Bipolar disorder — statistics quick reference¶
Last updated: 2026-10-23
Figures are transcribed from live-retrieved PubMed records, specified full-text tables, or named statistics portals. Estimates that appear to conflict are shown separately because populations, criteria, settings and methods differ.
Prevalence, service use and disability¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Global bipolar-disorder prevalence, 2023 | 36 million; 0.4%; approximately 32 million adults | WHO fact sheet dated 11 Sep 2026, citing GBD 2023; modelled population estimates, not the GBD 2021 series. Re-fetched 2026-10-23; the figures were unchanged | WHO — “Bipolar disorder”, https://www.who.int/news-room/fact-sheets/detail/bipolar-disorder, accessed 2026-10-23 |
| Lifetime bipolar I prevalence | 0.6% | 61,392 adults, 11 countries, WMH-CIDI cross-sectional surveys | Merikangas 2011, PMID 21383262 |
| Lifetime bipolar II prevalence | 0.4% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| Lifetime subthreshold bipolar prevalence | 1.4% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| Lifetime bipolar-spectrum prevalence | 2.4% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| 12-month bipolar I prevalence | 0.4% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| 12-month bipolar II prevalence | 0.3% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| 12-month subthreshold bipolar prevalence | 0.8% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| 12-month bipolar-spectrum prevalence | 1.5% | Same WMH sample/method | Merikangas 2011, PMID 21383262 |
| Severe role impairment during depression | 74.0% | WMH bipolar-spectrum respondents with depression | Merikangas 2011, PMID 21383262 |
| Severe role impairment during mania | 50.9% | WMH bipolar-spectrum respondents with mania | Merikangas 2011, PMID 21383262 |
| Any comorbid disorder | ~75% | WMH lifetime bipolar-spectrum cases | Merikangas 2011, PMID 21383262 |
| Mental-health-system contact, low-income countries | 25.2% | WMH lifetime bipolar-spectrum cases | Merikangas 2011, PMID 21383262 |
| Median bipolar treatment gap | 50.2% untreated | 37 community epidemiology studies | Kohn 2004, PMID 15640922 |
| Australian 12-month prevalence, recalibrated CIDI | 0.9% | 2007 national survey | Mitchell 2013, PMID 22906117 |
| Australian 12-month prevalence, unrecalibrated CIDI | 1.7% | Same survey; older algorithm | Mitchell 2013, PMID 22906117 |
| Singapore lifetime prevalence | 1.2% | Nationally representative CIDI survey | Subramaniam 2013, PMID 23017543 |
| Singapore 12-month prevalence | 0.6% | Same survey | Subramaniam 2013, PMID 23017543 |
| Singapore any other lifetime mental disorder | 69.4% | Lifetime bipolar cases | Subramaniam 2013, PMID 23017543 |
| Singapore chronic physical condition | 52.6% | Lifetime bipolar cases | Subramaniam 2013, PMID 23017543 |
| African community lifetime prevalence range | 0.1–1.83% | Systematic review of Nigeria/Ethiopia and other African evidence | Esan 2016, PMID 26155900 |
| African missed-diagnosis rate | Up to 36.2% | African systematic review | Esan 2016, PMID 26155900 |
| Disability association | OR 3.41 (95% CI 1.04–11.12) | Portugal WMH survey | Antunes 2018, PMID 29366849 |
Diagnosis, spectrum and mixed features¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| DSM-5 mixed features in major depressive episodes | 11.6% (95% CI 7.9–16.7) | 17 studies/20 samples, literature 2013–2020 | Na 2021, PMID 33418368 |
| DSM-5 mixed features in mania/hypomania | 26.8% (17.0–39.5) | Same meta-analysis | Na 2021, PMID 33418368 |
| ≥3 opposite-polarity symptoms, all mood episodes | 27.8% (27.2–28.5) | 17 reports, 19,198 participants | Vázquez 2018, PMID 28922738 |
