What does the research say?
In a study of 35 people with bipolar depression, participants were assigned by lottery to take CBD or a dummy pill (placebo) in addition to their current treatment. After eight weeks, depression scores had dropped by about 15 points in both groups, and the difference between them was not significant. A positive signal was seen in those who took a higher dose; however, this was an additional analysis and needs to be confirmed. The study was also ended earlier than planned because of the COVID-19 pandemic.
In two patients in a manic episode, CBD was found to be ineffective. Information on the safety of pure CBD in bipolar disorder also rests largely on a single small study. A larger study planned to include 360 people is still recruiting participants; according to its registry entry, the first results are not expected before the end of 2029.
The picture is different for cannabis containing THC. According to a 2026 review, THC-dominant cannabis can worsen manic symptoms and day-to-day functioning in bipolar disorder; for this reason, people with bipolar or psychotic disorders are advised to avoid cannabis. In a review of 19 follow-up studies published online in October 2026 (dated 2027 in the journal), a later diagnosis of bipolar disorder and manic symptoms were about twice as common in people who used cannabis. This is an association; the rates were not adjusted for confounding factors and do not show cause and effect. In a more recent study in California that followed 463,000 young people and took factors such as alcohol and other substance use into account, a later diagnosis of bipolar disorder was also about twice as common among cannabis users; this, too, is an association.
Most studies that track the link between cannabis use and the course of illness in people with bipolar disorder paint a negative picture. In a study in Europe that followed about 1900 people for two years after they started treatment for a manic or mixed episode, those who kept using cannabis recovered less and had new episodes more often; the outcomes of those who stopped using cannabis during the episode were similar to those of people who had never used it. In a study that followed people hospitalized for a first manic episode, cannabis use was associated with more time spent in episodes and with rapid cycling. There are signals for depression as well: among patients followed for two years, depressive symptoms eased less often in regular cannabis users; in a study that followed about 800 people for years, periods of increased cannabis use were followed by a small increase in suicidal thoughts; and in a small study in which 24 people kept a daily diary for six days, cannabis use was followed by increases in manic and depressive symptoms, while low mood did not predict cannabis use. On the other hand, in two follow-up studies cannabis did not on its own predict new episodes or rehospitalization. All of these studies are observational: cannabis users also smoke and drink alcohol more often, and it has not been shown that cannabis causes these outcomes.
THC itself has hardly been tested as a medicine in bipolar depression. In two small, double-blind studies in the US in the 1970s with 21 hospitalized patients with depression (9 of them bipolar), THC did not improve depression; about half of the patients could not continue treatment because of intense anxiety, a sense of detachment from oneself, panic or psychosis-like reactions. The original articles of these studies could not be obtained; the information was taken from a review that examined them. A laboratory study published in 2026, in which people with and without bipolar disorder were given a single dose of THC or placebo by lottery, looked only at attention and time-perception tests; it reported no results on mood.
Strength of evidence: ●○○○ Very weak evidence. Finding: No effect found. What does this mean?
Ongoing research on this topic
Trials testing CBD for this topic on ClinicalTrials.gov: 2 under way (2 recruiting). A trial being under way does not show that CBD works.
- NCT05867849: Phase 3, recruiting, target 360 participants
- NCT05457465: Phase 2, recruiting, target 25 participants
Possible side effects and drug interactions
In the bipolar depression study, people taking CBD did not have more side effects than those on placebo, and there was no difference in the development of mania. On the other hand, mania developed in a person with no psychiatric history after vaping a product of unverified purity at a high dose; it resolved after CBD was stopped and the person was treated.
CBD can slow down how the body breaks down some medicines. In healthy adults, the amount of the antidepressant citalopram in the body rose by 43% with CBD. A single case was also reported of a patient taking lithium whose lithium level rose to a harmful degree (lithium toxicity) weeks after CBD was added.
In two old, small studies in which THC was tested in patients with depression, 10 of 21 patients could not continue treatment because of side effects such as intense anxiety, a sense of detachment from oneself, panic or psychosis-like reactions. This information was taken from a review that examined these studies, whose original articles could not be obtained.
Interactions with medicines used for this condition
Interactions reported for these medicines in peer-reviewed publications and official prescribing information. A medicine that is not listed can still interact with CBD; tell your doctor about every medicine you take.
| Medicine | Importance | Warning | Source |
|---|---|---|---|
| Citalopram (SSRI) (Cipram, Citol, Relaxol) | Caution | Citalopram, escitalopram: CBD raised citalopram blood levels: a significant rise in a series of 6 patients; in healthy volunteers, citalopram exposure (AUC) rose by 43% after 12 days of approved CBD (2.5 mg/kg twice a day). If you take these medicines, talk to your doctor. | PMID 34121064, PMID 41622700 |
| Amitriptyline (tricyclic) (Laroxyl) | Caution | Amitriptyline: In healthy volunteers, a single 30 mg dose of CBD oil raised the blood level (AUC) of amitriptyline by 13% and its peak level by 17%; the effect is said to possibly be larger with ongoing use. Tell your doctor or pharmacist. | PMID 41398552 |
| Lithium (Lithuril, Litomaks) | Caution | Lithium: In a 13-year-old patient who had taken the same dose of lithium for a long time, lithium toxicity developed a few weeks after CBD was started (blood level 2.4 mmol/L; earlier levels 1–1.3); it resolved when lithium was stopped. The patient was also taking other medicines and drinking little fluid; CBD's role was rated "possible", and the cause is not known. Do not start without talking to your psychiatrist. | PMID 32851114, PMID 39719830 |
To check interactions with the medicines you take: CBD drug interaction checker.
How was CBD used in studies?
In the bipolar depression study, CBD was started at 150 mg a day (study dose; not a personal recommendation); after two weeks it was raised to 300 mg a day (study dose; not a personal recommendation) in those who had not responded well enough, and it was used for 8 weeks in addition to current treatment. In two patients in a manic episode, 600–1200 mg a day (study dose; not a personal recommendation) was tried.
