BACKGROUND The relationship between psychomotor agitation in unipolar depression and mood-switching from depression to manic, hypomanic and mixed states has been controversial. We investigated the future risk of initial mood-switching as a function of psychomotor agitation in unipolar depression. METHODS We identified 189 participants diagnosed with major depressive disorder (MDD). We divided all patients with MDD into two categories (1) agitated patients (n=74), and (2) non-agitated patients (n=115). These groups were prospectively followed and compared by time to mood-switching. Kaplan-Meier survival curves, log-rank test for trend for survivor functions, and Cox proportional hazard ratio estimates for a multivariate model were conducted to examine the risk of mood-switching by psychomotor agitation. RESULTS During follow-up, mood-switching occurred in 20.3% of the agitated patients and 7.0% of the non-agitated patients. In the Kaplan-Meier survival estimates for time to incidence of mood-switching with agitated or non-agitated patients, the cumulative probability of developing mood-switching for agitated patients was higher than those for non-agitated patients (log-rank test: χ(2)=7.148, df=1, p=0.008). Survival analysis was also performed using Cox proportional hazards regression within a multivariate model. The agitation remained significantly associated with incidence of mood-switching (HR=2.98, 95% CI: 1.18-7.51). CONCLUSIONS We did not make a clear distinction between antidepressant-induced mood-switching and spontaneous switching. CONCLUSIONS The main finding demonstrated that MDD patients with agitation were nearly threefold as likely to experience mood-switching, suggesting that psychomotor agitation in MDD may be related to an indicator of bipolarity.