Skip to content

Not yet independently audited

This page was written from verified sources but has not yet passed the independent citation audit. Treat its claims as provisional.

Suicide, mortality and physical health in bipolar disorder

TL;DR — Bipolar disorder is associated with approximately doubled all-cause mortality, a suicide mortality ratio around 14, pooled life expectancy near 67 years, and nearly 13 years of potential life lost, although estimates vary markedly by population and method. Natural causes account for much of the absolute loss: cardiovascular, respiratory, infectious and metabolic disease are all overrepresented, while screening and treatment after cardiovascular events are less frequent. Metabolic syndrome affects roughly one-third of bipolar populations, and obesity, hypertension and insulin resistance track adverse course. Suicide prevention cannot be reduced to medication choice: mixed states, prior attempts, substance use, trauma, access to means and transitions in care all matter. Lithium’s randomized-trial suicide estimate is directionally favorable but imprecise and nonsignificant, so causal certainty should not exceed the evidence.

Scale of premature mortality

A 31-study bipolar-specific meta-analysis estimated all-cause standardized mortality ratio (SMR) 2.05 (95% CI 1.89–2.23), with very high heterogeneity (I²=96.2%) (Hayes 2015, PMID 25735195). The average conceals differences by country, calendar period, health system, age and cohort severity.

Outcome Pooled estimate 95% CI Source
All-cause mortality SMR 2.05 1.89–2.23 Hayes 2015
Natural-cause mortality SMR 1.64 1.47–1.83 Hayes 2015
Unnatural-cause mortality SMR 7.42 6.43–8.55 Hayes 2015
Suicide mortality SMR 14.44 12.43–16.78 Hayes 2015
Pooled life expectancy 66.88 years 64.47–69.28 Chan 2022
Years of potential life lost 12.89 years 12.72–13.07 Chan 2022

Life-expectancy synthesis included 96,601 people; the YPLL synthesis included 128,989. Pooled life expectancy was 70.51 years in women and 64.59 in men, and was lowest in African samples (Chan 2022, PMID 35184778).

Natural causes accounted for 5.94 YPLL (95% CI 5.81–6.07) and unnatural causes 5.69 (5.59–5.79), showing that suicide and physical disease are coequal prevention priorities rather than competing narratives (Chan 2022, PMID 35184778).

Across mental disorders more broadly, 203 studies from 29 countries yielded an all-cause mortality RR of 2.22 (95% CI 2.12–2.33) and median 10 YPLL (Walker 2015, PMID 25671328). These broader estimates contextualize bipolar disorder but should not replace bipolar-specific values.

Suicide burden

Suicide mortality was the most elevated cause-specific ratio in the bipolar meta-analysis: SMR 14.44 (95% CI 12.43–16.78) (Hayes 2015, PMID 25735195). Relative risk is largest where background risk is low; absolute risk still depends on age, sex, prior behavior, episode state and setting.

Early-onset bipolar cohorts reported five-year cumulative suicide-attempt prevalence of 18–20% (Cirone 2021, PMID 33800274). In South Asian samples combining psychosis and bipolar disorder, pooled attempt prevalence was 22% (95% CI 17–27) and ideation prevalence 38% (27–51), but diagnoses were combined and only one study reported suicide deaths (Khoso 2023, PMID 37814419).

An umbrella review of pediatric bipolar-spectrum meta-analyses, read from the abstract only, estimated that 26.2% (95% CI 24.4–28.0) of children and adolescents had attempted suicide at least once, with higher odds than in depressed youth (OR 1.59, 1.24–2.05) and than in adults with bipolar-spectrum disorder (OR 1.68, 1.29–2.18). It pools 52 meta-analyses (k=754, N=59,967, mean age 13.0, 44.8% female). That is a synthesis of existing reviews, not a new cohort, and the retrieved abstract does not grade those source reviews (Torres-Cortes 2026, PMID 42901238).