Study details (for health professionals)
As add-on treatment in bipolar depression, 150–300 mg/day of pure CBD did not lower the depression score (MADRS) differently from placebo at 8 weeks; both groups had a drop of about 15 points (strong placebo effect). An exploratory subanalysis shows a signal in favor of 300 mg; no difference was seen in the development of mania. In mania, 600–1200 mg/day of CBD was ineffective in 2 patients. A phase 3 study of 360 people is underway (according to its registry entry, the estimated completion for primary outcomes is December 2029). Treatment data for THC-containing products in bipolar depression consist only of two small double-blind studies from the 1970s in hospitalized patients with depression (21 people, 9 of them bipolar): THC did not improve depression, and about half of the patients could not continue treatment because of side effects such as intense anxiety or psychosis-like reactions (Kotin 1973, Ablon 1974; original articles could not be read, details from Stanciu 2021). The 2026 single-dose THC laboratory study (Miranda 2026) looked only at cognitive tests. The most recent meta-analysis of RCTs did not find enough data in bipolar disorder to pool (Wilson 2026). Most studies following cannabis use in people with bipolar disorder are negative: in a 2-year follow-up of 1922 people in Europe, those who kept using had less recovery and more frequent new episodes, while outcomes in those who stopped during the episode were similar to those who never used (Zorrilla 2015); among those followed after a first hospitalization for mania, cannabis was associated with more time spent in episodes and with rapid cycling (Strakowski 2007); over 24 months, remission of depressive symptoms was less frequent in regular users (Kim 2015, unadjusted); in a long follow-up of 788 people, periods of increased cannabis use were followed by a small increase in suicidal thoughts (Smith 2025); in diaries kept by 24 people for 6 days, cannabis use was followed by increases in manic and depressive symptoms, while depressive symptoms did not predict use (Tyler 2015). On the other hand, in two follow-up studies cannabis did not independently predict new episodes or rehospitalization (Lagerberg 2021, Olivier 2025). Systematic reviews point in the same direction (Sorkhou 2024; Tourjman 2023, with low certainty of evidence for most outcomes); in cross-sectional data, cannabis use disorder is weakly associated with suicide attempts (OR 1.35; Bartoli 2019). Cannabis users also smoke and drink alcohol more often, and a reverse relationship, such as symptoms increasing use, cannot be ruled out; cause and effect has not been shown. October 2026 search (PubMed, Cochrane, ClinicalTrials.gov): 4 controlled studies with cannabinoids on this topic were found (1 CBD RCT in bipolar depression; 2 old THC studies in which some patients were bipolar; 1 single-dose THC laboratory study), along with 6 systematic reviews or meta-analyses and 1 umbrella review; there is no Cochrane review and no unpublished study with registered results, and 1 phase 3 study is underway. All of them, together with 8 observational studies, 2 case reports and 1 clinical review, are in the table below.
Evidence assessment: Very weak: a single small RCT that was ended early and gave results no different from placebo (Pinto 2024)
CBD alone
Pure CBD was not superior to placebo in bipolar depression; safety data rest on a single small RCT.
Products containing THC
THC-dominant cannabis can worsen manic symptoms and functioning in bipolar disorder; people with bipolar or psychotic disorders are advised to avoid cannabis (Kansagara 2026; the systematic review by Canada's CANMAT guideline group makes the same recommendation: Tourjman 2023). In two small, old controlled studies in which THC was tested in patients with depression (some of them bipolar), depression did not improve and about half of the patients could not continue treatment because of side effects (Kotin 1973, Ablon 1974; information from a review). In most studies following people with bipolar disorder, cannabis users recovered less and had episodes more often (Zorrilla 2015, Strakowski 2007); these findings are observational and do not show cause and effect.
Evidence: Very weak; with THC products there are only two small controlled studies from the 1970s in patients with depression (some of them bipolar) (no benefit, frequent distressing reactions; Kotin 1973, Ablon 1974, details from a secondary source); the harm signal rests on observational studies: with continued use in bipolar disorder, less recovery and more frequent episodes (Zorrilla 2015; Sorkhou 2024 and Tourjman 2023 reviews, with low certainty of evidence for most outcomes in Tourjman), and a later diagnosis of bipolar disorder and mania more common in cannabis users (Sanjuan-Ortiz 2027, unadjusted; 1.7–2-fold in a large adjusted cohort: Young-Wolff 2026); a recommendation to avoid it in a guideline and a clinical review (Tourjman 2023, Kansagara 2026)
Dose range in studies
150–300 mg/day in the RCT; 600–1200 mg/day in a case report in mania
Starting dose and titration
In Pinto 2024, 150 → 300 mg/day (the dose was increased at week 2 in non-responders)
These are doses researchers gave under medical supervision; they are not a recommended dose for you. "mg/kg/day" means milligrams per kilogram of body weight per day.
