Objective: To compare antipsychotic and mood stabilizer (MS) efficacy and tolerability in youth and adults with bipolar mania.
Methods: Medline/PubMed search for studies including: (i) youth (< 18 years) or adults (> or = 18 years); (ii) bipolar I disorder; (iii) double-blind, randomized, placebo-controlled trial (DB-RPCT); (iv) < or = 12 weeks of treatment; and (v) calculable effect sizes (ES) and/or numbers needed to treat/harm (NNT/NNH) +/- 95% confidence intervals (CI). Non-overlapping 95% CIs determined significant group differences.
Results: We identified nine DB-RPCTs in youth (n = 1,609), 5 evaluating second-generation antipsychotics (SGAs) (n = 1,140) and 4 evaluating MSs (n = 469). We also identified 23 DB-RPCTs in adults (n = 6,501), 14 including SGAs (n = 3,297), 5 using haloperidol as an active comparator (n = 580), and 11 including MSs (n = 2,581). Young Mania Rating Scale scores improved significantly more with SGAs than MSs in youth (ES = 0.65, CI: 0.53-0.78 versus 0.24, CI: 0.06-0.41) and adults (ES = 0.48, CI: 0.41-0.55 versus 0.24, CI: 0.17-0.31). After excluding topiramate studies, SGAs had larger ES than MSs only in youth (ES = 0.65, CI: 0.53-0.78 versus 0.20, CI: 0.02-0.39), but not adults (ES = 0.48, CI: 0.41-0.55 versus 0.46, CI: 0.37-0.55). However, in adults SGAs had significantly larger ES regarding Clinical Global Impressions scores than MSs, even without topiramate (ES = 0.75, CI: 0.68-0.82 versus 0.24, CI: 0.07-0.41). Rates of response, remission, and discontinuation due to any reason compared to placebo were similar between medication and age groups, except for more favorable NNTs for remission with SGAs than MSs in adults after excluding topiramate. SGAs caused more weight gain than MSs in youth (ES = 0.53, CI: 0.41-0.66 versus 0.10, CI: -0.12-0.33), but not in adults (ES = 0.13, CI: 0.05-0.22 versus 0.00, CI: -0.08-0.08). However, results were heterogeneous and not significant in either age group after excluding topiramate. Nevertheless, SGA-related weight gain was significantly greater in youth than adults. In youth, SGA-related somnolence was greater than with MSs (NNH = 4.7, CI: 3.9-6.0 versus 9.5, CI: 6.3-23.5), and more likely than in adults (NNH = 7.1, CI: 6.1-8.8). Conversely, youth experienced less akathisia with SGAs than adults (NNH = 20.4, CI: 14.1-36.5 versus 10.2, CI: 8.1-13.7), likely due to lower doses/slower titration.
Conclusions: In treating mania, potentially greater short-term efficacy compared to placebo with SGAs versus MS needs to be balanced against increased adverse events, especially in youth.