Skip to content

Comorbidity and differential diagnosis in bipolar disorder

TL;DR — Comorbidity is the rule rather than the exception: three-quarters of bipolar-spectrum cases in the World Mental Health surveys had another disorder. Anxiety, ADHD and substance-use disorders are especially common and generally mark earlier onset, more complex treatment and worse outcomes. Differential diagnosis depends primarily on longitudinal timing: bipolar mood and activation changes are episodic, whereas ADHD symptoms are developmentally persistent and borderline-personality instability is usually reactive and trait-like. Screening scales can structure inquiry but cannot adjudicate bipolar disorder from overlapping syndromes. Substance effects, medications and medical conditions must be assessed before attributing activation or mood instability to a primary bipolar disorder.

The comorbidity burden

In the World Mental Health surveys, approximately 75% of people with bipolar-spectrum disorder met criteria for at least one other disorder, most often an anxiety disorder (Merikangas 2011, PMID 21383262). Comorbidity may represent shared causes, one disorder increasing risk for another, treatment effects, or diagnostic overlap; a cross-sectional co-occurrence rate cannot distinguish these mechanisms.

Domain Quantitative estimate Population/source Main caution
Any additional mental disorder ~75% 61,392 adults in 11-country WMH surveys Structured lay interview; cross-sectional
Adult ADHD among bipolar disorder 17.11% (95% CI 13.05–21.59) 71-study meta-analysis High between-study heterogeneity
Bipolar disorder among adult ADHD 7.95% (95% CI 5.31–11.06) Same meta-analysis Referral setting affects prevalence
Any anxiety disorder in pediatric bipolar disorder 44.7% 37-study meta-analysis Pediatric nosology and onset definitions vary
Anxiety disorders in pediatric review 54% weighted mean 167 studies Broader inclusion than the focused meta-analysis
ADHD in pediatric bipolar disorder 48% weighted mean 167 studies Symptom overlap can inflate estimates
Substance-use disorder in pediatric bipolar disorder 31% weighted mean 167 studies Age and ascertainment vary

The difference between the two pediatric anxiety estimates—44.7% and 54%—should be displayed rather than averaged because study inclusion and diagnostic composition differed (Yapıcı Eser 2020, PMID 31899546; Frías 2015, PMID 25545605).

ADHD: overlap and genuine comorbidity

Adult ADHD and bipolar disorder co-occur well above chance. A 71-study meta-analysis covering 646,766 participants estimated ADHD in 17.11% of adults with bipolar disorder and bipolar disorder in 7.95% of adults with ADHD (Schiweck 2021, PMID 33515607).

Comorbid ADHD was associated with bipolar onset 3.96 years earlier (95% CI 2.65–5.26) (Schiweck 2021, PMID 33515607). Genetic evidence supports genuine shared liability: the SNP-based genetic correlation was 0.64 overall and 0.71 for early-onset bipolar disorder (van Hulzen 2017, PMID 27890468).

Feature ADHD Bipolar mania/hypomania Diagnostic use
Time course Persistent from childhood Distinct episode with change from baseline Most important discriminator
Sleep Difficulty settling or irregular habits Reduced need for sleep without expected fatigue Ask about next-day energy, not only hours slept
Speech/activity Chronic talkativeness/restlessness Episodic pressure, acceleration and goal-directed activation Establish baseline and onset
Mood Frustration-linked lability Sustained elevated, expansive or irritable episode Trigger and duration matter
Grandiosity Not a defining ADHD symptom Can be prominent in mania Distinguish confidence from impairment-producing grandiosity
Between episodes ADHD symptoms remain Activation should substantially recede Obtain collateral longitudinal history

A clinical review estimated ADHD in about 20% of adults with bipolar disorder or borderline personality disorder and emphasized that ADHD symptoms remain apparent between bipolar episodes (Asherson 2014, PMID 24804976). This is consensus synthesis rather than a diagnostic-accuracy study.

Large population datasets also show the reverse association: adults with ADHD had pooled OR 8.7 (95% CI 5.47–13.89) for bipolar disorder versus adults without ADHD (Hartman 2023, PMID 37149075). This relative estimate should not be confused with absolute prevalence.