| ≥3 opposite-polarity symptoms, bipolar depression | 35.2% (33.8–36.5) | Same systematic review | Vázquez 2018, PMID 28922738 |
| ≥3 opposite-polarity symptoms, bipolar mania/hypomania | 35.1% (32.9–37.3) | Same systematic review | Vázquez 2018, PMID 28922738 |
| ≥3 opposite-polarity symptoms, MDD depression | 23.8% (23.0–24.5) | Same systematic review | Vázquez 2018, PMID 28922738 |
| Mixed features and attempt history | OR 2.37 (1.42–3.94) | 8 studies, 3,070 bipolar participants | Bartoli 2020, PMID 32697704 |
| Mixed features and anxiety disorder | OR 2.67 (1.28–5.57) | Manic/hypomanic subgroup | Bartoli 2020, PMID 32697704 |
| Mixed features and rapid cycling | OR 4.23 (1.29–13.81) | Manic/hypomanic subgroup | Bartoli 2020, PMID 32697704 |
| Mixed-feature manic severity | SMD −0.40 (−0.65 to −0.16) | Manic/hypomanic subgroup | Bartoli 2020, PMID 32697704 |
| HCL-32 sensitivity | 81% | Psychiatric services; 53-study accuracy meta-analysis, n=21,542 | Carvalho 2015, PMID 25451435 |
| HCL-32 specificity | 67% | Same | Carvalho 2015, PMID 25451435 |
| MDQ sensitivity | 66% | Psychiatric services | Carvalho 2015, PMID 25451435 |
| MDQ specificity | 79% | Psychiatric services | Carvalho 2015, PMID 25451435 |
| BSDS sensitivity | 69% | Psychiatric services | Carvalho 2015, PMID 25451435 |
| BSDS specificity | 86% | Psychiatric services | Carvalho 2015, PMID 25451435 |
| MDQ sensitivity at cutoff 7 | 43% | Primary care/general population | Carvalho 2015, PMID 25451435 |
| MDQ specificity at cutoff 7 | 95% | Primary care/general population | Carvalho 2015, PMID 25451435 |
Course, recurrence and functioning¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Ten-year sustained remission/improvement | About 50% | 34 bipolar I participants, prospective follow-up | Goldberg 2004, PMID 15306137 |
| Ten-year functional decline | 30–40% | Same cohort | Goldberg 2004, PMID 15306137 |
| Normal work performance at every follow-up | <50% | Same cohort | Goldberg 2004, PMID 15306137 |
| Remained syndromally remitted at 12 months | 87.6% | 398 remitted bipolar I patients, prospective Spanish cohort | Montoya 2010, PMID 20557947 |
| Free of subsyndromal symptoms at 12 months | 79.9% | Same cohort | Montoya 2010, PMID 20557947 |
| Normal functioning at 12 months | 53.5% | Same cohort | Montoya 2010, PMID 20557947 |
| Early-onset diagnostic stability | 73–100% over 10 years | 42 publications, 15 cohorts, 10,187 patients | Cirone 2021, PMID 33800274 |
| Recovery from index early-onset episode | 81.5–100% | Same review | Cirone 2021, PMID 33800274 |
| Early-onset recurrence | 35–67% | Same review | Cirone 2021, PMID 33800274 |
| Early-onset five-year attempt prevalence | 18–20% | Five cohorts in review | Cirone 2021, PMID 33800274 |
| Hypersomnia prevalence | 29.9% (25.8–34.1) | 10 studies, 1,824 bipolar patients | Grigolon 2019, PMID 30611064 |
| Maltreatment and earlier onset | OR 1.85 (1.43–2.40) | Meta-analysis, 14 studies/5,733 participants | Agnew-Blais 2016, PMID 26873185 |
| Maltreatment and rapid cycling | OR 1.89 (1.45–2.48) | 8 studies/3,010 participants | Agnew-Blais 2016, PMID 26873185 |
| Maltreatment and more manic episodes | OR 1.26 (1.09–1.47) | 7 studies/3,909 participants | Agnew-Blais 2016, PMID 26873185 |
| Maltreatment and more depressive episodes | OR 1.38 (1.07–1.79) | 8 studies/4,025 participants | Agnew-Blais 2016, PMID 26873185 |
Psychiatric comorbidity¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| ADHD among adults with bipolar disorder | 17.11% (13.05–21.59) | 71 studies, 646,766 participants | Schiweck 2021, PMID 33515607 |
| Bipolar disorder among adults with ADHD | 7.95% (5.31–11.06) | Same meta-analysis | Schiweck 2021, PMID 33515607 |
