Lithium is a cornerstone in treatment of bipolar disorder. Findings are conflicting as to whether acquired unresponsiveness occurs following the discontinuation. Retrospective life chart data were evaluated to investigate the incidence of loss of response. Sixty-five patients chosen from a larger cohort, followed with prospective life charts, who discontinued lithium and had a second lithium treatment. Patients who had at least 2 mood episodes when they were drug naive to describe the natural frequency of illness and 3 mood episodes before the discontinuation were included. The type of response was defined as excellent, partial, or poor according to mirror design method. Eighteen of 65 patients (27.6%) had a decreased response to lithium following its discontinuation. Nine of these patients (13.8%) were unresponsive and nine patients (13.8%) had attenuated response to second lithium treatment. The mean time of discontinuation was longer in the patients who show decreased response (245.8+268.2 vs. 117.9+149.8 days, p=.01). Those who had episode recurrences during the discontinuation were more likely to show reduced responsiveness upon re-treatment. After discontinuation of lithium treatment, more than a quarter of the patients showed an attenuated response or unresponsiveness, and initial partial responders more likely to show unresponsiveness than excellent responders.