| Study | Participants | Product | Dose and titration | Duration | Result | Source |
|---|---|---|---|---|---|---|
| Pinto 2024 Double-blind, placebo-controlled pilot RCT (ended early because of the COVID-19 pandemic) | 35 randomized (CBD 19, placebo 16); 36 participants in the ClinicalTrials.gov registry Acute bipolar depression (add-on treatment) | Pure CBD (99.6% purity, THC-free; capsules in corn oil, oral) | 150–300 mg/day Started at 150 mg/day; raised to 300 mg/day at week 2 in non-responders (if there was no drop of at least 50% on the MADRS) | 8 weeks efficacy + safety up to week 12 | MADRS drop similar to placebo (−15.38 versus −14.56; difference −0.82, 95% CI −7.81 to 6.17; p = 0.82); no difference in anxiety; in an exploratory analysis, a signal in favor of 300 mg (among 11 patients per group who had not responded at week 2 and had their dose raised to 300 mg, MADRS drop over weeks 2–8 −13.72 versus −6.64; p = 0.01; the authors interpret this only as a signal to be tested in the future). Six of the authors are co-inventors on CBD-related patents (the article's conflict-of-interest statement) Side effects: No more side effects than placebo (1 person in the CBD group stopped because of a skin rash); no difference in the development of mania | PMID 37920963, doi:10.1177/07067437231209650, NCT03310593 |
| Zuardi 2010 Case report (2 patients; inpatient; 5 days of placebo each at the start and end) | 2 Manic episode in bipolar I disorder (two women aged 34 and 36; inpatient) | CBD | 600–1200 mg/day Started orally at 600 mg/day and raised to 1200 mg/day; the first patient also took olanzapine on days 6–20 | 25 days of CBD (days 6–30) | CBD was found to be ineffective in the manic episode: the first patient improved while on olanzapine, but CBD alone gave no additional improvement; the second patient did not improve at any CBD dose Side effects: No side effects reported; both patients tolerated CBD well | PMID 18801823, doi:10.1177/0269881108096521 |
| Auf 2025 Case report | 1 A 31-year-old man with no psychiatric history | Vaped CBD (product purity not verified) | Increasing, 'high dose' (mg not reported) Self-escalation | 3 months | Manic episode; resolved after CBD was stopped and treatment was given Side effects: Mania | PMID 40630971, doi:10.64719/pb.4542 |
| NCT05867849 Phase 3, double-blind, placebo-controlled RCT (ongoing) | Target 360 Acute bipolar depression that has not responded to at least one first-line treatment (bipolar I or II, aged 19–70; add-on treatment; 6 centers in Canada, 3 of them recruiting) | CBD | 200–600 mg/day, added to current treatment (according to the registry) not reported | 6 weeks (according to the registry) | Recruiting (University of British Columbia; started October 2023; registry last updated April 23, 2025; estimated completion for the primary outcome December 2029, study completion December 2030); no results | NCT05867849 |
| Sanjuan-Ortiz 2027 Systematic review and meta-analysis (19 longitudinal observational studies; PubMed and Web of Science, up to October 2025) | 6,960,517 (total); bipolar disorder analysis 206,572 (8 studies), manic symptoms 57,243 (6 studies) People without bipolar disorder at baseline (mean age 23.8) | Cannabis use | Cannabis use is associated with later bipolar disorder (RR 2.46; 95% CI 1.57–3.84) and manic/hypomanic symptoms (RR 2.11; 1.07–4.18); unadjusted ratios, observational data, does not show cause and effect | PMID 42838231, doi:10.1016/j.jad.2026.122588 | ||
| Kotin 1973 Double-blind, placebo-controlled small study (inpatient; 7 days placebo, 7 days THC, 7 days placebo) | 8 (4 bipolar, 4 unipolar depression); 4 people completed the THC period Hospitalized, medication-free adults with moderate to severe depression (US National Institute of Mental Health research unit) | THC (oral) | Oral THC; dose conflicting in the secondary source (Stanciu 2021's text: 0.3 mg/kg twice a day; the same review's table: from 5 mg a day to 2 × 20 mg a day) not reported (original article could not be read) | 3 weeks (7 days of THC) | THC did not improve depression on subjective and observer-rated mood measures; scale scores not reported. No separate result for the bipolar subgroup. The original article and its abstract could not be read; information from the Stanciu 2021 review Side effects: 4 of 8 patients stopped THC because of side effects (two had severe anxiety with a sense of detachment from oneself after the first dose); drowsiness in those who completed | PMID 4569675, doi:10.1001/archpsyc.1973.01750330041007 |
| Ablon 1974 Double-blind, placebo-controlled small study (inpatient; THC or placebo after 3 weeks of observation) | 13 (5 bipolar, 8 unipolar depression) Hospitalized adults with depression (US National Institute of Mental Health research unit) | THC (oral) | Oral THC; dose conflicting in the secondary source (Stanciu 2021's text: from 5 mg a day to 2 × 20 mg a day; the same review's table: 0.3 mg/kg twice a day, 7 days) Conflicting in the secondary source: according to Stanciu 2021's text, raised from 5 mg a day to 2 × 20 mg a day; according to its table, 0.3 mg/kg twice a day (original article could not be read) | not reported (unclear in the secondary source) | Depressive symptoms did not improve; standard scale scores not reported. A distressing (dysphoric) reaction was seen in 7 of 13 patients: 6 had unipolar and 1 bipolar depression. The original article and its abstract could not be read; information from the Stanciu 2021 review Side effects: 6 patients could not continue the study after the first dose because of panic or psychosis-like reactions | PMID 4592555, doi:10.1176/ajp.1974.131.4.448 |
| Miranda 2026 Randomized, placebo-controlled single-dose THC laboratory study; also a cross-sectional comparison of cannabis users and non-users | not reported (not in the abstract; the full text will become open on PMC in October 2027) Randomized part: adults with bipolar disorder who do not use cannabis, and healthy volunteers; cross-sectional part: bipolar patients who do and do not use cannabis, and healthy volunteers (US) | THC (Δ9-THC; route of administration not in the abstract) | not reported (not in the abstract) None (single dose) | Single dose | Bipolar participants who received a single dose of THC had faster time perception than healthy volunteers who received placebo; the abstract does not compare bipolar participants given THC with those given placebo. In the cross-sectional part, the time perception and attention tests of bipolar patients who used cannabis were similar to those of healthy volunteers, while bipolar patients who did not use cannabis had less precise time perception; the authors interpret this as consistent with an attempt at 'self-medication.' Only laboratory cognitive tests were measured; mood, depression or clinical outcomes are not in the abstract Side effects: not reported (in the abstract) | PMID 42481733, doi:10.1038/s41386-026-02489-w, NCT04231643, bipolar patients and healthy volunteers, in the registry, a single oral dose of 5 mg dronabinol, 600 mg CBD or placebo, ended at 19 people for budget reasons, no results in the registry), the abstract gives no registry number, so the match could not be verified |