| Earlier bipolar onset with comorbid ADHD | 3.96 years (2.65–5.26) | Same meta-analysis | Schiweck 2021, PMID 33515607 |
| Bipolar odds in adults with versus without ADHD | OR 8.7 (5.47–13.89) | Large studies: ADHD n=550,748; controls n=14,546,814 | Hartman 2023, PMID 37149075 |
| Any anxiety disorder in pediatric bipolar disorder | 44.7% | 37-study meta-analysis | Yapıcı Eser 2020, PMID 31899546 |
| Pediatric panic disorder | 12.7% | Same | Yapıcı Eser 2020, PMID 31899546 |
| Pediatric generalized anxiety disorder | 27.4% | Same | Yapıcı Eser 2020, PMID 31899546 |
| Pediatric social phobia | 20.1% | Same | Yapıcı Eser 2020, PMID 31899546 |
| Pediatric separation anxiety | 26.1% | Same | Yapıcı Eser 2020, PMID 31899546 |
| Pediatric OCD | 16.7% | Same | Yapıcı Eser 2020, PMID 31899546 |
| Pediatric ADHD, weighted mean | 48% | 167-study systematic review | Frías 2015, PMID 25545605 |
| Pediatric substance-use disorder | 31% | Same review | Frías 2015, PMID 25545605 |
| SUD association with male sex | OR 2.191 (1.121–4.281) | Adult bipolar systematic review/meta-analysis | Messer 2017, PMID 28419959 |
| SUD association with prior suicidality | OR 1.758 (1.156–2.674) | Same | Messer 2017, PMID 28419959 |
| Cannabis-use disorder and attempt history | OR 1.35; P=0.01 | 11 studies, 6,375 participants; cross-sectional | Bartoli 2019, PMID 31121199 |
Mortality, suicide and cardiovascular health¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| All-cause mortality SMR | 2.05 (1.89–2.23) | 31-study bipolar meta-analysis | Hayes 2015, PMID 25735195 |
| Natural-cause SMR | 1.64 (1.47–1.83) | Same | Hayes 2015, PMID 25735195 |
| Unnatural-cause SMR | 7.42 (6.43–8.55) | Same | Hayes 2015, PMID 25735195 |
| Suicide SMR | 14.44 (12.43–16.78) | Same | Hayes 2015, PMID 25735195 |
| Circulatory mortality SMR | 1.73 (1.54–1.94) | Same | Hayes 2015, PMID 25735195 |
| Respiratory mortality SMR | 2.92 (2.00–4.23) | Same | Hayes 2015, PMID 25735195 |
| Infection mortality SMR | 2.25 (1.70–3.00) | Same | Hayes 2015, PMID 25735195 |
| Neoplasm mortality SMR | 1.14 (1.10–1.21) | Same | Hayes 2015, PMID 25735195 |
| Pooled life expectancy | 66.88 years (64.47–69.28) | 11 studies, 96,601 people | Chan 2022, PMID 35184778 |
| Female life expectancy | 70.51 years (68.61–72.41) | Subgroup meta-analysis | Chan 2022, PMID 35184778 |
| Male life expectancy | 64.59 years (61.16–68.03) | Subgroup meta-analysis | Chan 2022, PMID 35184778 |
| Years of potential life lost | 12.89 (12.72–13.07) | 13 studies, 128,989 people | Chan 2022, PMID 35184778 |
| Stroke incidence HR | 1.43 (1.24–1.66) | 7 studies, 13,305,007 participants | Yuan 2022, PMID 34052937 |
| Stroke mortality HR | 1.54 (1.09–2.18) | Same | Yuan 2022, PMID 34052937 |
| Hypertension incidence-rate ratio | 1.27 (1.15–1.40) | Prospective-study meta-analysis | Ayerbe 2018, PMID 28915505 |
| Post-ACS mortality RR, SMI | 1.40 (1.21–1.62) | 22 studies, 12,235,501 participants | Chan 2022, PMID 35786737 |
| Post-ACS revascularization OR, SMI | 0.57 (0.49–0.67) | Same | Chan 2022, PMID 35786737 |
| Cardioprotective medication RR, SMI | 0.89 (0.85–0.94) | Same | Chan 2022, PMID 35786737 |
| Incident heart failure RR | 1.95 (1.31–2.91); I²=80.1% | 4 retrospective cohorts contributing the bipolar estimate (7 cohorts in the review overall); adjusted estimates | Dong 2026, PMID 42668911 |
| Lifetime suicide attempts, bipolar I with severe cannabis-use disorder | 33.71% (21.64–48.36) | NESARC-III; n=753 bipolar I / 35,556 without bipolar I; adjusted | Uelman 2026, PMID 42664567 |
| Lifetime suicide attempts, severe cannabis-use disorder without bipolar I | 16.65% (13.17–20.83) | Same NESARC-III comparison | Uelman 2026, PMID 42664567 |