Risk marker Quantitative association Interpretation
DSM-5 mixed features during mania/hypomania Attempt history OR 2.37 (95% CI 1.42–3.94) Mixed presentations warrant direct risk assessment
Childhood maltreatment Attempt OR 2.25 (1.88–2.70) Long-term vulnerability marker, not deterministic
Cannabis-use disorder Attempt-history OR 1.35; P=0.01 Weak, cross-sectional association
Substance-use disorder Prior suicidality OR 1.758 (1.156–2.674) Bidirectional relationship is plausible
Early-onset cohorts 18–20% attempted within five years High-risk developmental group

Mixed features also associated with anxiety (OR 2.67) and rapid cycling (OR 4.23), which may cluster risk rather than act independently (Bartoli 2020, PMID 32697704).

Childhood maltreatment was associated with attempt history, greater depressive and manic severity, psychosis, early onset and rapid cycling across 30 publications (Agnew-Blais 2016, PMID 26873185). Trauma-informed assessment should therefore include current risk without assuming a single causal pathway.

Cannabis-use disorder’s OR 1.35 came from 11 cross-sectional studies with 6,375 participants; lack of suitable longitudinal evidence prevents causal inference (Bartoli 2019, PMID 31121199). Severity-stratified NESARC-III estimates are compatible with that association: lifetime suicide-attempt prevalence was 33.71% (95% CI 21.64–48.36) among people with bipolar I and severe cannabis-use disorder versus 16.65% (13.17–20.83) among people with severe cannabis-use disorder without bipolar I (Uelman 2026, PMID 42664567). These are adjusted cross-sectional rates, not a longitudinal switch or attempt-incidence effect.

Lithium and suicide: evidence and uncertainty

Lithium is often described as antisuicidal, but study design matters. An updated meta-analysis found 15 randomized trials; compared with placebo, suicide attempts were numerically fewer with lithium (25 versus 63; OR 0.73, 95% CI 0.41–1.31) and completed suicides were 4 versus 13 (OR 0.61, 0.25–1.48), neither statistically significant (Wang 2025, PMID 40441661).

The wide intervals remain compatible with clinically important benefit and with substantially smaller effects. Rare events, heterogeneous diagnoses, subtherapeutic levels and inadequate treatment fidelity reduced power (Wang 2025, PMID 40441661).

Observational evidence may capture long-duration mood stabilization and patient selection but is vulnerable to confounding. Three 2026 reports show both how much observational evidence there now is and where its limits lie:

Study Design Result Caveat
Wei 2026 (PMID 42765379) Hong Kong EHR cohort, 1,379 lithium vs 5,556 valproate monotherapy initiators (IPTW), plus pairwise and network meta-analysis (review identified 3 RCTs and 12 observational studies) Cohort HR 0.76 (95% CI 0.54–1.08) for lithium vs valproate; pooled observational HR 0.87 (0.69–1.09). In the network, lithium HR 0.46 (0.33–0.64) and valproate HR 0.63 (0.45–0.88) vs no treatment, with no significant difference between them (HR 0.73, 0.52–1.02) The RCTs could not be pooled. The "vs no treatment" contrast is observational and inherits confounding by indication
Xu 2026 (PMID 42671279) Retrospective cohort, 15 Chinese psychiatric centres, ages 12–45; 290 lithium vs 295 unexposed 1-year suicidal ideation aOR 0.57 (0.38–0.86); suicidal behaviours aOR 0.60 (0.39–0.90); no difference in non-suicidal self-injury Exposure defined as lithium for ≥80% of 6 months; unexposed controls, not an active comparator
Guan 2026 (PMID 42713736) Cross-sectional, 14 Chinese centres, ages 12–45; 585 analysed Composite suicide-related behaviours OR 0.55 (0.34–0.89) adjusted; 0.53 (0.33–0.86) after propensity matching; aggression statistically accounted for ~21% of the association Cross-sectional; the authors flag residual confounding by indication

Xu and Guan both analyse 585 participants aged 12–45 recruited from 14–15 Chinese centres. They should be read as two analyses of what appears to be the same multicentre sample, not as independent replications. The head-to-head estimate is the one that addresses a real clinical choice. Lithium was numerically but not significantly favoured over valproate (Wei 2026, PMID 42765379), which is consistent with the imprecise randomized estimates above.

The defensible synthesis is that lithium remains important for bipolar maintenance, with a plausible suicide benefit that randomized evidence has not estimated precisely.

