OBJECTIVE: The authors explored the prevalence and predictors of nonresponse to reinstituted lithium prophylaxis in bipolar patients who had relapsed after discontinuation of successful lithium treatment. METHOD: The study was conducted with 54 bipolar patients for whom lithium had been reintroduced after one or more recurrences following discontinuation of successful prophylaxis. They were followed up, through bimonthly personal interviews, for 1 year after recovery from the episode during which lithium treatment had been resumed, or up to the first recurrence with onset after lithium reinstitution. RESULTS: During the follow-up period, 44 patients did not have any affective episodes, whereas 10 had at least one recurrence. The only significant difference between the two patient groups was the longer duration of prediscontinuation lithium treatment for the patients who relapsed. CONCLUSIONS: Nonresponse to reinstituted prophylaxis should be considered among the possible risks of the interruption of effective long-term lithium treatment.Abstract Teaser