| Wilson 2026 Systematic review and meta-analysis of RCTs (1980 to May 13, 2025; PROSPERO CRD42023392718) | 54 RCTs; 2,477 participants (all mental disorders and substance use disorders) People whose primary diagnosis is a mental disorder or substance use disorder; bipolar disorder under a separate heading | CBD, THC, THC and CBD mixtures and other cannabinoids | Not enough data were found for a meta-analysis in bipolar disorder; no RCTs at all were found for the treatment of depression (the search starts in 1980, so the two 1970s THC studies fall outside its scope). Across all disorders, there was no significant difference from placebo for anxiety, psychosis and post-traumatic stress disorder outcomes; significant benefit was seen only in cannabis use disorder, insomnia, tics and Tourette syndrome, and autism; the certainty of evidence is low for most outcomes, and 44% of studies are at high risk of bias. The correction published in the journal (June 16, 2026) corrected the axis of a figure and the study counts in the table; the results did not change Side effects: Any side effect more common with cannabinoids (OR 1.75; 95% CI 1.25–2.46; number needed to harm 7); no difference in serious adverse events or study discontinuation | PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged) | ||
| Stanciu 2021 Systematic review of controlled studies (1970 to February 5, 2020; 8 databases; no meta-analysis) | 8 controlled studies; two in patients with depression (21 people, 9 of them bipolar) People with mood disorders, anxiety disorders and post-traumatic stress disorder | Defined doses of CBD, THC or nabilone (a synthetic THC-like cannabinoid) | In two small 1970s studies with THC in bipolar and unipolar depression, depression did not improve, and anxiety and psychosis-like symptoms emerged in more than half of the patients. No placebo-controlled studies were found with THC or cannabis in bipolar mania or long-term treatment, or with CBD in mood disorders (the search ended in February 2020; Pinto 2024 was published later). The authors conclude that cannabinoids should not be recommended for these disorders Side effects: In the depression studies, intense anxiety, a sense of detachment from oneself, panic and psychosis-like reactions | PMID 33530732, doi:10.1176/appi.ps.202000189 | ||
| Sorkhou 2024 Systematic review (inception to November 2023; EMBASE, PsycINFO, PubMed, MEDLINE; no meta-analysis) | 78 studies (1 RCT, 52 longitudinal, 25 cross-sectional); 30 related to bipolar disorder or mania The general population and people with major depression and bipolar disorder | Cannabis use and cannabis use disorder; one CBD RCT | 4 of 5 prospective studies following cannabis use in bipolar disorder reported negative outcomes (less recovery and remission, more episodes, rapid cycling, mania severity); in 11 of 13 cross-sectional studies, cannabis went together with more severe depressive and manic symptoms, more hospitalizations, poorer medication adherence and more suicide attempts. In a small study, bipolar patients reported that their mood improved right after smoking cannabis. In the only RCT (Pinto 2024), there was no difference between CBD and placebo. Most studies are observational, and cannabis use was mostly self-reported; cause and effect cannot be shown | PMID 38655516, doi:10.3389/fpubh.2024.1346207 | ||
| Tourjman 2023 (CANMAT) Systematic review and clinical recommendation (Canadian CANMAT task force; inception to October 2020; PubMed, Embase, PsycINFO, CINAHL, Cochrane; GRADE; no meta-analysis) | 56 studies (1 RCT, 3 prospective, 26 longitudinal or cohort, 26 cross-sectional); 23 in bipolar disorder, 11 in both bipolar disorder and major depression; 73,891 people in the bipolar studies (2761 with cannabis use or cannabis use disorder) People with bipolar disorder or major depression | Cannabis use and cannabis use disorder | In bipolar disorder, cannabis use was associated with more manic and mixed episodes, rapid cycling and psychotic symptoms, more severe depressive and manic symptoms, earlier onset, more suicide attempts and suicide deaths, and lower functioning; in one study, less remission and more frequent recurrence. The associations were more consistent in bipolar disorder than in major depression. The certainty of evidence is low for most outcomes (studies mostly of low quality, cannabis amount measured in different ways, THC/CBD content not determined, confounders not controlled for some outcomes). Recommendation, also informed by expert clinical opinion: people with bipolar disorder should avoid cannabis (in the recommendation table, moderate certainty of evidence, strong recommendation). No studies were found on treating bipolar disorder and cannabis use disorder together | PMID 35711159, doi:10.1177/07067437221099769 | ||
| Bartoli 2019 Systematic review and meta-analysis of observational studies (up to July 2018) | 11 studies; 6375 people People with bipolar disorder | Cannabis use disorder | Cannabis use disorder was weakly but significantly associated with a history of suicide attempts (OR 1.35; p=0.01; cross-sectional data); no publication bias was detected. Prospective data were insufficient for meta-analysis; cause and effect could not be assessed | PMID 31121199, doi:10.1016/j.neubiorev.2019.05.017 | ||
| Solmi 2023 Umbrella review (meta-analyses of RCTs and observational studies; up to February 9, 2022) | 101 meta-analyses (50 observational, 51 RCT) General population and various patient groups | Cannabis, cannabinoids and cannabis-based medicines | In the general population, the evidence for an association between cannabis use and manic symptoms was weak in the main analysis and 'suggestive' in the sensitivity analysis (based on the Gibbs 2015 meta-analysis); the strongest harm association was with psychosis in adolescents. In RCTs, cannabis-based medicines increased psychiatric side effects (moderate certainty). It includes no meta-analysis on the treatment of bipolar disorder | PMID 37648266, doi:10.1136/bmj-2022-072348 | ||
| Kansagara 2026 Clinical (narrative) review; not a systematic review | People with mental illness (aimed at physicians) | THC-dominant cannabis and CBD | The benefit of cannabis in mental illness has not been studied enough; THC-dominant cannabis can worsen manic symptoms and functioning in bipolar disorder and is associated with a risk of self-harm in young people with mood disorders. The authors recommend that people with bipolar or psychotic disorders avoid cannabis. Not a systematic review; the primary studies it relies on are not specified in the abstract | PMID 41801216, doi:10.1001/jamainternmed.2025.8215 | ||
