The effects of milnacipran (50 mg bid) on sleep patterns of eight depressed inpatients, treated for 4 weeks, were studied during the initial (days 1-3) and terminal (days 26-28) treatment periods and compared with those obtained from three sleep recordings performed just prior to the initiation of the treatment. The clinical evolution of patients was evaluated weekly using the MADRS depression scale and the Spiegel and Norris sleep scales. Clinical improvement, shown by a mean reduction of 58% in MADRS scale scores, was accompanied by an improvement of disturbed sleep parameters. From the beginning of treatment, there was an increase in the total duration of sleep and stage II sleep, a decrease in sleep latency and an increase in sleep efficiency. Total REM sleep was not modified although, since there was an increase in total sleep time, the percent REM sleep was significantly reduced. REM latency was increased early in the study, an effect classically associated with antidepressant treatment. This study suggests that milnacipran improves disturbed sleep parameters in depressed patients without any additional disturbance at the onset of treatment.