Anxiety disorders

Anxiety is not merely a symptom of mood episodes. In pediatric bipolar disorder, lifetime prevalence estimates were: panic disorder 12.7%, generalized anxiety disorder 27.4%, social phobia 20.1%, separation anxiety 26.1%, and OCD 16.7% (Yapıcı Eser 2020, PMID 31899546).

Age pattern mattered: childhood-onset samples had more generalized and separation anxiety, while adolescent-onset samples had more panic, OCD and social phobia (Yapıcı Eser 2020, PMID 31899546). Separate diagnoses require symptoms that are not confined to mood episodes.

DSM-5 mixed features can further blur the boundary. During manic/hypomanic episodes, mixed features were associated with co-occurring anxiety disorders at OR 2.67 (95% CI 1.28–5.57) (Bartoli 2020, PMID 32697704).

Childhood maltreatment is a shared risk marker rather than a diagnosis. Among people with bipolar disorder, maltreatment history was associated with PTSD (OR 3.60, 95% CI 2.45–5.30) and anxiety disorders (OR 1.90, 1.39–2.61) (Agnew-Blais 2016, PMID 26873185).

Substance use and substance-induced states

Substance use can be comorbid, consequential, causal for an episode-like presentation, or all three. The assessment must establish the timing of intoxication, withdrawal and sustained abstinence relative to mood symptoms.

Substance/exposure Possible bipolar mimic or modifier Key longitudinal question
Stimulants/cocaine Insomnia, pressured speech, grandiosity, agitation, psychosis Did activation begin during exposure and resolve after clearance?
Cannabis Anxiety, paranoia, psychosis; associated with attempt history Is there a dose/time relation and a sober baseline?
Alcohol Intoxication-related disinhibition; withdrawal insomnia/agitation Are symptoms confined to use/withdrawal cycles?
Sedatives Disinhibition during use; anxiety/insomnia during withdrawal Was withdrawal medically supervised and sufficiently observed?
Prescribed corticosteroids Activation, insomnia, psychosis Did symptoms follow initiation or dose escalation?
Antidepressants Activation or treatment-emergent mania in susceptible patients Do syndromal symptoms persist beyond the pharmacological effect?

Within bipolar disorder, substance-use disorder was associated with male sex (OR 2.191, 95% CI 1.121–4.281) and prior suicidality (OR 1.758, 1.156–2.674) (Messer 2017, PMID 28419959). These are correlates, not causal predictors.

Cannabis-use disorder had a weak cross-sectional association with suicide-attempt history (OR 1.35; P=0.01; I²=41.7%) across 11 studies and 6,375 participants; suitable longitudinal data were absent (Bartoli 2019, PMID 31121199).

Maltreatment history was also associated with substance misuse (OR 1.84, 95% CI 1.41–2.39) and alcohol misuse (OR 1.44, 1.13–1.83), showing how early adversity can confound simple comorbidity models (Agnew-Blais 2016, PMID 26873185).

Treatment evidence for dual diagnosis is thin. Across 15 RCTs with 628 active-treatment and 622 placebo participants, adjunctive drugs produced small effects on illness severity (g −0.25, 95% CI −0.44 to −0.06), substance use (g −0.23, −0.44 to −0.02), and abstinence (g 0.21, 0.04–0.38), with low or very-low certainty (Radua 2024, PMID 37689524).

Borderline personality disorder

Bipolar II disorder and borderline personality disorder (BPD) are a recurrent diagnostic dilemma because both can include affective lability, impulsivity, anger, suicidality and unstable functioning. A 28-study systematic review found both overlap and group-level differences across affective, cognitive, behavioral and somatic domains, but did not establish an individually validated discriminator (Massó Rodriguez 2021, PMID 34177664). A cross-sectional machine-learning study of 134 volunteers (82 bipolar, 52 borderline) using cognitive/behavioral constructs, emotion-regulation strategies and recalled parental behaviors reported 73.1–73.9% overall accuracy — but that average conceals an asymmetry: 84.1–87.8% for bipolar disorder against only 50.0–57.7% for borderline personality disorder, the latter no better than chance in a small, bipolar-II-weighted sample. Without external prospective validation this is a research classifier, not a diagnostic test (Bayes 2021, PMID 33845326).