Cause-specific natural mortality

Cause Bipolar SMR 95% CI Source
Circulatory disease 1.73 1.54–1.94 Hayes 2015
Respiratory disease 2.92 2.00–4.23 Hayes 2015
Infection 2.25 1.70–3.00 Hayes 2015
Neoplasm 1.14 1.10–1.21 Hayes 2015
Other violent death 3.68 2.77–4.90 Hayes 2015

Respiratory and infectious ratios exceed the circulatory ratio, but cardiovascular disease produces a large absolute burden because it is common. Relative inequality and excess deaths should both inform priorities.

A five-country European cohort of severe mental illness found cardiovascular disease produced the largest absolute excess death rate in bipolar disorder: 8.36 per 10,000 person-years (95% CI 2.96–13.75). Relative bipolar SMRs were especially high for respiratory disease 2.72 (2.04–3.63), endocrine/metabolic disease 2.60 (2.26–2.98), and gastrointestinal disease 2.34 (2.00–2.74) (Gustafsson 2026, PMID 42480565).

Cardiovascular and cerebrovascular disease

A stroke meta-analysis of seven studies and 13,305,007 participants found incidence HR 1.43 (95% CI 1.24–1.66) and stroke-mortality HR 1.54 (1.09–2.18) in bipolar disorder (Yuan 2022, PMID 34052937). Incident heart failure is also elevated: a synthesis of adjusted cohort estimates reported RR 1.95 (95% CI 1.31–2.91; I²=80.1%) for bipolar disorder, higher than the schizophrenia estimate in the same review (RR 1.50, 1.35–1.67), with heterogeneity that limits precision (Dong 2026, PMID 42668911). Seven retrospective cohorts met the review's inclusion criteria overall, but only four contributed the bipolar estimate (McDermott 2005, Chen 2022, Yoo 2023, Lee 2024); leave-one-out estimates ranged 1.73–2.40, and the bipolar and schizophrenia confidence intervals overlap, so the ordering between the two diagnoses is not established by a formal between-group test.

The larger severe-mental-illness cardiovascular review included at least 1.84 million cases and 29.32 million controls in mortality analyses. For bipolar disorder, pooled cerebrovascular mortality SMR was 1.55 (95% CI 1.33–1.81); coronary heart disease incidence was also elevated (Lambert 2022, PMID 35439243).

A 32-study observational meta-analysis, read from the abstract only, associated bipolar disorder with higher cardiovascular incidence in its controlled studies (IRR 1.66, 95% CI 1.32–2.08). The abstract also said coronary and cerebrovascular incidence estimates were elevated; those numeric ratios were not in the retrieved text, so they are not quoted here. Pooled prevalence was 7.70% (5.07–11.52) for any cardiovascular disease, 6.89% (4.05–11.47) for coronary disease, 4.45% (2.13–9.05) for cerebrovascular disease and 3.30% (1.66–6.45) for peripheral arterial disease. The authors describe that prevalence pool as extremely heterogeneous and as a broad indicator of burden, not a precise rate for one bipolar population (Motta 2026, PMID 42902526). The medication and lifestyle sentence was truncated in the retrieved abstract and is omitted.

Hypertension incidence was elevated in bipolar disorder at incidence-rate ratio 1.27 (95% CI 1.15–1.40) (Ayerbe 2018, PMID 28915505). The same review found lower screening, prescribing and adherence in some bipolar and schizophrenia cohorts.

Inequity after cardiovascular disease

Across 47 studies and 24.4 million patients with mental disorders, odds of any cardiovascular screening or treatment were 0.773 (95% CI 0.742–0.804) versus people without mental disorders. Estimates were 0.734 for coronary disease and 0.810 for cerebrovascular disease (Solmi 2021, PMID 34256605).

After acute coronary syndrome, severe mental illness was associated with mortality RR 1.40 (95% CI 1.21–1.62), one-year mortality RR 1.68, revascularization OR 0.57 and cardioprotective-medication receipt RR 0.89 (Chan 2022, PMID 35786737). Both schizophrenia and bipolar groups had lower revascularization; bipolar mortality elevation was clearest in incident-ACS cohorts.