Metabolic and reproductive health¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Obesity association | OR 1.77 (1.40–2.23) | 9 cross-sectional studies; 12,259 cases/615,490 controls | Zhao 2016, PMID 27262632 |
| Global metabolic-syndrome prevalence | 32.7% (25.9–40.2) | 56 studies, 780,099 participants | Salari 2025, PMID 40921519 |
| Chinese inpatient metabolic-syndrome prevalence | 33% (29–37) | 11 studies, 1,562 participants | Tao 2022, PMID 35738391 |
| Impaired glucose metabolism and chronic course | OR 2.96 (1.69–5.17) | Fixed-effect synthesis, k=2 | Miola 2023, PMID 37086806 |
| Impaired glucose metabolism and rapid cycling | OR 2.88 (1.59–5.21) | Fixed-effect synthesis, k=2 | Miola 2023, PMID 37086806 |
| Impaired glucose metabolism and poor stabilizer response | OR 6.74 (1.04–43.54) | Fixed-effect synthesis, k=2 | Miola 2023, PMID 37086806 |
| Valproate and PCOS | OR 6.74 (1.66–27.32) | Women with bipolar disorder, meta-analysis | Zhang 2016, PMID 27160812 |
| Valproate and menstrual disorder | OR 1.81 (1.02–3.23) | Same | Zhang 2016, PMID 27160812 |
| Valproate and hyperandrogenism | OR 2.02 (1.11–3.65) | Same | Zhang 2016, PMID 27160812 |
Pregnancy and postpartum¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Reported effective contraception | 487/796 (61.2%) | French MONTRA care sample, women aged 18–50; data collected 2010–2011, published 2026; Table 1, full text. Definition includes condoms/diaphragm; not contemporary coverage | Rodríguez-Toscano 2026, PMID 42850618 |
| Pregnancy recurrence | 71% | 89 euthymic pregnant women, prospective cohort | Viguera 2007, PMID 18056236 |
| Pregnancy recurrences depressive/mixed | 74% | Same cohort | Viguera 2007, PMID 18056236 |
| Pregnancy recurrences in first trimester | 47% | Same cohort | Viguera 2007, PMID 18056236 |
| Recurrence with lamotrigine continuation | 30% | 26 pregnant women | Newport 2008, PMID 18402631 |
| Recurrence after stopping all mood stabilizers | 100% | Same small cohort | Newport 2008, PMID 18402631 |
| Overall postpartum relapse | 35% (29–41) | 37 articles, 5,700 deliveries | Wesseloo 2016, PMID 26514657 |
| Postpartum relapse, medication-free pregnancy | 66% (57–75) | Bipolar subgroup | Wesseloo 2016, PMID 26514657 |
| Postpartum relapse, prophylaxis | 23% (14–37) | Bipolar subgroup | Wesseloo 2016, PMID 26514657 |
| Bipolar I postpartum relapse | 39% (29–49) | 12 studies, 3,595 deliveries | Sharma 2024, PMID 38258551 |
| Manic/mixed among postpartum relapses | 38% (28–50) | Same | Sharma 2024, PMID 38258551 |
| Lithium-exposed cardiac malformations | 2.41% | 663 first-trimester exposures, Medicaid 2000–2010 | Patorno 2017, PMID 28591541 |
| Unexposed cardiac malformations | 1.15% | 1,322,955 unexposed pregnancies | Patorno 2017, PMID 28591541 |
| Lithium cardiac-malformation adjusted RR | 1.65 (1.02–2.68) | Same cohort | Patorno 2017, PMID 28591541 |
| Lithium >900 mg/day cardiac RR | 3.22 (1.47–7.02) | Dose subgroup | Patorno 2017, PMID 28591541 |
| Any lithium-associated congenital anomaly | OR 1.81 (1.35–2.41); NNH 33 | 29-study systematic review; 13 quantitative | Fornaro 2020, PMID 31623458 |
| Lithium-associated cardiac anomaly | OR 1.86 (1.16–2.96); NNH 71 | Same | Fornaro 2020, PMID 31623458 |
| Valproate-exposed autism by age eight | 2.7%; aHR 2.4 (1.7–3.3) | Nordic epilepsy cohort | Bjørk 2022, PMID 35639399 |
| Valproate-exposed intellectual disability | 2.4%; aHR 2.5 (1.7–3.7) | Same | Bjørk 2022, PMID 35639399 |
| Induction/planned caesarean, treated bipolar | 37.5%; OR 2.12 (1.68–2.67) | Swedish registry, 320 treated women | Bodén 2012, PMID 23137820 |
| Induction/planned caesarean, untreated bipolar | 30.9%; OR 1.57 (1.30–1.90) | Swedish registry, 554 untreated women | Bodén 2012, PMID 23137820 |