Feature Bipolar II disorder Borderline personality disorder Caveat
Pattern Episodic hypomania/depression Pervasive instability across contexts They can coexist
Triggering Episodes may arise without interpersonal trigger Shifts often follow rejection/abandonment cues Not absolute
Duration Syndromal episodes lasting days to weeks Shifts often minutes to hours Recall bias is common
Baseline Interepisode symptoms may persist, but activation changes from baseline Identity/interpersonal instability is trait-like Chronic bipolar symptoms complicate this
Family history Bipolar loading supports bipolar diagnosis Less specific Family history is probabilistic
Identity Mood-linked changes possible Identity diffusion more characteristic Comparative studies remain limited

The most useful discriminators are family history, developmental antecedents, longitudinal course, mood phenomenology, personality style and relationship patterns; impulsivity, sex distribution, neuropsychology and treatment response are less discriminating (Bayes 2019, PMID 31749106).

Identity-focused research suggests BPD self-concept shifts are more closely tied to interpersonal triggers, whereas bipolar shifts may track internal mood and motivation; this remains a proposed clinical discriminator rather than a validated test (Wright 2022, PMID 33811707).

High reported BD–BPD comorbidity may combine true pleiotropic/environmental overlap with measurement error from transdiagnostic symptoms (Parker 2022, PMID 35397334). A forced either/or diagnosis can therefore be as misleading as uncritical dual diagnosis.

Unipolar depression

Bipolar II is especially difficult to distinguish from recurrent MDD because patients commonly present during depression and may not identify hypomania as pathological (Phillips 2013, PMID 23663952).

Features that should prompt a structured lifetime hypomania/mania assessment include episodic reduced need for sleep, distinct activation, psychotic depression, postpartum onset, antidepressant-associated activation, early onset, recurrent brief episodes and bipolar family history. None is individually diagnostic.

Screening instruments help retrieve overlooked history but perform differently by setting. In mental-health services, HCL-32 sensitivity/specificity was 81%/67%, MDQ 66%/79%, and BSDS 69%/86%; in primary care/general populations, MDQ cutoff 7 had sensitivity 43% and specificity 95% (Carvalho 2015, PMID 25451435).

Thus, a negative MDQ in primary care does not reliably exclude bipolar disorder, and a positive HCL-32 in specialty care has substantial false positives. Every positive screen requires a diagnostic interview and episode chronology (Sayyah 2022, PMID 35588536).

Psychotic disorders

Mania and severe bipolar depression can include hallucinations, delusions and disorganization. The central distinction from schizophrenia-spectrum disorders is the longitudinal relationship between psychosis and syndromal mood episodes, not the content of psychosis alone.

Genetic findings reinforce overlap: bipolar I is strongly genetically correlated with schizophrenia, especially when psychosis is present (Stahl 2019, PMID 31043756). Shared biology should not erase the clinical importance of mood–psychosis timing.

First-episode services can under-recognize affective psychoses. A 2026 meta-analysis of 83 studies and 30,946 participants estimated affective psychoses at 18.0% (95% CI 15.4–20.6) of early-intervention presentations, with a wide 95% prediction interval of 3.6–39.4% (Catalán 2026, PMID 42057639).

Medical and medication mimics

Medical evaluation should be driven by presentation, age, examination and exposure history rather than an indiscriminate panel. A new first manic syndrome, fluctuating consciousness, focal neurological signs, autonomic instability or atypical late onset increases concern for a secondary cause.

Category Examples Clues against primary bipolar disorder
Endocrine/metabolic Thyroid disease, Cushing syndrome, electrolyte disturbance Systemic signs; temporal laboratory abnormality
Neurological Seizure disorder, tumor, stroke, autoimmune encephalitis Focal signs, seizures, delirium, rapid cognitive change
Infectious/inflammatory CNS infection, systemic inflammatory disease Fever, neurological signs, altered consciousness
Medication-induced Corticosteroids, dopaminergic drugs, stimulants Onset follows exposure or dose change
Sleep/circadian Severe sleep deprivation, shift disruption Symptoms remit with sleep restoration and lack full syndrome

The differential is particularly important because “bipolar-like” activation plus inattention, affective instability or substance use can lead both to overdiagnosis and underdiagnosis. The evidentiary anchor is repeated longitudinal observation.

