analysis of MDD cohorts identifies prior suicidal behavior, symptom severity, hopelessness, comorbidity, and acute stressors as risk correlates, not deterministic rules (Li 2022, PMID 35101521). Mental disorders overall carry about twice the all-cause mortality risk (Walker 2015, PMID 25671328), while GBD assigns nearly no direct years of life lost to mental disorders, exposing an attribution gap (GBD 2019, PMID 35026139). Safety depends less on a score than on direct inquiry, dynamic reassessment, access-to-means reduction, continuity after crises, and rapid response to change.

Risk architecture

Domain Examples Interpretation
History prior attempt, self-harm, hospitalization strongest broad markers, still nonspecific
Current state intent, plan, preparation, agitation, psychosis dynamic and decision-relevant
Illness severe depression, mixed features, substance use raises concern; reassess diagnosis
Context loss, violence, legal/financial crisis, isolation can change rapidly
Means access to lethal methods modifiable immediate risk
Protection reasons for living, support, engagement not guarantees

The cohort meta-analysis of predictors in MDD found multiple statistically significant factors but substantial heterogeneity and limited predictive certainty (Li 2022, PMID 35101521). Cognitive functioning also relates to suicidal outcomes across severe mental disorders, but evidence is not specific enough for a clinical test (Le 2024, PMID 39168166).

Why categories fail

Suicide is rare relative to suicidal thoughts, so even a high-risk label generates many false positives. Low-risk labels are unsafe because risk fluctuates and uncommon outcomes still occur. Risk stratification should organize actions, not claim certainty.

Weak practice Stronger practice
“Contract for safety” collaborative safety plan with concrete steps
One-time screen reassessment after transitions and changes
Score-only disposition clinical formulation plus means/access and support
Generic advice named contacts, crisis pathway, follow-up ownership

Treatment and suicidality

Rapid symptom change is not identical to proven suicide prevention. Ketamine/esketamine can reduce suicidal ideation quickly in some trials, but trials are not powered for suicide death. ECT is used for severe suicidal depression because of rapid, high acute efficacy, but continuity planning remains essential. Antidepressant initiation and dose changes require monitoring, especially in younger people and those with activation or mixed features.

Genomic meta-analysis of suicide attempt identified 12 genome-wide significant loci and substantial shared psychiatric genetic liability; this advances biology but not near-term individual prediction (Docherty 2023, PMID 37777856).

Mortality beyond suicide

Excess mortality can arise through cardiovascular/metabolic disease, smoking, substance use, accidents, treatment effects, and reduced access to medical care. Across mental disorders, pooled mortality risk was approximately doubled and life expectancy shortened, though estimates varied by diagnosis and setting (Walker 2015, PMID 25671328). TRD is associated with excess deaths in population data (Brenner 2021, PMID 33912340), but confounding by severity and comorbidity limits causal interpretation.

High-risk transitions

Transition Why vulnerable
Emergency discharge acute crisis may persist; handoff failure
Psychiatric discharge abrupt loss of containment
Treatment start/change activation, disappointment, early adverse effects
Treatment nonresponse hopelessness and worsening
Relationship/legal/financial shock rapid intent change
Intoxication/withdrawal impaired judgment and disinhibition

Risk prediction and intervention evidence

Suicide risk is dynamic and low-base-rate even in high-risk groups. High relative risks can coexist with poor positive predictive value.

Evidence question Quantitative direction Limitation
Do risk scales identify future suicide? Meta-analyses find associations between high-risk categories and outcomes Positive predictive values are low and thresholds miss many events
Do clinician predictions outperform scales? Clinician judgment identifies some elevation Accuracy remains insufficient for binary discharge or coercion decisions
Does suicidal ideation predict suicide? Ideation raises relative risk Sensitivity and PPV are too low for ideation absence to imply safety
Do antidepressants alter suicidality? FDA analyses show age-dependent effects, with concern in younger adults Trial events are rare and ideation/behavior composites vary
Does ECT reduce suicide outcomes? Observational and meta-analytic evidence suggests lower suicidality/mortality Confounding by indication and immortal-time bias remain possible
Do brief interventions work? Safety planning and follow-up interventions reduce suicidal behavior in some trials Effects depend on implementation and service continuity

Mortality synthesis must separate suicide from cardiovascular, respiratory, metabolic, substance-related, and care-access pathways. Depression-associated all-cause mortality is not equivalent to mortality caused directly by an episode.

Additional live-search evidence ledger

The records below were added after full PubMed E-utilities retrieval on 2026-08-30. The ledger states the evidentiary role of each record and preserves the design limitation that should travel with its citation.

