Sequential Multiple Assignment Randomized Treatment (SMART) for Bipolar Disorder at Any Phase of Illness and at least Mild Symptom Severity.

Objectives To sequentially study the effectiveness of lithium and divalproex monotherapy and adjunctive therapy with quetiapine or lamotrigine in the acute and continuation treatment of bipolar I or II disorder at any phase of illness and at least mild symptom severity. Methods From June 2011 to December 2016, patients with bipolar I or II disorder (using DSM-IV diagnostic criteria) and CGI-S (Clinical Global Impression-Severity) ⩾ 3 were randomized to receive lithium or divalproex monotherapy for 2 weeks. Patients who had CGI-S-depression ⩾ 3 for 2 weeks at any time after 2-week monotherapy were randomly assigned to receive quetiapine or lamotrigine, or remaining on monotherapy for a total of 26 weeks. Results The rates of early termination due to lack of efficacy and side effects and changes in BISS (Bipolar Inventory of Symptoms Scale) and CGI-S total score were not significantly different between lithium and divalproex. The completion rate was significantly higher with adjunctive therapy than with monotherapy. BISS and CGI-S total scores, and their sub-scores were significantly reduced with adjunctive therapy compared to monotherapy. Adjunctive therapy significantly increased survival times compared to monotherapy (hazard ratio = 6.8), and the monotherapy group had a significantly increased risk for not reaching sustained recovery from depression (hazard ratio = 12.7). Patients who did not need the 2nd randomization and remained on monotherapy had a significantly reduced hazard for discontinuation (hazard ratio = 3.8). Conclusions The efficacy of lithium and divalproex as monotherapy was modest. Adjunctive lamotrigine and quetiapine to either one was well-tolerated and equally effective in reducing bipolar symptomatology, but adjunctive therapy should be initiated as early as possible when depression symptoms are present.