Association between circadian activity rhythms and mood episode relapse in bipolar disorder: a 12-month prospective cohort study.

A significant proportion of patients with bipolar disorder experience mood episode relapses. We examined whether circadian activity rhythms were associated with mood episode relapses in patients with bipolar disorder. This prospective cohort study included outpatients with bipolar disorder who participated in a study titled "Association between the Pathology of Bipolar Disorder and Light Exposure in Daily Life (APPLE) cohort study." The participants' physical activity was objectively assessed using a wrist-worn accelerometer over 7 consecutive days for the baseline assessment and then at the 12-month follow-up for mood episode relapses. The levels and timing of the circadian activity rhythms were estimated using a cosinor analysis and a nonparametric circadian rhythm analysis. Of the 189 participants, 88 (46%) experienced mood episodes during follow-up. The Cox proportional hazards model adjusting for potential confounders showed that a robust circadian activity rhythm, including midline-estimating statistic of rhythm (MESOR) and amplitude by cosinor analysis and 10 consecutive hours with the highest amplitude values (M10) by the nonparametric circadian rhythm analysis, was significantly associated with a decrease in mood episode relapses (per counts/min, hazard ratio [95% confidence interval]: MESOR, 0.993 [0.988-0.997]; amplitude, 0.994 [0.988-0.999]; and M10, 0.996 [0.993-0.999]). A later timing of the circadian activity rhythm (M10 onset time) was significantly associated with an increase in the depressive episode relapses (per hour; 1.109 [1.001-1.215]). We observed significant associations between circadian activity rhythms and mood episode relapses in bipolar disorder.