June 1993

Outcome After Rapid vs Gradual Discontinuation of Lithium Treatment in Bipolar Disorders

Author Affiliations

From the Consolidated Department of Psychiatry and Neuroscience Program, Harvard Medical School, Boston, Mass, Laboratories for Psychiatric Research-Psychotic Disorders Program, Mailman Research Center, McLean Division of Massachusetts General Hospital, Belmont, and Bipolar Disorder Research Program, McLean and Massachusetts General Hospitals, Boston, (Drs Faedda, Baldessarini, Suppes, and Tohen); Centro Lucio Bini, Cagliari, Italy (Drs Faedda and Tondo); and the Department of Epidemiology, Harvard School of Public Health, Boston (Dr Tohen).

Arch Gen Psychiatry. 1993;50(6):448-455. doi:10.1001/archpsyc.1993.01820180046005

Objective:  Withdrawal of bipolar mood disorder (BP-I) patients from prolonged, stable lithium maintenance has a high risk of early recurrence, particularly of mania. We thus compared risks of stopping lithium rapidly vs gradually.

Design:  Outpatients undergoing clinically determined discontinuation of lithium treatment at different rates were followed up prospectively to 5 years. Risks and timing of new episodes were analyzed.

Patients:  Subjects (N=64) with a DSM-III-R BP disorder, previously stable on lithium monotherapy for 18 to 120 months (mean, 3.6 years) were followed up clinically after discontinuing lithium (elected in prolonged wellbeing in 67%). None was unavailable for follow-up, and subtyping (BP-I or BP-II) remained stable.

Results:  Within 5 years, 75% had a recurrent episode; BP-I patients were 1.5-times less likely than BP-II to remain in remission. Polarity of first-recurrent and onset episodes was 80.8% concordant. Overall risk of a new episode of mania was significantly greater after rapid (<2) than gradual (2 to 4 weeks discontinuation (5-year hazard ratio=2.8); the difference in risk of depression was even greater hazard ratio=5.4). Recurrence rate was more elevated within months of rapid discontinuation (12month hazard ratio=5.4). Recurrence rate was more elevated within months of rapid discontinuation (12-month hazard ratio=4.3) than at later times (2 to 5 years), when courses of "survival" over time were nearly parallel in both discontinuation groups.

Conclusions:  Risk of early recurrence of BP disorder following discontinuation of lithium maintenance is elevated, but may be both predictable (timing and polarity) and modifiable by gradual discontinuation.