| Zorrilla 2015 (EMBLEM) Prospective observational cohort (EMBLEM; Europe, multicenter) | 1922 analyzed (3459 people in the 12-month analysis of the same cohort) Adults with bipolar disorder starting treatment for a manic or mixed episode | Cannabis use | 2 years | Those who continued using (6.9%) had less recovery (p=0.004) and remission (p=0.014), more frequent new episodes (p=0.014) and more impairment in work life than never-users. Outcomes of those who stopped during the episode (4.6%) were similar to never-users (remission 68.1%, recovery 38.7%). In the 12-month analysis of the same cohort, cannabis users had lower treatment adherence and higher illness severity, mania and psychosis (van Rossum 2009). Observational; stopping was not assigned by lottery | PMID 25430820, doi:10.1111/acps.12366, 12-month analysis of the same EMBLEM cohort: van Rossum 2009, PMID 19155808, PMID 20458676 | |
| Strakowski 2007 Inception cohort (prospective follow-up after a first hospitalization for mania) | 144 (in 33, cannabis use disorder began before bipolar disorder, in 36 after; 75 with bipolar disorder only) Bipolar I patients aged 12–45 hospitalized for the first time with a manic or mixed episode, with no previous hospitalization and very little prior treatment (US) | Cannabis use disorder | Up to 5 years | Cannabis use was associated with more time spent in mood episodes during follow-up and with rapid cycling. Recovery at first appeared better in those whose cannabis use disorder began before bipolar disorder, but the difference disappeared after adjustment for mediating variables. Most cannabis use disorders improved right after hospitalization, then quickly returned | PMID 17199055, doi:10.1001/archpsyc.64.1.57 | |
| Kim 2015 (BCOS) Prospective observational cohort (Australia, Bipolar Comprehensive Outcomes Study; 9 assessments over 24 months) | 239 enrolled; 234 analyzed (25 regular cannabis users) Adults with bipolar I disorder or bipolar-type schizoaffective disorder taking mood stabilizers and/or olanzapine | Regular cannabis use (3 or more days a week; self-reported at baseline) | 24 months | Among regular cannabis users, overall remission and remission of depressive symptoms (HAMD-21) were significantly less frequent over 24 months; remission of manic symptoms was also less frequent, but the difference was borderline (p=0.051). Less remission of depressive symptoms in women and of manic symptoms in men. Only 3 of the cannabis users did not smoke cigarettes daily; remission was lowest in those who used both tobacco and cannabis. Cannabis use was asked about only at baseline; comparisons were not adjusted for confounders | PMID 26207128, doi:10.4306/pi.2015.12.3.349 | |
| Lagerberg 2021 (TOP) Prospective cohort (Oslo, TOP study; bipolar disorder in the first treatment period) | 220 enrolled; 112 completed follow-up Adults with bipolar I, II or not otherwise specified bipolar disorder receiving their first treatment (Norway) | Harmful cannabis use (above the DUDIT threshold) | 1 year | Harmful cannabis use fell from 29% to 15%. The course of cannabis use (none, stopping, continuing) did not independently predict the risk of a new episode; a borderline trend in the crude analysis (p=0.058) disappeared after adjustment for age at onset, alcohol and smoking. Continued harmful alcohol use, however, increased the risk of a new episode. In an earlier, smaller analysis of the same cohort, continued cannabis use was associated with a higher mania score and lower functioning at 1 year (Kvitland 2015) | PMID 34017275, doi:10.3389/fpsyt.2021.656912, Earlier analysis of the same Oslo TOP cohort: Kvitland 2015, PMID 25651990, continued use associated with the mania score measured with the YMRS and functioning measured with the GAF) | |
| Olivier 2025 Prospective cohort (Barcelona, Hospital Clínic acute psychiatric unit; admissions 2015–2019) | 279 Patients with bipolar disorder hospitalized for a manic or mixed episode (Spain) | Cannabis use | 3 years | Cannabis use was not associated with emergency visits or rehospitalization over 3 years (this association was seen only for tobacco); but it went together with earlier onset of illness, a current episode with manic polarity, psychotic symptoms and a higher likelihood of stopping treatment. The pattern of use was not recorded in detail; follow-up was incomplete for those who moved out of the area or to the private sector | PMID 40221053, doi:10.1016/j.jad.2025.04.026 | |
| Smith 2025 (Prechter cohort) Long-term prospective cohort (Prechter Longitudinal Study of Bipolar Disorder, US); within-person temporal model | 788 (bipolar I 565, bipolar II 162, not otherwise specified 61) Adults with bipolar spectrum disorder (67% women, mean age 39.8) | Cannabis use frequency and intensity (modified AUDIT; self-report) | Median 72 months of participation; substance use measured every 6 months, suicidal ideation every 2 months | When a person's cannabis use rose above their own average, suicidal ideation was slightly higher at the next measurement (time-lagged coefficient 0.042; 95% credible interval 0.020–0.063; small effect); suicidal ideation did not predict later cannabis use. In the concurrent analysis, there was no association between cannabis and suicidal ideation. When alcohol and substance use disorder diagnoses were added, the cannabis model could not be estimated (did not converge); with these diagnoses removed, age, sex and bipolar subtype did not change the association. Suicidal ideation was measured with a single PHQ-9 item (thoughts that you would be better off dead or of hurting yourself) | PMID 39793525, doi:10.1016/j.psychres.2025.116357 | |
| Tyler 2015 Experience sampling study (diary entries at 10 random times a day for 6 days) | 29 enrolled; 24 analyzed Adults with bipolar I or II disorder, outside an episode, who used cannabis at least 2 times a week (England) | Cannabis (their own use in daily life) | 6 days | The likelihood of using cannabis increased at moments of high positive affect (OR 1.25; 95% CI 1.06–1.47); negative affect, manic or depressive symptoms did not predict cannabis use. Cannabis use was followed by increases in positive affect (β = 0.35), manic symptoms (β = 0.20) and depressive symptoms (β = 0.17). According to the authors, there is no evidence of use to 'self-medicate' depression in daily life. Very small sample, short duration | PMID 25738578, doi:10.1371/journal.pone.0118916, Correction: PLoS One 2015, 10(3):e0123953, PMID 25822166 | |
