Bipolar disorder (BD) is a disabling condition characterised by episodes of mania or hypomania, alternating or co-occurring with depression (American Psychiatric Association, 2013; GBD, 2019).
Research suggests that depressive symptoms are predominant in BD and may present greater burden compared to elevated features (Judd et al., 2002; Judd et al., 2003; Miller et al., 2014). Nevertheless, the use of antidepressants for bipolar depression is debatable, as their safety and efficacy remain uncertain (McGirr et al., 2016; Pacchiarotti et al., 2013; Sachs et al., 2007).
As highlighted in a Mental Elf blog post by Prof Joseph Hayes, the National Institute for Health and Care Excellence (NICE) 2014 guidelines for BD (NICE, 2014) indicate that fluoxetine may be preferable to other antidepressants. The British Association for Psychopharmacology 2016 guidelines suggest that antidepressants in BD may be effective, but only in combination with mood stabilisers (Goodwin et al., 2016).
Evidence suggests that the risk of mania may be greater for tricyclic antidepressants (TCA), compared to selective serotonin reuptake inhibitors (SSRIs) (Goodwin et al., 2016; Tondo et al., 2010). Although findings from randomised controlled trials (RCTs) are conflicting, there is some evidence that prolonged treatment with antidepressants may worsen manic or hypomanic symptoms in BD (Ghaemi et al., 2021; McGirr et al., 2016; Yatham et al., 2023).
A recent target trial emulation by Rohde et al. 2024, aimed to shed some light on the risk of antidepressant-induced mania in people with bipolar depression.
Methods
Rohde et al.â€s (2024) study employed a ‘target trial emulation†model (Matthews et al., 2022), whereby they were able to leverage data accumulated by Danish national registers while incorporating core characteristics of the gold standard in evidence-based medicine: the randomised controlled trial (RCT).
In a target trial emulation, pre-existing observational data is compiled and manipulated such that the principles that govern a traditional RCT are upheld: participants are identified, screened according to rigorous criteria and then the data is stratified as a proxy for randomisation. Authors assessed mania as psychiatric admission with a primary diagnosis of a manic or hypomanic episode.
Results
The final study cohort consisted of 979 patients with bipolar depression, 36.6% (n = 358) of whom received treatment with antidepressants. The sex and age characteristics were comparable between the antidepressant and non-antidepressant groups.
Interestingly, of those treated with antidepressants, only 62.6% used them concurrently with mood stabilisers. Among patients in the antidepressant group, around 50.5% (n=181) received an SSRI, 14.3% (n=51) a TCA and 11.5% (n=41) a serotonin and norepinephrine reuptake inhibitors (SNRIs). The remaining 85 patients were treated with another type of antidepressant.
Rohde et al (2024) claim that considering the effect size of a hazard ratio of 1.77 (based on a meta-analysis by McGirr et al., 2016), their study achieved power of 90%.
- The fully adjusted models which included the entire sample, showed no statistically significant associations between antidepressant treatment and mania or hypomania, (hazard rate ratio=1.08, 95% CI=0.72 to 1.61) (a hazard ratio of 1 indicates the lack of an association).
- Antidepressant treatment was not significantly associated with risk of mania or hypomania
- among those treated with mood stabilisers (hazard rate ratio=1.16, 95% CI=0.63 to 2.13)
- or individuals who did not receive a mood stabiliser (hazard rate ratio=1.16, 95% CI=0.65 to 2.07).
- Secondary analyses indicated that the risk of bipolar depression recurrence was not associated with antidepressant use (hazard ratio=0.91, 95% CI=0.65 to 1.27).
Conclusions
This was a target trial emulation which used observational data from Danish health registers to assess the risk of mania in bipolar depression following antidepressant use over a period of 1 year.
Authors concluded: “findings suggest that the risk of antidepressant-induced mania is negligible†and highlighted the need for further research. However, evidence from other studies suggests that prolonged treatment with antidepressants is linked with increased risk of manic or hypomanic symptoms in BD (McGirr et al., 2016; Yatham et al., 2023; Tondo et al., 2010).
Although the study by Rohde and colleagues (2024) improved our understanding of the so-called “mood switch†following treatment with antidepressants in BD, given its methodological limitations and the conflicting findings from the literature, further research on this topic is warranted.
Strengths and limitations
Rohde et al.â€s implementation of a target trial emulation model provided them with some notable methodological benefits:
- Large sample size: their use of Danish national registers provided them with a final cohort of 979 – this, they note, meant their study was larger than any of the previous antidepressant trials with bipolar depression patients.
- Extended ‘follow-up†period: this study followed patients for a year, unless either readmission or death occurred. This represents one of the longest follow-up periods for a study of this issue.
However, despite the various upsides to their study design, there remain numerous limitations and drawbacks. After all, while target trial emulations seek to approximate the scientific rigour of RCTs by applying their principles to observational data, they are not a perfect substitute for properly controlled ‘live†studies.
- Restricted management of participants: Rohde et al. were able to assign participants to conditions, but could not restrict their course of treatment. More than a quarter of those initially assigned to the non-antidepressant condition in this trial eventually started a course of antidepressants.
- Restricted utility of findings: Rohde et al.â€s data did not facilitate distinctions between bipolar I and II disorders. This means that their study cannot inform as to whether there is a difference in rates of antidepressant-induced [hypo]mania between the two types.
- Reduced generalisability: the eligibility criteria for this study excluded participants who experienced an episode without being hospitalised, thereby limiting the studyâ€s relevance to patients who have been admitted at least once.