Point in pathway Documented disparity Potential intervention target
Risk-factor detection Lower screening in mental-disorder cohorts Integrated blood pressure, lipid and diabetes monitoring
Acute coronary care Revascularization OR 0.57 in SMI Standardized pathways and diagnostic parity
Secondary prevention Cardioprotective medication RR 0.89 Medication reconciliation and follow-up
Long-term outcomes One-year post-ACS mortality RR 1.68 Cardiology–psychiatry shared care

These estimates are not proof that healthcare disparity explains the whole mortality gap; smoking, poverty, illness severity, medication effects and competing comorbidity remain relevant.

Obesity and metabolic syndrome

Nine cross-sectional studies with 12,259 bipolar cases and 615,490 controls found obesity associated with bipolar disorder at OR 1.77 (95% CI 1.40–2.23; I²=82.1%) (Zhao 2016, PMID 27262632). Cross-sectional evidence cannot establish direction.

A later bidirectional meta-analysis estimated obesity-to-bipolar association OR 1.32 (95% CI 1.01–1.62) and bipolar-to-obesity OR 1.68 (1.35–2.00), but residual confounding and medication exposure remain important (Kambey 2023, PMID 36806758).

The most recent global metabolic-syndrome meta-analysis included 56 studies and 780,099 participants; pooled prevalence was 32.7% (95% CI 25.9–40.2), with I²=99.3% (Salari 2025, PMID 40921519). A Chinese inpatient meta-analysis estimated 33% (29–37%), compared with 10.82% in controls (Tao 2022, PMID 35738391).

Metabolic outcome Estimate Evidence caveat
Global metabolic-syndrome prevalence 32.7% (95% CI 25.9–40.2) Extreme heterogeneity
Chinese inpatient prevalence 33% (29–37) Inpatient and country-specific
Obesity association OR 1.77 (1.40–2.23) Cross-sectional
Hypertension incidence IRR 1.27 (1.15–1.40) Few prospective studies

Insulin resistance and illness course

In a 10-report systematic review with 1,183 participants, impaired glucose metabolism was associated with chronic bipolar course (OR 2.96, 95% CI 1.69–5.17), rapid cycling (OR 2.88, 1.59–5.21), and poor mood-stabilizer response (OR 6.74, 1.04–43.54) (Miola 2023, PMID 37086806).

These small, largely cross-sectional studies cannot determine whether insulin resistance worsens mood course, severe illness worsens metabolism, or both share treatment and social determinants.

Appetite-regulating hormones are not a validated mechanistic bridge. Leptin did not differ significantly from controls during mania, depression or euthymia in an 11-study meta-analysis (Fernandes 2016, PMID 27065008).

Medication contributions

Treatment can reduce episode-related mortality while increasing particular physical risks. This is a comparative, not binary, problem.

Exposure Physical-health concern Evidence boundary
Second-generation antipsychotics Weight, glucose and lipid changes differ by drug Long-term cardiovascular-event RCT data are sparse
Lithium Renal, thyroid and toxicity risks Monitoring can detect many harms; discontinuation can destabilize illness
Valproate Weight/metabolic and reproductive endocrine effects Especially important in women of reproductive potential
Carbamazepine Interactions, hyponatremia and hematologic risk Requires drug-specific monitoring

In women with bipolar disorder, valproate was associated with PCOS (OR 6.74, 95% CI 1.66–27.32), menstrual disorder (OR 1.81, 1.02–3.23) and hyperandrogenism (OR 2.02, 1.11–3.65) (Zhang 2016, PMID 27160812).

Long-term antipsychotic RCT evidence was thin: among 92 trials lasting at least 52 weeks, myocardial infarction was reported in only 11, stroke in six and cardiovascular death in 24 (Rotella 2020, PMID 31917068). Absence of detected differences is not proof of long-term cardiovascular equivalence.

Integrated prevention framework

  1. Assess suicide risk directly during every acute mood episode, especially mixed, agitated, psychotic and post-discharge states.
  2. Treat prior attempts, current intent, access to lethal means and rapidly changing agitation as separate data, not a single score.
  3. Screen tobacco, alcohol and other substance use repeatedly.
  4. Record weight/BMI, waist measure where used, blood pressure, glucose or HbA1c, and lipids at clinically appropriate intervals.
  5. Ensure ordinary cardiovascular and cancer prevention is not displaced by psychiatric care.
  6. Reconcile psychiatric and cardiometabolic medicines after every transition.
  7. Address sleep, activity, food security, housing and access barriers alongside prescriptions.
  8. Investigate renal, thyroid and reproductive effects specific to the chosen mood stabilizer.