| Young-Wolff 2026 Retrospective cohort (Kaiser Permanente Northern California electronic health records; 2016–2023) | 463,396 adolescents (26,345 reported cannabis use in the past year) Adolescents aged 13–17 asked about cannabis use at a routine well-child visit (US) | Cannabis use in the past year (self-reported on a confidential screening form) | Mean follow-up 3.7 years (up to age 25 at most) | Cannabis use was associated with an increased risk of a later new diagnosis of bipolar disorder (HR adjusted for sex, ethnicity, deprivation, insurance, alcohol and other substance use 2.01; 95% CI 1.82–2.22); when other mental disorders at baseline were also taken into account, HR 1.73 (1.57–1.90). Observational; cannabis use self-reported, diagnoses from records | PMID 41719031, doi:10.1001/jamahealthforum.2025.6839 |
Reported side effects
- Pinto 2024: no more side effects than placebo, no difference in the development of mania
- A single case of mania after high-dose vaped CBD (Auf 2025)
- THC (1970s, hospitalized patients with depression): 10 of 21 patients could not continue because of side effects such as intense anxiety, a sense of detachment from oneself, panic or psychosis-like reactions (Kotin 1973, Ablon 1974; details from a secondary source, Stanciu 2021)
Monitoring done in studies
- Monitoring by a doctor for symptoms of mania (monitored in the studies)
Who should be careful
- THC-containing products are in the group to be avoided in bipolar disorder (Kansagara 2026)
- In the general population, cannabis use is associated with psychosis, mania, depression and suicide attempts (observational; Solmi 2023)
- Safety data for pure CBD are limited and rest on a single small RCT
- Mood stabilizers/antipsychotics should not be stopped on one's own
- In bipolar disorder, continued cannabis use is associated in most follow-up studies with less recovery and more frequent episodes (observational; Zorrilla 2015, Strakowski 2007); the CANMAT guideline group recommends avoiding it (Tourjman 2023)
Drug interactions
- Citalopram: in healthy adults, CBD (2.5 mg/kg twice a day, 12 days) increased citalopram exposure (AUC) by 43% (Salcedo 2026, PMID 41622700)
- Amitriptyline: a single 30 mg dose of CBD increased amitriptyline AUC by 13% and Cmax by 17%; the effect may be larger with continuous use (Gorbenko 2026, PMID 41398552)
- In an emergency department survey, 69.7% of people using CBD at home were taking a medicine with interaction potential; the most common group was antidepressants (64.4%) (Geneau 2025, PMID 41227172)
- Reports of interactions with antidepressants, antiepileptics, opioids and alcohol were reviewed (Balachandran 2021, PMID 33515191)
- Lithium: a single case of lithium toxicity developing weeks after CBD was added (Singh 2020, PMID 32851114)
Abbreviations: RCT, randomized controlled trial (a study in which participants are assigned to groups by chance); placebo, an inactive substance; SMD, OR and RR, effect measures used in analyses that pool studies; CYP, liver enzymes that break down medicines.
If you are considering CBD
- Do not stop or reduce your mood stabilizer or antipsychotic medicines on your own; CBD does not replace them.
- If you take lithium or an antidepressant, do not add CBD without talking to your doctor; medicine levels may change.
- Stay away from products that contain THC; check the label and the certificate of analysis to see whether the product contains THC.
- If you use CBD, tell your doctor about any unusual changes in your mood; you may need to be monitored for symptoms of mania.
- If you feel very unwell or have thoughts of harming yourself, contact your doctor or the nearest emergency department without delay.
Choose your age group, medicines and situation to prepare a short summary you can take to your doctor. Nothing you enter is sent anywhere.
Bottom line
The only study of CBD in bipolar depression showed no result different from a dummy pill; results from a larger study are awaited. Cannabis containing THC, on the other hand, was linked with more frequent episodes in most studies that followed people with bipolar disorder; this association is observational. Throughout this process, the most important thing is to continue your treatment and to discuss every change with your doctor.
Frequently asked questions
Does CBD oil work for bipolar depression?
So far, no results show this. In the only small study, depression scores in patients who added CBD to their current treatment dropped to a similar degree as in those taking a dummy pill (placebo). A larger study of 360 people is underway (according to its registry entry, it is not expected to finish before the end of 2029); its results are needed before anything definite can be said.
Can CBD trigger a manic episode?
In the bipolar depression study, there was no increase in the development of mania among people taking CBD. However, a case was reported of a person with no psychiatric history who developed mania with a high dose of a product of unverified purity. In two patients in a manic episode, CBD was found to be ineffective.
Can I use CBD with lithium or my mood medicine?
Always make this decision with your doctor. CBD can change the blood levels of some medicines; a case was reported of a patient taking lithium who developed lithium toxicity after CBD was added. Do not stop or reduce your medicines on your own.
Can THC-containing products be used in bipolar disorder?
A 2026 review stated that THC-dominant cannabis can worsen manic symptoms and functioning in bipolar disorder and should therefore be avoided. In the general population, cannabis use has been linked with psychosis, mania, depression and suicide attempts; these findings are observational. In two small, old studies in which THC was tested as a medicine in patients with depression (some of them bipolar), depression did not improve and about half of the patients could not continue treatment because of side effects such as intense anxiety or psychosis-like reactions. In most studies that followed people with bipolar disorder over a long time, those who kept using cannabis also recovered less and had new episodes more often; these studies are observational too. An expert guideline group in Canada, based on its systematic review, also recommends avoiding cannabis in bipolar disorder.
Can CBD oil be obtained with a prescription for bipolar depression in Turkey?
In Turkey, products derived from cannabis are regulated by the Kenevirden Elde Edilen Ürünlere Dair Yönetmelik (Regulation on Products Derived from Hemp) dated January 31, 2026. These products, including CBD, can be sold only in pharmacies. Those classed as medicines and health products are dispensed with a prescription; the regulation does not state a prescription requirement for support products and cosmetics with a THC content of at most 0.3%. As of October 2026, there is no approved (licensed) CBD medicine in Turkey. If you are thinking of using it, talk to your doctor first.
