Predictors of switching from mania to depression in a large observational study across Europe (EMBLEM).

BACKGROUND: The risk of switching from mania to depression in bipolar disorder has been poorly studied. Large observational studies may be useful in identifying variables that predict switch to depression after mania and provide data on medication use and outcomes in "real world" patients.

METHOD: EMBLEM (European Mania in Bipolar Longitudinal Evaluation of Medication) is a 2-year, prospective, observational study of patients with a manic/mixed episode. Symptom severity measures included Clinical Global Impression-Bipolar Disorder scale (CGI-BP), Young Mania Rating Scale (YMRS) and 5-item Hamilton Depression Rating Scale. Switching was defined using CGI-BP mania and depression such that patients changed from manic and not depressed to depressed but not manic over two consecutive observations within the first 12 weeks of follow-up. Cox proportional hazards models identified baseline variables independently associated with switch to depression.

RESULTS: Of 2390 patients who participated in the maintenance phase (i.e. up to 24 months), 120 (5.0%) switched to depression within the first 12 weeks. Factors associated with greater switching to depression include previous depressive episodes, substance abuse, greater CGI-BP overall severity and benzodiazepine use. Factors associated with lower switching rates were greater CGI-BP depression, lower YMRS severity and atypical antipsychotic use.

LIMITATIONS: The definition of switching biased against patients with mixed episodes being likely to switch.

CONCLUSIONS: Strictly defined, switch to depression from mania occurs in a small proportion of bipolar patients. Clinical history, illness severity, co-morbidities and treatment patterns are associated with switching to depression. Atypical antipsychotics may protect against switch to depression.