Breakthrough depression is a common problem in the treatment of bipolar disorder. Only one, recently published, double-blind, placebo-controlled trial has examined the efficacy of divalproex in the prevention of depressive episodes in bipolar patients. This report describes, in further detail, the findings from that trial of the effect of divalproex on multiple dimensions of depressive morbidity in bipolar disorder. A randomized, double-blind, parallel-group, multicenter study was conducted over a 52-week maintenance period. Bipolar I patients, who may have been treated with open-label lithium or divalproex and who met recovery criteria within 3 months of onset of an index manic episode, were randomized to maintenance treatment with divalproex, lithium, or placebo in a 2 : 1 : 1 ratio. Adjunctive paroxetine or sertraline for breakthrough depression was allowed in maintenance phase. Outcome measures were the rate of early discontinuation for depression, time to depressive relapse, proportion of patients with depressive relapse, mean change in Depressive Syndrome Scale score, proportion of patients receiving antidepressants, and time in the study. Among patients taking an antidepressant, a higher percentage of patients on placebo than divalproex discontinued early for depression. Patients who were previously hospitalized for affective episodes or took divalproex in the open period relapsed later on divalproex than on lithium during the maintenance period. Divalproex-treated patients had less worsening of depressive symptoms than lithium-treated patients during maintenance. Indices of severity of prestudy illness course predicted worse outcome in all treatment groups. Divalproex improved several dimensions of depressive morbidity and reduced the probability of depressive relapse in bipolar disorder, particularly in patients who had responded to divalproex when manic, and among patients with a more severe course of illness.