Behavior, exposure and care are separable pathways

Across 69 severe-mental-illness studies (n=35,682), waking sedentary time averaged 476 minutes/day (95% CI 407–545) and moderate/vigorous activity 38.4 minutes/day (32.0–44.8). People with severe mental illness were less likely to meet activity guidance (OR 1.5 for nonadherence, 95% CI 1.1–2.0; I²=95.8%); bipolar participants were the most active diagnostic group yet accumulated the most sedentary time (Vancampfort 2017, PMID 28941119). Activity and sedentary time are not interchangeable targets. Fitness itself is lower. A 2026 meta-analysis estimated adult cardiorespiratory fitness at 27.0 mL·kg⁻¹·min⁻¹ (k=8, n=190; I²=97%) and 7.4 mL·kg⁻¹·min⁻¹ below controls (95% CI −12.7 to −2.1; k=7, n=333). It warned that the between-group estimate was not robust to excluding one study, and that only two uncontrolled pilots measured fitness before and after an exercise intervention, so no conclusion about exercise efficacy is possible (Arida 2026, PMID 42702273).

That fitness limit is about measured cardiorespiratory fitness. It is separate from whether activity changes mood. The ISBD Nutrition and Exercise Task Force position statement searched PsycINFO, SPORTDiscus, Embase, PubMed and Web of Science through 1 March 2026 and concluded that, despite limited high-quality evidence, physical activity and exercise show promising physical and mental health benefits in bipolar disorder, and that multicenter randomized trials are the next step (Lafer 2026, PMID 42545207). The statement was PMC-live on 4 August 2026, before this sweep window, and is backfilled here. It supplements bipolar studies with evidence from depression and the general population. It does not change the finding that only two uncontrolled pilots have measured fitness before and after exercise.

Muscle quantity is a further, separate outcome. In South-Eastern Australia, 106 women with bipolar disorder recruited from healthcare settings were compared with 300 age-matched women without bipolar disorder from the Geelong Osteoporosis Study; both groups were aged at least 20, and Table 1 lists the combined sample as 406. Women with bipolar disorder were twice as likely to meet sarcopenia criteria (FNIH OR 2.24, 95% CI 1.11–4.55; EWGSOP2 OR 2.78, 1.47–5.20), independent of age, smoking, activity, alcohol, diet and socioeconomic status. Low handgrip strength was more common (15/106, 14.2%, versus 21/300, 7.1%). The paper’s counts of 36 women (8.9%; 34 probable and 2 confirmed) by FNIH criteria and 46 women (11.4%; 43 probable and 3 confirmed) by EWGSOP2 criteria sit with Table 1’s combined sample of 406. Thirty-six of 406 is 8.9%, and 46 of 406 is 11.3%, which the paper prints as 11.4%. They are not rates among the 106 women with bipolar disorder. Confirmed sarcopenia, defined as low handgrip plus low muscle mass, was those 2 and 3 women, so the association is carried by probable sarcopenia. The authors ask for replication (Williams 2026, PMID 42897866).

Smoking is associated with both illness and suicide outcomes, but direction is unsettled. In 22 psychosis/bipolar studies (N=11,452), smoking was associated with attempt history (OR 1.70, 95% CI 1.48–1.95), with evidence of publication inflation (Pietschnig 2024, PMID 39328818). Separately, 25 cohort studies (2.92 million people) associated current smoking with subsequent bipolar disorder (RR 1.54, 1.22–1.95), but residual confounding and reverse causation prevent a causal conclusion (Hu 2025, PMID 40018351).

Environmental evidence is even less bipolar-specific. A 22-study air-pollution review could meta-analyze depression and suicide exposures but not establish a bipolar-specific effect; long-term PM2.5 was associated with depression (OR 1.102 per 10 μg/m³, 95% CI 1.023–1.189) and short-term PM10 with suicide at a 0–2-day lag (Braithwaite 2019, PMID 31850801). These findings motivate bipolar-stratified cohorts rather than direct risk attribution.