| Gestational diabetes with bipolar diagnosis | aOR 1.75 (1.15–2.70) | Hong Kong 2003–2018 cohort | Chan 2024, PMID 38914040 |
| Pregnant bipolar cohort receiving psychotropics | 59.6% | Hong Kong, 302 women, 2003–2018 | Kan 2022, PMID 35151952 |
| Polypharmacy during pregnancy | 23.3% | Same cohort | Kan 2022, PMID 35151952 |
Evidence-broadening additions¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Mean specialist-care diagnostic delay | 6.46 years | Hungarian administrative cohort, n=8,935 | Lublóy 2020, PMID 32075625 |
| Pediatric bipolar symptom-index separation | g 1.30 (95% CI 0.98–1.62) | 28 studies, 115 effect sizes; accuracy moderated by informant/content/comparator | Alcaíno 2024, PMID 39103165 |
| Pediatric bipolar psychotic-feature prevalence | 16%–87.5% | Review; range driven by age, sampling and interview method | Pavuluri 2004, PMID 15094254 |
| Twelve-month mood-episode relapse | 46% | Prospective actigraphy cohort, 88/189 | Esaki 2021, PMID 34645802 |
| Pooled recurrence across heterogeneous studies | 43.6% | 34 studies, 33,500 participants | Bai 2026, PMID 41176250 |
| Sleep disturbance and recurrence | OR 3.69 | Same synthesis; nonrandomized prognostic association | Bai 2026, PMID 41176250 |
| Maintenance monotherapy versus placebo | OR 0.42 (0.34–0.51) | 22 enriched RCTs, n=7,773; 24–104 weeks | Nestsiarovich 2022, PMID 34489127 |
| Valproate acute-mania response | RR 1.42 (1.19–1.71) | High-quality review within overview of 26 systematic reviews | Mari 2024, PMID 39500601 |
| Youth risperidone manic-symptom effect | SMD −1.18 (−1.45 to −0.92) | Network meta-analysis, 18 RCTs/2,844 participants overall | Vita 2025, PMID 39128561 |
| Continued adjunctive antidepressant: any episode | HR 0.68 (0.43–1.10) | 52-week versus 8-week continuation after bipolar I depression remission | Yatham 2023, PMID 37530824 |
| Continued adjunctive antidepressant: depressive episode | HR 0.43 (0.25–0.75) | Same prematurely stopped RCT | Yatham 2023, PMID 37530824 |
| Continued adjunctive antidepressant: mania/hypomania | HR 2.28 (0.86–6.08) | Same prematurely stopped RCT | Yatham 2023, PMID 37530824 |
| Lithium-associated mean eGFR difference | −11.14 mL/min/1.73 m² (−16.61 to −5.68) | Lithium versus non-lithium meta-analysis, n=1,622; I²=86% | Macaron 2026, PMID 41727809 |
| Lithium-associated annual eGFR decline difference | 0.13 mL/min/1.73 m²/year (0.06–0.20) faster | Meta-analysis, n=13,280 | Macaron 2026, PMID 41727809 |
| Group psychoeducation and relapse | OR 0.43 (0.28–0.62) | 11 group-therapy RCTs, n=1,191; I²=41% | Tan 2022, PMID 36421612 |
| OCD prevalence in bipolar disorder | 10.9% lifetime (7.8–14.4); 11.2% cross-sectional (7.6–15.3) | 39 lifetime and 29 cross-sectional studies | Ferentinos 2020, PMID 31818777 |
| Waking sedentary time in severe mental illness | 476 min/day (407–545) | 69 studies, n=35,682; mixed severe-mental-illness diagnoses | Vancampfort 2017, PMID 28941119 |
| Smoking and suicide-attempt history | OR 1.70 (1.48–1.95) | 22 psychosis/bipolar studies, n=11,452; publication inflation detected | Pietschnig 2024, PMID 39328818 |
| Infant serum lithium | Mean 0.23 mEq/L | 13 case reports/series, 39 dyads | Imaz 2019, PMID 31551795 |
| Passive relapse-prediction AUC range | 0.70–0.88 | 52 studies, 4,814 participants; 75% high risk of bias, mostly internal validation | Fang 2026, PMID 42106724 |
Periodic sweep additions — 2026-10-09¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Global bipolar prevalent cases, ages 15–49 | 25.28 million (2021; 16.74 million in 1990) | GBD 2021 modelled estimates, ages 15–49 only | Liang 2026, PMID 42688390 |
