Key Point: Antidepressants posed "negligible" risk of inducing mania in bipolar depression, in a target trial emulation with a larger cohort than in RCTs that have assessed this effect.
The risk of antidepressants inducing a switch to mania in patients with bipolar depression was found to be "negligible", in a 1-year nationwide target trial emulation.1 The study cohort was larger than in randomized controlled trials (RCTs) that have assessed the effect, and the length of follow-up longer than in most.
"Although mania occurs frequently among patients with bipolar depression treated with antidepressants, our findings imply that this may not be caused by the antidepressants but rather by the recurrent nature of bipolar disorder," indicate Christopher Rohde, MD, PhD, and colleagues, Departments of Clinical Medicine and of Affective Disorders, Aarhus University, Aarhus, Denmark.
While the investigators acknowledge limits on generalizing from their findings, they tout the target trial emulation and use of nationwide Danish health registers as a means to overcome the high costs and other challenges of conducting a comparable RCT. These include the requirement for high numbers of patients to be retained for statistical power throughout a sufficiently long term for mania to arise in the course of bipolar depression with and without antidepressant treatment.
"This study demonstrates the power of using nation-wide register-based data for target trial emulations," said Rohde et al. "With a total sample size of 979 patients, this study was much larger than any of the previous antidepressant treatment trials with patients with bipolar depression, and it achieved good statistical power."
In an accompanying editorial,2 Natalie Gottlieb, PhD, and Allan Young, FRCPsych, Institute of Psychiatry, Psychology, and Neuroscience, King's College London, London, UK, agree with the investigators that using longitudinal observation data from the health registers, within comprehensive eligibility criteria and adjustment for baseline confounders, did overcome some shortfalls of the RCTs, but note that it poses other limitations.
"Data that are not collected for research purposes may include inconsistencies, such as ambiguous medical coding and a lack of standardized measurements," Gottlieb and Young point out. "Additionally, they are unable to emulate RCTs with regard to placebo control or tight monitoring of treatment adherence."
Target Trial Emulation
Rohde et al drew on data from health registries to identify 979 patients 18 years or older who were discharged from their first psychiatric admissions with bipolar depression, in the period from January 1, 1996, through March 1, 2018. They excluded those who were previously diagnosed with bipolar depression or had used an antidepressant in the 2 years preceding the admission, as well as patients with previous diagnosis of schizophrenia or schizoaffective disorder.
The antidepressant treatment group comprised 358 of the patients, identified for having filled a prescription for an antidepressant within 2 weeks after discharge. Among these, 181 were initially prescribed a selective serotonin reuptake inhibitor, 51 a tricyclic antidepressant, 41 a serotonin and norepinephrine reuptake inhibitors, and 85 received another antidepressant. They found that those treated with an antidepressant were more likely to have the index admission in an earlier calendar year, to receive a mood-stabilizing agent, to have severe (vs mild or moderate) bipolar depression, to have a prior diagnosis of unipolar depression, to be married and to be employed.
The occurrence of mania over 1 year was compared between groups with or without antidepressant treatment, with adjustment for baseline covariables to emulate randomized open-label treatment allocation.
Rohde et al reported no statistically significant associations between treatment with an antidepressant and the risk of mania (hazard rate ratio 1.08, 95% CI, 0.72-1.61); or in subgroup analyses of those treated with or without a concomitant mood stabilizing agent.In a secondary measure of the possible contribution of antidepressants to rapid cycling, there was no statistically significant association between antidepressant treatment and bipolar depression recurrence.
The investigators acknowledged that the internal validity of this trial emulation is less than that of an ideally conducted double-blind RCT for several reasons, including less capacity to control for baseline confounders. In the baseline comparison, for example, they found that patients not treated with antidepressants more likely to have a more severe course of illness marked by such events as repeat admissions, increased outpatient contacts, and manic episodes.
"This relationship may very well reflect clinicians' caution against prescribing antidepressants for patients with a higher probability of mania," they posited.
"Indeed, a major limitation of the present study is the lack of direct information on factors that influence treatment decisions, particularly the severity of the previous and current mood episodes," Rohde et al observe.
Gottlieb and Young note other considerations, including the difficulty in distinguishing between bipolar I and II in constituting the cohort of this trial emulation. "This distinction is of crucial interest because there is evidence that these groups may respond to antidepressants very differently," they point out.
Both the investigators and the commentators agree that additional studies are required to optimize treatment strategies for individuals with bipolar depression. "It remains clear, however, that new efficacious treatments for bipolar depression are urgently needed," Gottlieb and Young state.
Note: This article originally appeared on Psychiatric Times.
Comments