BACKGROUND This meta-analysis assessed the equipotency of carbamazepine and lithium prophylaxis in bipolar disorder.
METHOD We selected only randomised, double-blind, controlled studies comparing carbamazepine with lithium from a manual and computerised search, and subjected them to a quality inventory. Their statistical results were weighted by their quality score and combined.
RESULTS Four studies met our criteria, yielding, overall, P = 0.15. This result is not straightforward because the studies showed significant heterogeneity (P < 0.01).
CONCLUSION Differences in statistical power and sensitivity of outcome measure explain this heterogeneity and the conflicting results of the studies. Therefore, the prophylactic efficacy of carbamazepine remains questionable.