Lithium discontinuation: uncovering latent bipolar disorder?

Lithium discontinuation: uncovering latent bipolar disorder?

Authors: Faedda GL, Tondo L, Baldessarini RJ.
Lucio Bini Center, New York NY 10023; Department of Psychiatry, Harvard Medical School; and the International Consortium for Bipolar Disorder Research.
Am J Psychiatry. 2001 Aug;158(8):1337-9

Bauer and colleagues (1) add to already compelling evidence that supplementing antidepressant treatment with lithium can enhance the response of many treatment-resistant depressive patients (2). In their study, 41/75 (54.7%) patients with DSM-III-R major depression and no clinical evidence of bipolar disorder or currently suicidality, responded poorly to an apparently adequate antidepressant trial but recovered when open-label lithium was added for 8�10 weeks. Of the recovered patients, 29 were later randomized to continue lithium supplementation (N=14) or to discontinue to a placebo as antidepressant treatment continued (N=15) up to four months. Following lithium-discontinuation, 7/15 (46.7%) patients again became ill: 5/7 (71.4%) depressed and 2/7 (28.6%) manic, and one of the discontinued patients (1/15; 6.7%) committed suicide. Therefore, at least 2/15 (13.3%) of the previously unipolar depressed patients were rediagnosed as bipolar. These responses all emerged within 4 months (average of 4 weeks) following lithium discontinuation. Several findings of this study call for comment. First, relatively rapid discontinuation of lithium (over 1�7 days) led to a 47% risk of early return of affective illness. It may be that this antidepressant-unresponsive group simply fell ill again after removal of lithium and that the two cases of mania arose by chance. However, such high early risk of both depression and mania has been found in many studies of bipolar or mixed major affective disorder patients following lithium discontinuation (3�5). Moreover, some of the observed outcome differences in studies involving lithium vs. placebo, particularly after rapid discontinuation of lithium, may include treatment withdrawal-related responses (4,6). Recurrence of mania soon after discontinuing lithium in patients with previously nonbipolar depression is not widely documented. However, spontaneous expression of mania following episodes of depression is well known (7), and antidepressant treatment is sometimes associated with manic or mixed-agitated states in depressed patients (8,9). The spontaneous (or partly antidepressant-associated) rate of switching from apparent unipolar to bipolar disorder over prolonged follow-up was 12.5% (70/559) within ca. 5.5 years (2.28%/year) in a recent large study (7). In the reported experience, the observed rate was at least 35-times greater (2/30/0.0833 years, equivalent to 80.0 %/year) for mania in antidepressant-unresponsive depressed patients discontinued from lithium while antidepressants continued. The rediagnosed cases may represent latent bipolar disorder that followed an apparently unipolar course exposure to the combined pharmacodynamic stressors of lithium discontinuation plus unprotected antidepressant treatment. Antidepressant-unresponsive depressed patients represent a special group that may include a disproportionate risk of latent or potential bipolarity. Moreover, many of the approximately one-third to one-half (1,2) of treatment-resistant depressed patients who respond to lithium augmentation of antidepressant treatment may derive from this pseudo-unipolar or latent bipolar subgroup. In general, the new findings (1) suggest that lithium discontinuation may be dangerous in patients with unrecognized latent bipolar disorder, particularly among antidepressant-unresponsive depressed patients. Clinically, this experience encourages extra efforts to seek evidence (past and family history) of potential bipolarity among antidepressant-resistant depressives. Special caution is also warranted in the early weeks following rapid discontinuation of lithium (or any psychotropic agent), and gradual tapering of medication should be considered whenever feasible (3). The report also raises again the question of interpretation of drug/placebo contrasts in research involving drug discontinuation (4,6).