References
Show 23 references
- Pinto 2024. Double-blind, placebo-controlled pilot RCT (ended early because of the COVID-19 pandemic). PMID 37920963, doi:10.1177/07067437231209650, NCT03310593
- Zuardi 2010. Case report (2 patients; inpatient; 5 days of placebo each at the start and end). PMID 18801823, doi:10.1177/0269881108096521
- Auf 2025. Case report. PMID 40630971, doi:10.64719/pb.4542
- NCT05867849. Phase 3, double-blind, placebo-controlled RCT (ongoing). NCT05867849
- Sanjuan-Ortiz 2027. Systematic review and meta-analysis (19 longitudinal observational studies; PubMed and Web of Science, up to October 2025). PMID 42838231, doi:10.1016/j.jad.2026.122588
- Kotin 1973. Double-blind, placebo-controlled small study (inpatient; 7 days placebo, 7 days THC, 7 days placebo). PMID 4569675, doi:10.1001/archpsyc.1973.01750330041007
- Ablon 1974. Double-blind, placebo-controlled small study (inpatient; THC or placebo after 3 weeks of observation). PMID 4592555, doi:10.1176/ajp.1974.131.4.448
- Miranda 2026. Randomized, placebo-controlled single-dose THC laboratory study; also a cross-sectional comparison of cannabis users and non-users. PMID 42481733, doi:10.1038/s41386-026-02489-w, NCT04231643, bipolar patients and healthy volunteers, in the registry, a single oral dose of 5 mg dronabinol, 600 mg CBD or placebo, ended at 19 people for budget reasons, no results in the registry), the abstract gives no registry number, so the match could not be verified
- Wilson 2026. Systematic review and meta-analysis of RCTs (1980 to May 13, 2025; PROSPERO CRD42023392718). PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged)
- Stanciu 2021. Systematic review of controlled studies (1970 to February 5, 2020; 8 databases; no meta-analysis). PMID 33530732, doi:10.1176/appi.ps.202000189
- Sorkhou 2024. Systematic review (inception to November 2023; EMBASE, PsycINFO, PubMed, MEDLINE; no meta-analysis). PMID 38655516, doi:10.3389/fpubh.2024.1346207
- Tourjman 2023 (CANMAT). Systematic review and clinical recommendation (Canadian CANMAT task force; inception to October 2020; PubMed, Embase, PsycINFO, CINAHL, Cochrane; GRADE; no meta-analysis). PMID 35711159, doi:10.1177/07067437221099769
- Bartoli 2019. Systematic review and meta-analysis of observational studies (up to July 2018). PMID 31121199, doi:10.1016/j.neubiorev.2019.05.017
- Solmi 2023. Umbrella review (meta-analyses of RCTs and observational studies; up to February 9, 2022). PMID 37648266, doi:10.1136/bmj-2022-072348
- Kansagara 2026. Clinical (narrative) review; not a systematic review. PMID 41801216, doi:10.1001/jamainternmed.2025.8215
- Zorrilla 2015 (EMBLEM). Prospective observational cohort (EMBLEM; Europe, multicenter). PMID 25430820, doi:10.1111/acps.12366, 12-month analysis of the same EMBLEM cohort: van Rossum 2009, PMID 19155808, PMID 20458676
- Strakowski 2007. Inception cohort (prospective follow-up after a first hospitalization for mania). PMID 17199055, doi:10.1001/archpsyc.64.1.57
- Kim 2015 (BCOS). Prospective observational cohort (Australia, Bipolar Comprehensive Outcomes Study; 9 assessments over 24 months). PMID 26207128, doi:10.4306/pi.2015.12.3.349
- Lagerberg 2021 (TOP). Prospective cohort (Oslo, TOP study; bipolar disorder in the first treatment period). PMID 34017275, doi:10.3389/fpsyt.2021.656912, Earlier analysis of the same Oslo TOP cohort: Kvitland 2015, PMID 25651990, continued use associated with the mania score measured with the YMRS and functioning measured with the GAF)
- Olivier 2025. Prospective cohort (Barcelona, Hospital Clínic acute psychiatric unit; admissions 2015–2019). PMID 40221053, doi:10.1016/j.jad.2025.04.026
- Smith 2025 (Prechter cohort). Long-term prospective cohort (Prechter Longitudinal Study of Bipolar Disorder, US); within-person temporal model. PMID 39793525, doi:10.1016/j.psychres.2025.116357
- Tyler 2015. Experience sampling study (diary entries at 10 random times a day for 6 days). PMID 25738578, doi:10.1371/journal.pone.0118916, Correction: PLoS One 2015, 10(3):e0123953, PMID 25822166
- Young-Wolff 2026. Retrospective cohort (Kaiser Permanente Northern California electronic health records; 2016–2023). PMID 41719031, doi:10.1001/jamahealthforum.2025.6839
Points that could not be verified from the full text
- NCT05457465 (anxiety in bipolar disorder; single-arm, open-label pilot, 4 weeks, target 25) is recruiting (registry updated March 30, 2026; estimated completion December 2026); the product tested is a CBD-rich solution derived from hemp that, according to the registry, contains no THC; people in an acute depressive or manic episode are not enrolled; there is no control group
- Kotin 1973 and Ablon 1974 (THC, hospitalized patients with depression): the original articles could not be read; details were taken from the Stanciu 2021 review, and the review's text and table give the dose information for the two studies the other way round
- Miranda 2026 (single-dose THC in bipolar disorder): the full text could not be read (it will become open on PMC in October 2027); the number of participants, dose and route of administration are not in the abstract; it could not be verified whether it is the same study as the same center's ClinicalTrials.gov entry (NCT04231643: single oral dose of 5 mg dronabinol, CBD or placebo; 19 people; ended early for budget reasons)
Research history
Here we record what changed on this page, when and why, with dates. The newest change is at the top.
The last sentence of the page's short summary was rewritten in a calmer, more guiding tone: instead of "Talk to your doctor before you start.", it now says whom you can talk with, and about what, if you are considering CBD. The conclusion and the strength of evidence did not change.
The correction to the Wilson 2026 meta-analysis published in the journal was read: only the axis of one figure and the study counts in a table were corrected; the results on this page did not change.
Related studies: Wilson 2026, published April 2026 (PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged))
Display fix: internal working file names left in the text were removed; content and evidence did not change.