A practical diagnostic sequence

  1. Define the putative episode: change from baseline, duration, impairment, sleep need, cognition, behavior and psychosis.
  2. Build a lifetime polarity timeline, including depressions, activation, mixed symptoms, postpartum periods and treatment-emergent changes.
  3. Establish whether ADHD, anxiety, personality or substance symptoms persist outside mood episodes.
  4. Obtain collateral history when insight or recall is limited.
  5. Reconstruct all substance, prescribed-drug and withdrawal timing.
  6. Investigate medical causes proportionately to atypical features.
  7. Treat screening scales as prompts, not verdicts.
  8. Revise the formulation as longitudinal evidence accumulates.

Quantified overlap does not establish diagnostic identity

OCD co-occurrence is substantial: 29 studies (n=6,109) yielded cross-sectional prevalence 11.2% (95% CI 7.6%–15.3%) and 39 studies (n=8,205) yielded lifetime prevalence 10.9% (7.8%–14.4%), about 4.4 times the general-population estimate but not significantly different from major depression (Ferentinos 2020, PMID 31818777). Similar cross-sectional and lifetime estimates suggest persistence, yet extreme heterogeneity was not explained by setting, subtype, remission or diagnostic procedure.

From the opposite denominator, 96 autism studies estimated bipolar disorder in 5% (95% CI 3%–6%), alongside ADHD in 28% and anxiety in 20%; residual heterogeneity exceeded 95% (Lai 2019, PMID 31447415). Episodic change from baseline, developmental history and collateral information remain essential because prevalence overlap cannot distinguish autistic dysregulation from mania.

Epilepsy is both a comorbidity and a mimic context. A 27-study meta-analysis including 565,443 people with epilepsy and 13.4 million controls found bipolar-disorder odds 3.12 (95% CI 2.23–4.36), alongside elevated psychosis, depression, anxiety and substance-use disorders (Kwon 2025, PMID 39585664). Seizure phenomenology, antiseizure medication effects and postictal states must therefore be considered before assigning recurrent behavioral change to bipolar disorder.

Stimulants: temporal association versus causation

Across 16 studies of 391,043 people with ADHD exposed to stimulants, pooled occurrence was 2.76% for psychotic symptoms (95% CI 0.73–9.88; k=10), 2.29% for psychotic disorder (1.52–3.40; k=4) and 3.72% for bipolar disorder (0.77–16.05; k=4), all with I²>95%. The bipolar interval spans a twentyfold range across four studies, so the point estimate carries almost no precision. Amphetamine exposure was associated with more psychosis than methylphenidate (OR 1.57, 95% CI 1.15–2.16), but the designs could not establish causality or separate unmasking from induction (Salazar de Pablo 2025, PMID 40900605). Mixed states further complicate the boundary because DSM-5 broadened the phenotype while overlapping agitation and affective lability occur in borderline personality disorder (Betzler 2017, PMID 28417647).

Open questions

  • Which longitudinal features best discriminate bipolar II from BPD when both are present (Bayes 2019, PMID 31749106)?
  • Can prospective childhood cohorts separate persistent ADHD from prodromal bipolar activation without circular diagnostic definitions (Schiweck 2021, PMID 33515607)?
  • Does integrated treatment of anxiety improve bipolar recurrence, rather than anxiety symptoms alone (Yapıcı Eser 2020, PMID 31899546)?
  • Which interventions improve both mood and substance outcomes with more than small, low-certainty effects (Radua 2024, PMID 37689524)?
  • How much apparent BD–BPD comorbidity remains after blinded longitudinal adjudication (Parker 2022, PMID 35397334)?