  • Chan JKN 2025 — All-cause and cause-specific mortality in people with depression: a large-scale systematic review and meta-analysis of relative risk and aggravating or attenuating factors, including antidepressant treatment. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Chan JKN 2025, PMID 40948054)

  • Naismith H 2025 — Systematic review and meta-analysis of the effectiveness of ECT in reducing suicidal ideation, self-harm, suicide, and mortality. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Naismith H 2025, PMID 41165097)

  • Jabbi M 2020 — Neuro-transcriptomic signatures for mood disorder morbidity and suicide mortality. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Jabbi M 2020, PMID 32485434)

  • Rhee TG 2025 — Longitudinal associations of electroconvulsive therapy with all-cause mortality and suicide deaths in severe unipolar or bipolar depression: a systematic review and meta-analysis. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Rhee TG 2025, PMID 41407487)

  • Odermatt J 2025 — Electroconvulsive therapy reduces suicidality and all-cause mortality in refractory depression: A systematic review and meta-analysis of neurostimulation studies. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Odermatt J 2025, PMID 41323432)

  • Spittal MJ 2025 — Machine learning algorithms and their predictive accuracy for suicide and self-harm: Systematic review and meta-analysis. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Spittal MJ 2025, PMID 40934153)

  • Large M 2018 — Suicide risk assessment among psychiatric inpatients: a systematic review and meta-analysis of high-risk categories. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Large M 2018, PMID 28874218)

  • Carter G 2017 — Predicting suicidal behaviours using clinical instruments: systematic review and meta-analysis of positive predictive values for risk scales. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Carter G 2017, PMID 28302700)

  • McHugh CM 2019 — Association between suicidal ideation and suicide: meta-analyses of odds ratios, sensitivity, specificity and positive predictive value. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (McHugh CM 2019, PMID 30702058)

  • Woodford R 2019 — Accuracy of Clinician Predictions of Future Self-Harm: A Systematic Review and Meta-Analysis of Predictive Studies. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Woodford R 2019, PMID 28972271)

  • Kohnepoushi P 2026 — Diagnostic accuracy of machine learning approaches for suicide‑related outcomes: a meta‑analysis. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Kohnepoushi P 2026, PMID 42106827)

  • Harris IM 2019 — Predicting future self-harm or suicide in adolescents: a systematic review of risk assessment scales/tools. Systematic review; useful for mapping consistency and gaps, not automatically a pooled causal estimate. (Harris IM 2019, PMID 31494608)

  • Reeves RR 2010 — Antidepressant-induced suicidality: an update. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Reeves RR 2010, PMID 20553304)

  • Reeves RR 2009 — Antidepressant-induced suicidality: implications for clinical practice. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Reeves RR 2009, PMID 19488000)

  • Carpenter DJ 2011 — Meta-analysis of efficacy and treatment-emergent suicidality in adults by psychiatric indication and age subgroup following initiation of paroxetine therapy: a complete set of randomized placebo-controlled trials. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Carpenter DJ 2011, PMID 21367354)

  • Stone M 2009 — Risk of suicidality in clinical trials of antidepressants in adults: analysis of proprietary data submitted to US Food and Drug Administration. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Stone M 2009, PMID 19671933)

  • Vanderburg DG 2009 — A pooled analysis of suicidality in double-blind, placebo-controlled studies of sertraline in adults. Randomized comparison; population, control credibility, duration, and missingness bound transportability. (Vanderburg DG 2009, PMID 19552866)

  • Wightman DS 2010 — Meta-analysis of suicidality in placebo-controlled clinical trials of adults taking bupropion. Meta-analysis; pooled estimates depend on eligibility, heterogeneity, and reporting bias. (Wightman DS 2010, PMID 21274361)

  • Eugene AR 2024 — Country-specific psychopharmacological risk of reporting suicidality comparing 38 antidepressants and lithium from the FDA Adverse Event Reporting System, 2017-2023. Primary or secondary empirical evidence; interpretation should follow its design and comparator rather than the headline alone. (Eugene AR 2024, PMID 39575192)

Open questions

  • Which dynamic measures improve action without false reassurance (Li 2022, PMID 35101521)?
  • Do rapid-acting treatments reduce attempts or deaths, not only ideation scores?
  • Which post-discharge service models most reduce suicide mortality?
  • How should mortality attribution incorporate medical pathways without double counting (Walker 2015, PMID 25671328)?