| Global bipolar incident cases, ages 15–49 | 1.63 million (2021) | Same GBD 2021 analysis; incidence rate changed little since 1990 | Liang 2026, PMID 42688390 |
| Global bipolar DALYs, ages 15–49 | 5.50 million (2021; 3.65 million in 1990) | Same; increase mainly attributed to population size | Liang 2026, PMID 42688390 |
| Share of psychotic or bipolar diagnoses in former CAMHS attendees | 58.3% (combined psychosis/bipolar outcome) | Danish births 1990–2020, n=1,978,630, followed to 2021 | Gregersen 2026, PMID 42753132 |
| Incident bipolar disorder after cannabis exposure | RR 2.46 (1.57–3.84), unadjusted | 8 longitudinal studies, N=206,572 | Sanjuan-Ortiz 2026, PMID 42838231 |
| Cardiorespiratory fitness deficit vs controls | −7.4 mL·kg⁻¹·min⁻¹ (−12.7 to −2.1) | 7 studies, n=333; I²=89.9%; not robust to excluding one study | Arida 2026, PMID 42702273 |
| Adjunctive IV ketamine vs midazolam, treatment-resistant bipolar depression | MADRS −7.3 (−12.0 to −2.5); d 0.7 | Ket-BD RCT, 68 randomized / 63 analysed, day 14 | Orsini 2026, PMID 42684694 |
| LAI vs oral antipsychotic periods: psychiatric hospitalization | aIRR 0.67 (0.61–0.74) | Hong Kong self-controlled case series, n=2,086 | Deng 2026, PMID 42711749 |
| LAI vs oral: depressive-episode hospitalization | aIRR 1.34 (0.94–1.93) | Same | Deng 2026, PMID 42711749 |
| Two-year psychiatric hospitalization/ED risk after first psychotic mania | 42.3% mood stabilizer; 49.5% SGA; 42.2% combination | BD-CAUSAL target-trial emulation, n=371 | Szmulewicz 2026, PMID 42716056 |
| Lithium+AAP vs valproate+AAP: mood-episode rehospitalization after first mania | HR 1.18 (1.03–1.35) | Taiwan claims, propensity-matched n=3,909 | Li 2026, PMID 42765449 |
| Lithium vs valproate: suicide (network estimate) | HR 0.73 (0.52–1.02) | Network meta-analysis; review identified 3 RCTs and 12 observational studies, RCTs not poolable | Wei 2026, PMID 42765379 |
| ≥25% eGFR decline on long-term lithium | 13.7% (87/636) | Single-centre retrospective cohort, ≥1 year lithium, 2010–2022 | İnanlı 2026, PMID 42810282 |
| SILENT after lithium poisoning | 7.6% (49/644 episodes) | 23 hospitals in Spain, Ireland and Switzerland, 2015–2022 | Pineda-Torcuato 2026, PMID 42703744 |
| EDIT-B bipolar-vs-MDD discrimination | AUC 0.873 (0.837–0.909); sens 82.6%, spec 80.3% | External double-blind validation, 4 European centres, n=393 (39% bipolar) | Weissmann 2026, PMID 42833865 |
| Valproate discontinuation maintained at 1 year, women of child-bearing potential | 47.9% (France SNDS, n=9,943); 51.4% (UK CPRD, n=144) | Claims-database post-authorization safety study, 2014–2017 | Colas 2026, PMID 42700484 |
Periodic sweep additions — 2026-10-23¶
| Measure | Estimate | Population / method / data year | Source |
|---|---|---|---|
| Cardiovascular disease incidence IRR | 1.66 (1.32–2.08) | Controlled incidence studies inside a 32-study observational meta-analysis. Abstract only. Prevalence was pooled at 7.70% (5.07–11.52) and the authors call that pool extremely heterogeneous, so it is not given its own row. Coronary and cerebrovascular incidence ratios were not numeric in the retrieved abstract | Motta 2026, PMID 42902526 |
| Diagnostic delay, minority ethnic groups vs White | Median 556 days (IQR 89.8–1342) vs 50.5 days (IQR 6.25–222) | Retrospective cohort, n=88, South London and Maudsley NHS Foundation Trust. Adjusted for age, gender, substance use and illness severity. Deprivation not adjusted. No demographic association with time to lithium initiation | Richards-Brown 2026, PMID 42894408 |
