Key points are not available for this paper at this time.
BACKGROUND: Bipolar disorder (BD) is a chronic mental illness that requires lifelong treatment. We investigated the comparative efficacy and safety of pharmacological interventions for bipolar maintenance, considering dose effects across different age groups. METHODS: We conducted a network meta-analysis (NMA) of randomized controlled trials (RCTs) comparing pharmacological interventions against each other or placebo for the maintenance of BD, searching for records indexed in PubMed/MEDLINE, Embase, Web of Science, and Scopus (from inception through 2026.1.12). Co-primary outcomes were the relapse rate (into an acute episode of any mood polarity) and tolerability (discontinuation due to side effects). Relapse into episodes of specific mood polarities (namely, depressive/manic/mixed polarity), acceptability (discontinuation due to any cause), and rate of specific adverse events were secondary outcomes. Confidence-In-Network-Meta-Analysis was likewise appraised. RESULTS: Forty-four RCTs involving 23 distinct treatment combinations encompassed 10,867 participants. Sensitivity analysis retaining only low-risk-of-bias studies and excluding outliers for possible effect modifiers indicated that asenapine 20 mg/day, aripiprazole 20 mg/day, 30 mg/day, quetiapine 300 mg/day, 600 mg/day and 550 mg/day, lithium 800 mg/day, lurasidone 80 mg/day, carbamazepine 400 mg/day and 650 mg/day, valproate 71-125 μg/mL, olanzapine 20 mg/day, long-acting aripiprazole 400 mg/4 weeks, long-acting risperidone 25 mg/2 weeks, aripiprazole 30 mg/day+lamotrigine 200 mg/day outperformed placebo. Several additional drugs might be efficacious, although they either emerged as outliers for the mean age of participants/proportion of females/proportion of BD-I/II patients, baseline severity of mania/depression, and trial duration. CONCLUSIONS: Our findings are consistent with previous NMAs and current guidelines, yet they expand current knowledge by concurrently appraising different drugs, doses, and age groups.
Fornaro et al. (Tue,) studied this question.