Nine trials were included in the review (1,661 patients).
A statistically significant difference, favouring second-generation antipsychotics over mood stabilisers, was found for change in symptom severity ratings (SMD -0.22, 95% CI -0.32 to -0.12; Ι²=0%; nine trials). Small-sample bias was not evident on the corresponding funnel plot (data not shown).
Statistically significant differences, favouring second-generation antipsychotics over mood stabilisers, were found for drop-out rates (RD -0.05, 95% CI -0.10 to -0.01; Ι²=23%; nine trials) and responder rates (RD -0.07, 95% CI -0.13 to -0.01; Ι²=28%; six trials). Number needed to treat for responder rates (50% reduction in patient symptom severity by endpoint) was 17 patients (95% CI 9 to 122).
In exploratory analyses, the statistically significant difference that favoured second-generation antipsychotics over mood stabilisers was maintained when analysing lithium (SMD -0.23, 95% CI -0.40 to -0.07) and valproate (SMD -0.20, 95% CI -0.36 to -0.05) separately. Statistically significant differences were also maintained with subanalyses for trials that investigated olanzapine versus mood stabilisers (SMD -0.24, 95% CI -0.38 to -0.10) and for trials that enrolled mixed mania patients (SMD -0.25, 95% CI -0.38 to -0.13). Similar trends were shown with the other subanalyses, but they were not statistically significant. No substantial heterogeneity was indicated in any of the subanalyses.