In his book Psychiatry:A textbook for Students and Physicians published in 1899, Kraeplin (see giants, home page) mentioned that individuals with bipolar disorder transition or switch from depressive to manic or mixed states without an intervening period of euthymia or stable feeling. The extent of this risk has been hard to determine, as have the risk factors and the clinical features of the particular bipolar patient in question. Some older studies put the switch rate with antidepressants as high as 50% while some more recent randomized placebo-controlled trials have come in at less than 10% when antidepressants are combined with mood stabilizers (Keck et. al. 2005, Nemeroff et.al. 2001).
In spite of this being an active research area, a number of important questions remain. First, what is the incidence of switching of patients with bipolar I and II in large clinical cohorts? Second, are individuals at greatest risk for switching characterized by sociodemographic features, and if so, what are they? Third, If such features exist, do they influence antidepressant-treated patients differentially from those not treated with antidepressants?
These questions are addressed in the Systemic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) which was a multicenter study conducted in the US between 1999 and 2005. The methods are detailed by Perlis et al., 2006, and Sachs et al., 2003. The full cohort included 3640 subjects with 48,287 follow-up visits; among these 2166 (69.3% bipolar I, 59.3% female).
With regard to the first question above, the researchers found 21.3% transitioned directly from depression to mania (4.9%), hypomania (8.5%), or mixed (7.8%). This cohort included 19.6% of subjects who transitioned were antidepressant-exposed and 24.9% were antidepressant-unexposed.
The second question seeks to determine whether there are sociodemographic factors or clinical features that predict risk for switching. The study found that the factors significantly associated with increased risk of switching were younger age, earlier age of illness onset, history of rapid cycling in the past year, history of suicide attempts, and a greater proportion of days spent in an elevated, anxious, or irritable state, and current substance abuse (especially alcohol) in the past year. A meta-analysis of this data has been published by Bond et al., which suggested greater risks for bipolar I than bipolar II.
I should emphasize that these predictors do not differentiate significantly between antidepressant-treated and not-antidepressant-treated individuals. Therefore, one cannot conclude that individuals with greater subthreshold manic severity should not receive antidepressants as these subjects are at elevated risk independent of their treatment status.