Review ArticleAntidepressants in bipolar disorder:the case for cautionOver the last decade, reports generated in the US andCanada have pointed out the paucity of evidence onthe efficacy of antidepressants in bipolar disorder(1, 2). Further, recent North American-based treat-ment guidelines, including those of the AmericanPsychiatric Association, have been conservative,recommending antidepressants only for severe bipo-lar depression (3–8). Moreover, if antidepressantsare to be used, they should be withdrawn as early aspossible. This shift away from antidepressant usehas engendered criticism from some groups inEurope, particularly Germany (9). In that criticalarticle on US and Canadian-based treatment guide-lines, the authors assert that these guidelines are notbalanced, and should be rewritten to removerestrictions on the use of antidepressants in thetreatment of bipolar depression.The Munich group asserts that the argument forrestriction of antidepressants in US guidelines isbased on four premises (9), which they state as:(i) The risk of switching into mania/rapid cyclinginduced by antidepressants is an importantclinical phenomenon in bipolar depressionGhaemi SN, Hsu DJ, Soldani F, Goodwin FK. Antidepressants inbipolar disorder: the case for caution.Bipolar Disord 2003: 5: 421–433. ª Blackwell Munksgaard, 2003The 2002 American Psychiatric Association (APA) guidelines for thetreatment of bipolar disorder recommended more conservative use ofantidepressants. This change in comparison with previous APAguidelines has been criticized, especially from some groups in Europe.The Munich group in particular has published a critique of assumptionsunderlying the conservative recommendations of the recent APAtreatment guidelines. In this paper, we re-examine the argument putforward by the Munich group, and we demonstrate that indeed,conceptually and empirically, there is a strong rationale for a cautiousapproach to antidepressant use in bipolar disorder, consistent with, andperhaps even more strongly than, the APA guidelines. This rationale isbased on support for the following four propositions: (i) The risk ofantidepressant induced mood-cycling is high, (ii) Antidepressants havenot been shown to definitively prevent completed suicides and reducemortality, whereas lithium has, (iii) Antidepressants have not beenshown to be more effective than mood stabilizers in acute bipolardepression and have been shown to be less effective than mood stabilizersin preventing depressive relapse in bipolar disorder and (iv) Moodstabilizers, especially lithium and lamotrigine, have been shown to beeffective in acute and prophylactic treatment of bipolar depressiveepisodes. We therefore draw three conclusions from this interpretation ofthe evidence: (i) There are significant risks of mania and long-termworsening of bipolar illness with antidepressants, (ii) Antidepressantsshould generally be reserved for severe cases of acute bipolar depressionand not routinely used in mild to moderate cases and (iii)Antidepressants should be discontinued after recovery from thedepressive episode, and maintained only in those who repeatedly relapseafter antidepressant discontinuation (a minority we judge to representonly about 15–20% of bipolar depressed patients).S Nassir Ghaemia, Douglas J Hsua,Federico Soldaniaand Frederick KGoodwinbaBipolar Disorder Research Program, CambridgeHospital, Cambridge, MA and Harvard MedicalSchool, Boston, MA, USA,bCenter onNeuroscience, Medical Progress, and Society andDepartment of Psychiatry and Behavioral Sciences,George Washington University, Washington, DC,USAKey words: antidepressants – bipolar disorder –depression – mood stabilizers –psychopharmacologyReceived 29 April 2003, revised and accepted forpublication 2 July 2003Corresponding author: S Nassir Ghaemi, MD,Cambridge Hospital, 1493 Cambridge St., Cam-bridge, MA 02139, USA. Fax: 617 665 1623;e-mail: [email protected] Disorders 2003: 5: 421–433Copyright ª Blackwell Munksgaard 2003BIPOLAR DISORDERS421(ii) The risk of suicidality, suicide attempts andsuicide in bipolar depressive patients is ofminor clinical relevance(iii) The antidepressive efficacy of antidepressantsin bipolar depression is insufficiently proven(iv) The antidepressive efficacy of mood stabiliz-ers in bipolar depression is sufficiently provenThis paper will re-examine these four assertions,and make the case for US-based treatment guide-lines that de-emphasize the use of antidepressantsin the treatment of bipolar disorder.The case for cautionWe agree that antidepressants may be effective intreating acute bipolar depression, as there is someevidence to that effect. While all treatment guide-lines recognize this, the point of contention regard-ing North American-based treatment guidelinescenters on how often antidepressants should beused and for how long.We believe that the evidence of antidepressantefficacy in bipolar depression is not as definitive asmany assume. Given the risks of acute mania, theroutine use of antidepressants would appear to bemore risky without much added benefit over theuse of mood stabilizers alone for acute bipolardepression, a point which seems even more com-pelling with the appearance of lamotrigine. Fur-ther, available studies fail to provide any rigorousevidence of antidepressant prevention of depressiverelapse. In contrast, such evidence exists withmood stabilizers. Moreover, there is likely asignificant risk of more mood episodes over timeand possible rapid cycling with long-term antide-pressant use in bipolar disorder.Seen this way, the research evidence appears tosupport US and Canadian-based treatment guide-lines in which antidepressants use is restricted tocases of severe depression (or when the appropriatemood stabilizer combination has failed to preventor reverse a depression); further the guidelinesrecommend antidepressant discontinuation afteracute recovery.Critics fear that, if antidepressants are usedless aggressively, bipolar depression will go under-treated and the suicide risk will rise. The evidence,however, is not clear that there is anti-suicidalbenefit with antidepressants, whereas lithium,among all psychotropic agents, has by far themost extensive evidence of an anti-suicide effect.Moreover, these treatment guidelines account forpotential undertreatment of depression by sup-porting antidepressant use in cases of severedepression. The question is not whether or notantidepressants