The exact recurrence rate of bipolar disorder in patients receiving lithium maintenance phase treatment and the modifiers associated with recurrence are still unknown.
We searched Embase, PubMed, and CENTRAL from inception until April 28, 2020. Outcomes included recurrence rate of any mood episode, depressive episodes, and manic/hypomanic/mixed episodes; all-cause discontinuation rate; and discontinuation rate due to adverse events. A random-effects model, single-group summary meta-analysis was conducted. A meta-regression analysis to examine whether the modifiers (total number of patients, %female, mean age, duration of study, duration of preliminary phase, publication year, bipolar disorder type, mood status at recruitment, presence of a placebo arm, sponsorship, enrichment design, number of treatment arms, and risk of bias for blinding or randomization) were associated with the event rate of the outcomes was also performed.
We identified 21 randomized trials (n = 1,415; mean study duration, 78.40 ± 32.10 weeks; %female, 54.85%; mean age, 43.47 ± 4.88 years). The event rates (95% confidence interval [CI]) were as follows: recurrence of any mood episode, 39.8% (32.8%, 47.1%); depressive episodes, 25.6% (18.8%, 34.0%); manic/hypomanic/mixed episodes, 18.5% (13.7%, 24.7%); all-cause discontinuation rate, 67.0% (57.2%, 75.5%); and discontinuation rate due to adverse events, 8.7% (5.1%, 14.7%). After adjusting for multiple testing, our meta-regression analysis showed association only between the all-cause discontinuation rate and presence of a placebo arm.
The recurrence rate of depressive episodes seemed to be higher than the recurrence rate of manic/hypomanic/mixed episodes. The all-cause discontinuation rate was high. However, the studies included in our meta-analysis were of short duration.

Author