References

  1. Merikangas KR, et al. Prevalence and correlates of bipolar spectrum disorder in the world mental health survey initiative. Archives of General Psychiatry. 2011. PMID 21383262.
  2. Schiweck C, et al. Comorbidity of ADHD and adult bipolar disorder: A systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews. 2021. PMID 33515607.
  3. van Hulzen KJE, et al. Genetic Overlap Between Attention-Deficit/Hyperactivity Disorder and Bipolar Disorder. Biological Psychiatry. 2017. PMID 27890468.
  4. Asherson P, et al. Differential diagnosis, comorbidity, and treatment of attention-deficit/hyperactivity disorder in relation to bipolar disorder or borderline personality disorder in adults. Current Medical Research and Opinion. 2014. PMID 24804976.
  5. Hartman CA, et al. Anxiety, mood, and substance use disorders in adult men and women with and without attention-deficit/hyperactivity disorder. Neuroscience & Biobehavioral Reviews. 2023. PMID 37149075.
  6. Yapıcı Eser H, et al. Anxiety disorders comorbidity in pediatric bipolar disorder: a meta-analysis and meta-regression study. Acta Psychiatrica Scandinavica. 2020. PMID 31899546.
  7. Frías Á, et al. Comorbidity in pediatric bipolar disorder: prevalence, clinical impact, etiology and treatment. Journal of Affective Disorders. 2015. PMID 25545605.
  8. Bartoli F, et al. Clinical correlates of DSM-5 mixed features in bipolar disorder: A meta-analysis. Journal of Affective Disorders. 2020. PMID 32697704.
  9. Agnew-Blais J, et al. Childhood maltreatment and unfavourable clinical outcomes in bipolar disorder. Lancet Psychiatry. 2016. PMID 26873185.
  10. Messer T, et al. Substance abuse in patients with bipolar disorder: A systematic review and meta-analysis. Psychiatry Research. 2017. PMID 28419959.
  11. Bartoli F, et al. Cannabis use disorder and suicide attempts in bipolar disorder: A meta-analysis. Neuroscience & Biobehavioral Reviews. 2019. PMID 31121199.
  12. Radua J, et al. Meta-analysis of the effects of adjuvant drugs in co-occurring bipolar and substance use disorder. Spanish Journal of Psychiatry and Mental Health. 2024. PMID 37689524.
  13. Bayes A, et al. Differential Diagnosis of Bipolar II Disorder and Borderline Personality Disorder. Current Psychiatry Reports. 2019. PMID 31749106.
  14. Wright L, et al. Differential diagnosis of borderline personality disorder and bipolar disorder: Self-concept, identity and self-esteem. Clinical Psychology & Psychotherapy. 2022. PMID 33811707.
  15. Parker G, et al. Why might bipolar disorder and borderline personality disorder be bonded? Journal of Psychiatric Research. 2022. PMID 35397334.
  16. Phillips ML, Kupfer DJ. Bipolar disorder diagnosis: challenges and future directions. Lancet. 2013. PMID 23663952.
  17. Carvalho AF, et al. Screening for bipolar spectrum disorders: A comprehensive meta-analysis of accuracy studies. Journal of Affective Disorders. 2015. PMID 25451435.
  18. Sayyah M, et al. Assessment of the diagnostic performance of two new tools versus routine screening instruments for bipolar disorder: a meta-analysis. Revista Brasileira de Psiquiatria. 2022. PMID 35588536.
  19. Stahl EA, et al. Genome-wide association study identifies 30 loci associated with bipolar disorder. Nature Genetics. 2019. PMID 31043756.
  20. Catalán A, et al. The proportion of people with primary affective disorders presenting to early intervention services: systematic review and meta-analysis. British Journal of Psychiatry. 2026. PMID 42057639.
  21. Massó Rodriguez A, et al. A systematic review of bipolar disorder and borderline personality disorder: overlap and discrimination of both diagnoses. Aust N Z J Psychiatry. 2021. PMID 34177664.
  22. Bayes A, et al. Differentiation of bipolar disorder versus borderline personality disorder: a machine learning approach. J Affect Disord. 2021;288:68–73. PMID 33845326.
  23. Ferentinos P, et al. Comorbidity of obsessive-compulsive disorder in bipolar spectrum disorders: systematic review and meta-analysis of its prevalence. J Affect Disord. 2020;263:193–208. PMID 31818777
  24. Lai MC, et al. Prevalence of co-occurring mental health diagnoses in the autism population: a systematic review and meta-analysis. Lancet Psychiatry. 2019;6:819–829. PMID 31447415
  25. Salazar de Pablo G, et al. Occurrence of psychosis and bipolar disorder in individuals with ADHD treated with stimulants: a systematic review and meta-analysis. JAMA Psychiatry. 2025;82:1103–1112. PMID 40900605
  26. Kwon CS, et al. Psychiatric comorbidities in persons with epilepsy compared with persons without epilepsy: a systematic review and meta-analysis. JAMA Neurol. 2025;82:72–84. PMID 39585664
  27. Betzler F, et al. Mixed states in bipolar disorder—changes in DSM-5 and current treatment recommendations. Int J Psychiatry Clin Pract. 2017;21:244–258. PMID 28417647