The physical-health gap also includes treatment inequity: a broad review found higher cardiometabolic, respiratory, infectious and other disease burden in severe mental illness together with lower receipt of standard medical care (De Hert 2011, PMID 21379357). Its age and mixed-diagnosis scope limit current bipolar-specific quantification, but it identifies healthcare delivery—not only patient behavior or psychotropic exposure—as a causal candidate.

Open questions

  • How much of the current mortality gap is mediated by healthcare disparities versus exposure and illness factors (Solmi 2021, PMID 34256605)?
  • Which suicide-prevention combinations reduce deaths rather than only ideation or attempts (Wang 2025, PMID 40441661)? Is lithium's antisuicidal advantage over valproate real but small, or absent? The 2026 head-to-head network estimate (HR 0.73, 0.52–1.02) cannot yet distinguish the two (Wei 2026, PMID 42765379).
  • Does active treatment of insulin resistance improve mood course in randomized bipolar samples (Miola 2023, PMID 37086806)?
  • Why have cardiovascular mortality inequalities not clearly narrowed in recent decades (Lambert 2022, PMID 35439243)?
  • Which antipsychotic strategies minimize long-term cardiovascular events while maintaining relapse prevention (Rotella 2020, PMID 31917068)?

References

  1. Hayes JF, et al. A systematic review and meta-analysis of premature mortality in bipolar affective disorder. Acta Psychiatrica Scandinavica. 2015. PMID 25735195.
  2. Chan JKN, et al. Life expectancy and years of potential life lost in bipolar disorder: systematic review and meta-analysis. British Journal of Psychiatry. 2022. PMID 35184778.
  3. Walker ER, et al. Mortality in mental disorders and global disease burden implications. JAMA Psychiatry. 2015. PMID 25671328.
  4. Cirone C, et al. What Do We Know about the Long-Term Course of Early Onset Bipolar Disorder? Brain Sciences. 2021. PMID 33800274.
  5. Khoso AB, et al. Prevalence of suicidal ideation and suicide attempts in individuals with psychosis and bipolar disorder in South Asia. BJPsych Open. 2023. PMID 37814419.
  6. Bartoli F, et al. Clinical correlates of DSM-5 mixed features in bipolar disorder. Journal of Affective Disorders. 2020. PMID 32697704.
  7. Agnew-Blais J, et al. Childhood maltreatment and unfavourable clinical outcomes in bipolar disorder. Lancet Psychiatry. 2016. PMID 26873185.
  8. Bartoli F, et al. Cannabis use disorder and suicide attempts in bipolar disorder. Neuroscience & Biobehavioral Reviews. 2019. PMID 31121199.
  9. Wang JX, et al. The efficacy of lithium in the treatment of suicidal ideation, behavior and suicide. Journal of Affective Disorders. 2025. PMID 40441661.
  10. Gustafsson TT, et al. Cause-specific mortality due to physical illness in severe mental disorders in Europe. Lancet Public Health. 2026. PMID 42480565.
  11. Yuan M, et al. Bipolar disorder and the risk for stroke incidence and mortality: a meta-analysis. Neurological Sciences. 2022. PMID 34052937.
  12. Lambert AM, et al. Temporal trends in associations between severe mental illness and risk of cardiovascular disease. PLoS Medicine. 2022. PMID 35439243.
  13. Ayerbe L, et al. Hypertension risk and clinical care in patients with bipolar disorder or schizophrenia. Journal of Affective Disorders. 2018. PMID 28915505.
  14. Solmi M, et al. Disparities in Screening and Treatment of Cardiovascular Diseases in Patients With Mental Disorders Across the World. American Journal of Psychiatry. 2021. PMID 34256605.
  15. Chan JKN, et al. Mortality, Revascularization, and Cardioprotective Pharmacotherapy After Acute Coronary Syndrome in Patients With Severe Mental Illness. Schizophrenia Bulletin. 2022. PMID 35786737.
  16. Zhao Z, et al. The potential association between obesity and bipolar disorder: A meta-analysis. Journal of Affective Disorders. 2016. PMID 27262632.
  17. Kambey PA, et al. The bi-directional association between bipolar disorder and obesity. International Journal of Obesity. 2023. PMID 36806758.