| Sarcopenia odds, women with bipolar disorder | FNIH OR 2.24 (1.11–4.55); EWGSOP2 OR 2.78 (1.47–5.20) | 106 women with bipolar disorder recruited from healthcare settings vs 300 age-matched women without bipolar disorder, Geelong Osteoporosis Study; both ≥20 years; South-Eastern Australia. Independent of age, smoking, activity, alcohol, diet and SES. Low handgrip 15/106 (14.2%) vs 21/300 (7.1%). Confirmed sarcopenia was 2 (FNIH) and 3 (EWGSOP2) women. The paper’s 8.9% and 11.4% figures accompany the combined sample of 406 (36/406 is 8.9%; 46/406 is 11.3%, printed as 11.4%), not the bipolar subgroup of 106 | Williams 2026, PMID 42897866 |
| Usual lithium level, older adults | 0.47–0.67 mmol/L (mean or median) | 14 observational studies, 8,808 older adults. 0.82–1.00 mmol/L associated with side effects or early toxicity; 1.20–1.25 mmol/L overtly toxic. Consistent with age-adjusted laboratory ranges; not an estimate of the concentration that prevents relapse | Nunes 2026, PMID 42460615 |
| Lifetime suicide attempt, pediatric bipolar spectrum | 26.2% (24.4–28.0) | Umbrella review, abstract only: 52 meta-analyses, k=754, N=59,967, mean age 13.0. OR 1.59 (1.24–2.05) vs depressed youth; OR 1.68 (1.29–2.18) vs adults with bipolar spectrum disorder. Source-review quality was not in the retrieved abstract | Torres-Cortes 2026, PMID 42901238 |
Known conflicts and caveats¶
- Diagnostic algorithms can nearly double prevalence in the same survey (Mitchell 2013, PMID 22906117).
- Mixed-feature prevalence differs substantially between strict DSM-5 criteria and broader opposite-polarity symptom counts (Na 2021, PMID 33418368; Vázquez 2018, PMID 28922738).
- Most mortality, metabolic and pregnancy estimates have high heterogeneity and should not be treated as universal patient-level probabilities.
- Pregnancy comparisons against the general population mix medication effects with bipolar illness, smoking, obesity, healthcare use and other confounding; treated-versus-untreated bipolar comparisons remain observational.
- Relative risks, absolute risks, prevalence and standardized mortality ratios answer different questions and are not interchangeable.
- Treatment effects from responder-enriched maintenance trials should not be treated as effectiveness estimates in unselected practice (Nestsiarovich 2022, PMID 34489127).
- Digital-prediction AUCs are predominantly internally validated and likely optimistic; calibration and clinical utility remain unestablished (Fang 2026, PMID 42106724).
- GBD incidence count conflict (flagged 2026-10-09). Jiang 2025 (PMID 40499832) reports global bipolar incidence of 53.9 million cases in 2021. Liang 2026 (PMID 42688390), from the same GBD 2021 round, reports 1.63 million incident and 25.28 million prevalent cases at ages 15–49. Jiang's incidence figure is about twice Liang's 15–49 prevalence, which an age restriction cannot explain. Treat Jiang's figure as unreconciled. The GBD Results Tool was not queried.
Burden-series boundary (2026-10-16). The new WHO estimate uses GBD 2023 and is retained separately from existing GBD 2021 estimates. It does not resolve the previously flagged GBD 2021 incidence/prevalence labelling conflict. The IHME GBD Results page was retrieved but exposed only an application shell; no underlying result table was obtained.
Re-checked 2026-10-23. The IHME GBD Results page was again an application shell (about 5.5 KB), so the Jiang versus Liang labelling conflict stays unresolved. The WHO fact sheet was re-fetched and still reports, for 2023, 36 million people (0.4%), including approximately 32 million adults; the page date is still 11 September 2026. Secondary GBD analyses returned by a follow-up query were not added and were not checked against a GBD results table. SciELO and LILACS could not be searched, so this file makes no claim that regional burden series are absent.