References

  1. Li X, et al. Predictors of suicidal ideation, attempt and death in MDD. Journal of Affective Disorders. 2022. PMID 35101521
  2. Walker ER, et al. Mortality in mental disorders and global disease burden implications. JAMA Psychiatry. 2015. PMID 25671328
  3. GBD 2019 Mental Disorders Collaborators. Global burden of 12 mental disorders. Lancet Psychiatry. 2022. PMID 35026139
  4. Docherty AR, et al. GWAS Meta-Analysis of Suicide Attempt. American Journal of Psychiatry. 2023. PMID 37777856
  5. Brenner P, et al. Excess deaths in treatment-resistant depression. Therapeutic Advances in Psychopharmacology. 2021. PMID 33912340
  6. Le GH, et al. Cognitive functioning and suicidal outcomes across severe mental disorders. Psychological Medicine. 2024. PMID 39168166
  7. Chan JKN, et al. All-cause and cause-specific mortality in people with depression: a large-scale systematic review and meta-analysis of relative risk and aggravating or attenuating factors, including antidepressant treatment. World psychiatry : official journal of the World Psychiatric Association (WPA). 2025;24:404-421. PMID 40948054
  8. Naismith H, et al. Systematic review and meta-analysis of the effectiveness of ECT in reducing suicidal ideation, self-harm, suicide, and mortality. Psychological medicine. 2025;55:e328. PMID 41165097
  9. Jabbi M, et al. Neuro-transcriptomic signatures for mood disorder morbidity and suicide mortality. Journal of psychiatric research. 2020;127:62-74. PMID 32485434
  10. Rhee TG, et al. Longitudinal associations of electroconvulsive therapy with all-cause mortality and suicide deaths in severe unipolar or bipolar depression: a systematic review and meta-analysis. BMJ mental health. 2025;28:e302083. PMID 41407487
  11. Odermatt J, et al. Electroconvulsive therapy reduces suicidality and all-cause mortality in refractory depression: A systematic review and meta-analysis of neurostimulation studies. Neuroscience applied. 2025;4:105520. PMID 41323432
  12. Spittal MJ, et al. Machine learning algorithms and their predictive accuracy for suicide and self-harm: Systematic review and meta-analysis. PLoS medicine. 2025;22:e1004581. PMID 40934153
  13. Large M, et al. Suicide risk assessment among psychiatric inpatients: a systematic review and meta-analysis of high-risk categories. Psychological medicine. 2018;48:1119-1127. PMID 28874218
  14. Carter G, et al. Predicting suicidal behaviours using clinical instruments: systematic review and meta-analysis of positive predictive values for risk scales. The British journal of psychiatry : the journal of mental science. 2017;210:387-395. PMID 28302700
  15. McHugh CM, et al. Association between suicidal ideation and suicide: meta-analyses of odds ratios, sensitivity, specificity and positive predictive value. BJPsych open. 2019;5:e18. PMID 30702058
  16. Woodford R, et al. Accuracy of Clinician Predictions of Future Self-Harm: A Systematic Review and Meta-Analysis of Predictive Studies. Suicide & life-threatening behavior. 2019;49:23-40. PMID 28972271
  17. Kohnepoushi P, et al. Diagnostic accuracy of machine learning approaches for suicide‑related outcomes: a meta‑analysis. Annals of general psychiatry. 2026;25:52. PMID 42106827
  18. Harris IM, et al. Predicting future self-harm or suicide in adolescents: a systematic review of risk assessment scales/tools. BMJ open. 2019;9:e029311. PMID 31494608
  19. Reeves RR, et al. Antidepressant-induced suicidality: an update. CNS neuroscience & therapeutics. 2010;16:227-34. PMID 20553304
  20. Reeves RR, et al. Antidepressant-induced suicidality: implications for clinical practice. Southern medical journal. 2009;102:713-8. PMID 19488000
  21. Carpenter DJ, et al. Meta-analysis of efficacy and treatment-emergent suicidality in adults by psychiatric indication and age subgroup following initiation of paroxetine therapy: a complete set of randomized placebo-controlled trials. The Journal of clinical psychiatry. 2011;72:1503-14. PMID 21367354
  22. Stone M, et al. Risk of suicidality in clinical trials of antidepressants in adults: analysis of proprietary data submitted to US Food and Drug Administration. BMJ (Clinical research ed.). 2009;339:b2880. PMID 19671933
  23. Vanderburg DG, et al. A pooled analysis of suicidality in double-blind, placebo-controlled studies of sertraline in adults. The Journal of clinical psychiatry. 2009;70:674-83. PMID 19552866
  24. Wightman DS, et al. Meta-analysis of suicidality in placebo-controlled clinical trials of adults taking bupropion. Primary care companion to the Journal of clinical psychiatry. 2010;12:PCC.09m00894. PMID 21274361
  25. Eugene AR. Country-specific psychopharmacological risk of reporting suicidality comparing 38 antidepressants and lithium from the FDA Adverse Event Reporting System, 2017-2023. Frontiers in psychiatry. 2024;15:1442490. PMID 39575192