Older changes (6)
For THC-containing products, the strength of evidence became 'Very weak' instead of 'Insufficient.' The October 2026 search found two small double-blind studies in which THC was tested in hospitalized patients with depression (21 people, 9 of them bipolar): no benefit, frequent and severe distressing reactions (original articles could not be read, information from a review). The main basis for harm is observational: 5 of 7 studies following the course of illness in bipolar disorder found a worse course in cannabis users, and 2 found no independent association; two systematic reviews, a meta-analysis and Canada's CANMAT guideline group point in the same direction and recommend avoiding cannabis. Because these findings do not show cause and effect, the evidence was not rated 'Weak.' For pure CBD, the conclusion and the strength of evidence did not change.
Studies that led to the change: Kotin 1973, published March 1973 (PMID 4569675, doi:10.1001/archpsyc.1973.01750330041007); Ablon 1974, published April 1974 (PMID 4592555, doi:10.1176/ajp.1974.131.4.448); Sorkhou 2024, published April 9, 2024 (PMID 38655516, doi:10.3389/fpubh.2024.1346207); Bartoli 2019, published August 2019 (PMID 31121199, doi:10.1016/j.neubiorev.2019.05.017); Tourjman 2023 (CANMAT), published May 2023 (online June 16, 2022) (PMID 35711159, doi:10.1177/07067437221099769)
PubMed, Cochrane and ClinicalTrials.gov were systematically searched again in October 2026 (1235 unique records in PubMed, 47 in ClinicalTrials.gov). No new controlled study of CBD in bipolar depression was found. Added: two small controlled studies from the 1970s in which THC was tested in patients with depression (some of them bipolar) (original articles could not be read, information from a review), a single-dose THC laboratory study in bipolar disorder, five systematic reviews or meta-analyses, one of them by Canada's CANMAT guideline group, an umbrella review, the clinical review cited as a source in the text, seven studies following the link between cannabis use and the course of bipolar disorder, and a large adolescent cohort; the conclusion text for THC-containing products was expanded accordingly. Corrected: the duration, dose increase and side effect details of the mania case report; the pilot study's reason for early termination (COVID-19 pandemic), number of participants, dose-increase schedule, product purity and the authors' conflict-of-interest statement (from the full text); the registry status, centers and dates of the ongoing phase 3 study; that the product tested in the anxiety pilot contains no THC according to the registry; the meta-analysis's online publication date and how many studies were pooled. For CBD, the conclusion and the strength of evidence did not change; for THC-containing products, the strength of evidence rose from 'Insufficient' to 'Very weak' (separate note).
Related studies: Kotin 1973, published March 1973 (PMID 4569675, doi:10.1001/archpsyc.1973.01750330041007); Ablon 1974, published April 1974 (PMID 4592555, doi:10.1176/ajp.1974.131.4.448); Miranda 2026, published October 2026 (online July 21, 2026) (PMID 42481733, doi:10.1038/s41386-026-02489-w, NCT04231643, bipolar patients and healthy volunteers, in the registry, a single oral dose of 5 mg dronabinol, 600 mg CBD or placebo, ended at 19 people for budget reasons, no results in the registry), the abstract gives no registry number, so the match could not be verified); Wilson 2026, published April 2026 (PMID 41856154, doi:10.1016/S2215-0366(26)00015-5, Correction: Lancet Psychiatry 2026, 13(8):e11, PMID 42302809, 2026, the axis of a figure and the study counts in the table were corrected, results unchanged)); Stanciu 2021, published April 2021 (PMID 33530732, doi:10.1176/appi.ps.202000189); Sorkhou 2024, published April 9, 2024 (PMID 38655516, doi:10.3389/fpubh.2024.1346207); Bartoli 2019, published August 2019 (PMID 31121199, doi:10.1016/j.neubiorev.2019.05.017); Solmi 2023, published August 30, 2023 (PMID 37648266, doi:10.1136/bmj-2022-072348); Kansagara 2026, published May 2026 (PMID 41801216, doi:10.1001/jamainternmed.2025.8215); Zorrilla 2015 (EMBLEM), published February 2015 (PMID 25430820, doi:10.1111/acps.12366, 12-month analysis of the same EMBLEM cohort: van Rossum 2009, PMID 19155808, PMID 20458676); Strakowski 2007, published January 2007 (PMID 17199055, doi:10.1001/archpsyc.64.1.57); Kim 2015 (BCOS), published July 2015 (PMID 26207128, doi:10.4306/pi.2015.12.3.349); Lagerberg 2021 (TOP), published May 4, 2021 (PMID 34017275, doi:10.3389/fpsyt.2021.656912, Earlier analysis of the same Oslo TOP cohort: Kvitland 2015, PMID 25651990, continued use associated with the mania score measured with the YMRS and functioning measured with the GAF)); Olivier 2025, published August 2025 (PMID 40221053, doi:10.1016/j.jad.2025.04.026); Smith 2025 (Prechter cohort), published February 2025 (PMID 39793525, doi:10.1016/j.psychres.2025.116357); Tyler 2015, published March 4, 2015 (PMID 25738578, doi:10.1371/journal.pone.0118916, Correction: PLoS One 2015, 10(3):e0123953, PMID 25822166); Young-Wolff 2026, published February 6, 2026 (PMID 41719031, doi:10.1001/jamahealthforum.2025.6839); Tourjman 2023 (CANMAT), published May 2023 (online June 16, 2022) (PMID 35711159, doi:10.1177/07067437221099769)
A new observational meta-analysis (Sanjuan-Ortiz 2027) examining the link between cannabis use and later bipolar disorder and mania was added. For CBD, the conclusion and the strength of evidence did not change.
The names of the finding labels were simplified: on this page the "No benefit seen" label became "No effect found." The assessment did not change.
The full texts of the studies were reread and the information on the page was checked against its sources. Corrected: duration and dose details in the study table. Points that could not be verified from the full text went down from 2 items to 1 item. The year of 3 publications named on the page was corrected to the journal's print year instead of the online publication year. The conclusion and the strength of evidence did not change.
Page published. This first review examined 4 studies.