Review of: Kishi T et al, Psychol Med 2020 Oct 13:1–9; PMID: 33046156

Type of study: Meta-analysis of double-blind, randomized placebo-controlled trials

Bipolar disorder is a lifelong illness whose treatment is an ongoing challenge, complicated not least by difficulties in treatment adherence. What do we know and what can we tell our patients about how long to continue medication?

This meta-analysis looked at recurrence rates of any mood episode in bipolar disorder after 6 months of maintenance treatment vs discontinuation of medication. The authors analyzed 22 double-blind, placebo-controlled studies of second-generation antipsychotics (LAI or oral) or a mood stabilizer in adults with bipolar I or II disorder that included a maintenance phase lasting at least 12 weeks. Some studies recruited stable subjects, while others were extensions of acute treatment studies after initial stabilization. Results were also broken down by relapse into episode subtypes: depressive, manic/hypomanic/mixed, as well as all-cause discontinuation from the study.

Compared to drug discontinuation, maintenance treatment was associated with a lower recurrence rate at 6 months for all mood episodes. The relative risks (RR) for those who remained on medication were 0.61 for any episode, 0.72 for depressive episodes, 0.45 for manic/hypomanic/mixed episodes, and 0.71 for all-cause discontinuation. The raw relapse rates into any mood episode at 6 months were 32.3% with maintenance vs 52.7% with discontinuation. In subgroup analyses, oral second-generation antipsychotics had better preventative properties than lithium or lamotrigine, based on their RRs. However, all but one of the eight SGA studies (nine groups) were “enriched”—that is to say, they included only subjects who had proven to be responsive to medication—while only two of the 15 lithium studies were enriched, making direct comparisons invalid. Oral antipsychotics were superior to long-acting injectables at preventing depression, which is not surprising considering none of the FDA-approved options for bipolar depression (cariprazine, lurasidone, quetiapine, and olanzapine-fluoxetine combination) are available in injectable form.

One limitation is the possibility that some of these relapses represented withdrawal phenomena, as most studies discontinued the medication abruptly. Also, the majority of relapses occurred at the very beginning of the follow-up period, which raises the possibility that withdrawal effects contributed to the relapse risk.

TCPR’s TAKE
Long-term maintenance medication is the norm in bipolar disorder, and this study backs up that standard. When patients want to stop their primary mood stabilizer, we can advise them that they are twice as likely to stay well if they stick with it. It may be feasible to come off an augmentation agent, however, if they have been well for at least 6 months and keep the primary mood stabilizer on board (see TCPR Jan 2020 on antipsychotic maintenance). Slow discontinuation (eg, over 1–6 months) also reduces the relapse risk, particularly with lithium.