Continuation versus discontinuation of lithium in recurrent bipolar illness: a naturalistic study


  • The authors of this paper do not have any commercial associations that might pose a conflict of interest in connection with this manuscript.

Matthew G Biel, MD, MSc, 550 First Avenue, NBV 21 South 6, New York, NY 10016-9196, USA.
Fax: +1 212 263 0202; e-mail:


Objectives:  Lithium's efficacy in prophylaxis of mood episodes in bipolar disorder (BD) is well established in the clinical trial setting, but may be less robust in routine clinical practice. We compared illness recurrence in bipolar patients naturalistically continued on or discontinued from lithium after an extended period of clinical stability on lithium monotherapy, and evaluated other potential risk factors for relapse.

Methods:  We followed 213 patients who were stable for 2 years on lithium monotherapy following resolution of acute symptoms marking their last manic episode. Based upon patient preference and clinical judgment, 159 patients were continued on lithium monotherapy and 54 patients were slowly discontinued. Survival differences between the continued and discontinued groups were assessed using the Kaplan–Meier product limit method, and risk factors for relapse were evaluated in Cox proportional hazards regression.

Results:  Patients continued on lithium prophylaxis experienced risk of recurrence equivalent to a third of that suffered by discontinued patients during the first year of treatment (0.15 versus 0.45), and significant survival differences persisted throughout follow-up. Median survival time to illness recurrence for patients continued on lithium was 7.33 years [95% confidence interval (CI) 5.67–9.67]; that for patients discontinued from lithium was 1.33 years (95% CI 0.33–2.33). After controlling for all significant covariates, lithium discontinuation was associated with a hazard ratio of 4.85. Inter-episode manic and depressive symptoms conferred increased risk for subsequent recurrence of illness in both groups, while lower lithium levels recorded during the two years of clinical stability preceding study onset were protective.

Conclusions:  Despite considerable rates of illness recurrence in both groups, those who continued on lithium sustained markedly lower rates of recurrence over a lengthy follow-up period. Lithium discontinuation in BD after successful maintenance monotherapy is not advisable. If discontinuation is considered, lithium levels previously required to maintain clinical stability, and breakthrough or residual mood symptoms experienced during remission, should inform clinical decision making.