  18. Salari N, et al. Global prevalence of metabolic syndrome among individuals with bipolar disorder. Archives of Psychiatric Nursing. 2025. PMID 40921519.
  19. Tao H, et al. A Systematic Review and Meta-Analysis of Metabolic Syndrome Prevalence in Chinese Inpatients with Bipolar Disorder. Hormone and Metabolic Research. 2022. PMID 35738391.
  20. Miola A, et al. Insulin resistance in bipolar disorder. Journal of Affective Disorders. 2023. PMID 37086806.
  21. Fernandes BS, et al. Leptin in bipolar disorder. European Psychiatry. 2016. PMID 27065008.
  22. Zhang L, et al. Reproductive and metabolic abnormalities in women taking valproate for bipolar disorder. European Journal of Obstetrics & Gynecology and Reproductive Biology. 2016. PMID 27160812.
  23. Rotella F, et al. Long-term metabolic and cardiovascular effects of antipsychotic drugs. European Neuropsychopharmacology. 2020. PMID 31917068.
  24. Uelman C, et al. Examination of the intersection of bipolar I disorder status and cannabis use disorder severity in relation to health-related quality-of-life and lifetime suicide attempts. J Psychiatr Res. 2026;202:206-210. PMID 42664567.
  25. Dong Q, Ye X. Risk of heart failure in patients with schizophrenia or bipolar disorder: A systematic review and meta-analysis. Pak J Med Sci. 2026;42:2129-2137. PMID 42668911.
  26. Vancampfort D, et al. Sedentary behavior and physical activity levels in people with schizophrenia, bipolar disorder and major depressive disorder: a global systematic review and meta-analysis. World Psychiatry. 2017;16:308–315. PMID 28941119
  27. Pietschnig J, et al. Smoking behavior is associated with suicidality in individuals with psychosis and bipolar disorder: a systematic quantitative review and meta-analysis. Front Psychol. 2024;15:1369669. PMID 39328818
  28. Hu Z, et al. Association between cigarette smoking and the risk of major psychiatric disorders: a systematic review and meta-analysis in depression, schizophrenia, and bipolar disorder. Front Med (Lausanne). 2025;12:1529191. PMID 40018351
  29. Braithwaite I, et al. Air pollution exposure and associations with depression, anxiety, bipolar disorder, psychosis and suicide risk: a systematic review and meta-analysis. Environ Health Perspect. 2019;127:126002. PMID 31850801
  30. De Hert M, et al. Physical illness in patients with severe mental disorders. I. Prevalence, impact of medications and disparities in health care. World Psychiatry. 2011;10:52–77. PMID 21379357
  31. Wei C, et al. Comparing lithium and valproate in reducing suicide risk: a territory-wide retrospective cohort study with pairwise and network meta-analysis. Psychol Med. 2026;56:e289. PMID 42765379
  32. Xu N, et al. Association of lithium use with suicide and self-harm among Chinese adolescents and adults with bipolar disorder: a nationwide multicenter retrospective cohort study. J Clin Psychiatry. 2026;87(3):25m16219. PMID 42671279
  33. Guan X, et al. Association between lithium use and suicide-related behaviors in bipolar disorder: a multicenter study. Bipolar Disord. 2026;28:e70180. PMID 42713736
  34. Arida VJP, et al. Cardiorespiratory fitness in people with bipolar disorder: a systematic review and meta-analysis. J Affect Disord. 2026;415:122480. PMID 42702273
  35. Torres-Cortes J, et al. Meta-analytic evidence of clinical, cognitive, biological and treatment characteristics of bipolar spectrum disorder in children and adolescents: results from an umbrella review. J Am Acad Child Adolesc Psychiatry. 2026. PMID 42901238.
  36. Motta FDS, et al. Bipolar disorder and the risk of cardiovascular diseases: a systematic review and meta-analysis of observational studies. Acta Psychiatr Scand. 2026. PMID 42902526.
  37. Lafer B, et al. Sedentary behavior and physical activity in individuals with bipolar disorder: an evidence synthesis and position statement from the ISBD Nutrition and Exercise Task Force (NExT). Bipolar Disord. 2026;28(5):e70164. PMID 42545207.
  38. Williams LJ, et al. Women with bipolar disorder are at increased risk of sarcopenia. Bipolar Disord. 2026;28(7):e70192. PMID 42897866.