Little is known about clinical features associated with the risk of recurrence in patients with bipolar disorder receiving treatment according to contemporary practice guidelines. The authors looked for the features associated with risk of recurrence. The authors examined prospective data from a cohort of patients with bipolar disorder participating in the multicenter Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) study for up to 24 months. For those who were symptomatic at study entry but subsequently achieved recovery, time to recurrence of mania, hypomania, mixed state, or a depressive episode was examined with Cox regression. Of 1,469 participants symptomatic at study entry, 858 (58.4%) subsequently achieved recovery. During up to 2 years of follow-up, 416 (48.5%) of these individuals experienced recurrences, with more than twice as many developing depressive episodes (298, 34.7%) as those who developed manic, hypomanic, or mixed episodes (118, 13.8%). The time until 25% of the individuals experienced a depressive episode was 21.4 weeks and until 25% experienced a manic/hypomanic/mixed episode was 85.0 weeks. Residual depressive or manic symptoms at recovery and proportion of days depressed or anxious in the preceding year were significantly associated with shorter time to depressive recurrence. Residual manic symptoms at recovery and proportion of days of elevated mood in the preceding year were significantly associated with shorter time to manic, hypomanic, or mixed episode recurrence. Recurrence was frequent and associated with the presence of residual mood symptoms at initial recovery. Targeting residual symptoms in maintenance treatment may represent an opportunity to reduce risk of recurrence.