Implications for practice
The authors explain how their findings indicate that the risk of antidepressants causing mania in bipolar disorder patients is negligible, and that while common in patients treated with antidepressants, manic episodes are potentially simply a consequence of the recurrent nature of the disorder and not a side effect of treatment. Moreover, the results of their emulation study support the notion that antidepressants do not necessarily cause mania.
As data continues to accumulate in this area, we may eventually witness a shift in clinicians†attitudes towards the prescription of antidepressants as treatment for bipolar depression, especially in the short-term. However, current guidelines for bipolar disorder do not recommend monotherapy of antidepressants (Goodwin et al 2016). Currently, as noted by Rohde et al., clinicians are particularly cautious when considering prescribing antidepressants for patients with bipolar disorder, given the prevailing view that there is an increased risk of mania. Rohde et al. suggest that this caution may be reflected in their data, with their findings indicating a more severe clinical course among patients that were not treated with an antidepressant. This cohort, on average, saw a higher number of hospital admissions, outpatient contacts and episodes of mania than those in their antidepressant-using counterparts.
While the findings of Rohde et al. do serve to bolster a growing body of evidence against the notion that antidepressants are necessarily at fault for increased rates of manic episodes, they are not conclusive. This is due in no small part to the limitations of the study described previously. To this end, we believe that further research is merited in order to fully understand whether:
- participants who refrain from taking antidepressants for the entirety of a future study (including follow-up) fare better or worse than those that are prescribed with antidepressants for the duration of a trial;
- distinctions between bipolar I and II are further reflected in terms of their response to antidepressant treatment; and
- the severity of manic symptoms is, on average, greater among those taking antidepressants – regardless of whether they have been or are hospitalised due to a bipolar episode.
Once these issues are addressed, not only will we know with greater confidence the overall impact of antidepressant treatment within the context of bipolar disorder, but clinicians will be more informed as to whether a given patient could stand to benefit from such a course of treatment, based on their unique circumstances (e.g., type of bipolar disorder and recent symptom severity).
Statement of interests
Paul Leeks declares no conflicts of interest. Michail Kalfas has received honoraria from Neurocentrx Pharma.
Links
Primary paper
Rohde C, Østergaard SD, Jefsen OH. (2024). A Nationwide Target Trial Emulation Assessing the Risk of Antidepressant-Induced Mania Among Patients With Bipolar Depression. The American Journal of Psychiatry 181(7) 630–638.Â
Other references
American Psychiatric Association. (2013) Diagnostic and statistical manual of mental disorders (5th ed.).
Bobo WV. (2017) The Diagnosis and Management of Bipolar I and II Disorders: Clinical Practice Update. Mayo Clinic proceedings, 92(10), 1532–1551.
GBD 2019 Mental Disorders Collaborators. (2022) Global, regional, and national burden of 12 mental disorders in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry 9(2) 137-150.
Ghaemi SN, Whitham EA, Vohringer PA. et al (2021) Citalopram for Acute and Preventive Efficacy in Bipolar Depression (CAPE-BD): A Randomized, Double-Blind, Placebo-Controlled Trial. The Journal of clinical psychiatry 82(1) 19m13136.
Goodwin GM, Haddad PM, Ferrier IN. et al (2016) Evidence-based guidelines for treating bipolar disorder: Revised third edition recommendations from the British Association for Psychopharmacology. Journal of psychopharmacology (Oxford, England) 30(6) 495–553.
Judd LL, Akiskal HS, Schettler PJ. et al (2003) A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Archives of general psychiatry 60(3) 261–269.
Judd LL, Akiskal HS, Schettler PJ. et al (2002) The long-term natural history of the weekly symptomatic status of bipolar I disorder. Archives of general psychiatry 59(6) 530-537.
Matthews A A, Danaei G, Islam N. et al (2022) Target trial emulation: applying principles of randomised trials to observational studies BMJ 378:e071108 doi:10.1136/bmj-2022-071108
McGirr A, Vöhringer PA, Ghaemi SN. et al (2016) Safety and efficacy of adjunctive second-generation antidepressant therapy with a mood stabiliser or an atypical antipsychotic in acute bipolar depression: a systematic review and meta-analysis of randomised placebo-controlled trials. Lancet Psychiatry 3(12) 1138-1146.
Miller S, Dell’Osso B, Ketter TA. (2014) The prevalence and burden of bipolar depression. Journal of affective disorders 169(1) S3-11.
National Institute for Health and Care Excellence [NICE] (2014)Â Bipolar Disorder: assessment and guidance. NICE clinical guideline CG185.
Pacchiarotti I, Bond DJ, Baldessarini RJ. et al (2013) The International Society for Bipolar Disorders (ISBD) task force report on antidepressant use in bipolar disorders. The American journal of psychiatry 170(11) 1249–1262.
Sachs GS, Nierenberg AA, Calabrese JR. et al (2007) Effectiveness of adjunctive antidepressant treatment for bipolar depression. The New England journal of medicine 356(17) 1711–1722.
Tondo L, Vázquez G, Baldessarini RJ. (2010) Mania associated with antidepressant treatment: comprehensive meta-analytic review. Acta psychiatrica Scandinavica 121(6) 404–414.
Yatham LN, Arumugham SS, Kesavan M. et al (2023) Duration of Adjunctive Antidepressant Maintenance in Bipolar I Depression. The New England journal of medicine 389(5) 430–440.