Antidepressant Discontinuation-Related Mania: Critical Prospective Observation and Theoretical Implications in Bipolar Disorder

Tina R. Goldstein; Mark A. Frye, M.D.; Kirk D. Denicoff, M.D.; Earlian Smith-Jackson, R.N.; Gabriele S. Leverich, L.C.S.W.; Ann L. Bryan; S. Omar Ali; and Robert M. Post, M.D.

Background: Development of manic symptoms on antidepressant discontinuation has primarily been reported in unipolar patients. This case series presents preliminary evidence for a similar phenomenon in bipolar patients.

Method: Prospectively obtained life chart ratings of 73 bipolar patients at the National Institute of Mental Health were reviewed for manic episodes that emerged during antidepressant taper or discontinuation. Medical records were utilized as a corroborative resource. Six cases of antidepressant discontinuation-related mania were identified and critically evaluated.

Results: All patients were taking conventional mood stabilizers. The patients were on antidepressant treatment a mean of 6.5 months prior to taper, which lasted an average of 20 days (range, 1-43 days). First manic symptoms emerged, on average, 2 weeks into the taper (range, 1-23 days). These 6 cases of antidepressant discontinuation-related mania involved 3 selective serotonin reuptake inhibitors (SSRIs), 2 tricyclic antidepressants (TCAs), and 1 serotonin-norepinephrine reuptake inhibitor. Mean length of the ensuing manic episode was 27.8 days (range, 12-49 days). Potential confounds such as antidepressant induction, phenomenological misdiagnosis of agitated depression, physiologic drug withdrawal syndrome, and course of illness were carefully evaluated and determined to be noncontributory.

Conclusion: These 6 cases suggest a paradoxical effect whereby antidepressant discontinuation actually induces mania in spite of adequate concomitant mood-stabilizing treatment. These preliminary observations, if replicated in larger and controlled prospective studies, suggest the need for further consideration of the potential biochemical mechanisms involved so that new preventive treatment approaches can be assessed.

(J Clin Psychiatry 1999;60:563567)

Received July 22, 1998; accepted Nov. 9, 1998. From the Biological Psychiatry Branch, National Institute of Mental Health, Bethesda, Md. (all authors); the Department of Clinical Psychology, University of Colorado, Boulder (Ms. Goldstein); and the Department of Psychiatry and Biobehavioral Sciences, University of California at Los Angeles School of Medicine, Los Angeles (Dr. Frye).

Supported in part by the Ted and Vada Stanley Foundation.

Presented in part at the 151st annual meeting of the American Psychiatric Association, Toronto, Ontario, Canada, June 4, 1998.

Reprint requests to: Robert M. Post, M.D., National Institute of Mental Health, Biological Psychiatry Branch, Building 10 Room 3N212, Bethesda, MD 20892-1272.

Standard clinical practice for treatment of bipolar depression invariably includes the addition of an antidepressant to a mood-stabilization regimen.1 Of concern, however, is the induction of mania; retrospective reviews have estimated the rate of mania to be as high as 35% for all antidepressant classes.2 Although not guided by controlled literature, clinicians usually taper and discontinue antidepressant agents if an adequate trial has not proved clinically efficacious or once some degree of euthymia has been maintained.

Discontinuation of a psychotropic medication resulting in mood relapse or physiologic drug withdrawal symptoms is common. Relapse upon discontinuation of maintenance antidepressant treatment in unipolar patients is well described,3 as is manic relapse upon discontinuation of maintenance mood-stabilizing agents such as lithium and carbamazepine in bipolar patients.4,5

An entire cluster of withdrawal symptoms (including gastrointestinal distress, movement disorders, insomnia, anxiety, and mood fluctuations) have been associated with discontinuation of tricyclic antidepressants (TCAs) and monoamine oxidase inhibitors (MAOIs).6-8 Additionally, Zajecka and colleagues9 have recently proposed an operationalized definition for the physical and psychological symptoms reported upon discontinuation of selective serotonin reuptake inhibitors (SSRIs), also known as the "serotonin reuptake inhibitor discontinuation syndrome"; hallmark to this definition are both somatic and psychic symptoms including dizziness, light-headedness, insomnia, fatigue, nausea, headache, sensory disturbance, and anxiety/agitation. The only controlled study to date on antidepressant discontinuation10 showed differential rates of withdrawal symptoms on discontinuation of various SSRIs. Presumably owing to a longer metabolic half-life, discontinuation of fluoxetine was associated with fewer withdrawal symptoms than either sertraline or paroxetine.

The overlap between symptoms of mania and psychiatric antidepressant withdrawal symptoms led us to question whether conventionally accepted antidepressants have mood-stabilizing (i.e., antimanic) properties. In what would now be considered an extraordinary study, Akimoto11 found 12 (86%) of 14 and 3 (100%) of 3 manic patients had a marked improvement or complete remission with imipramine and amitriptyline monotherapy, respectively. This observation prompted only one controlled study by Klein,12 in which imipramine was found to be no better than placebo for the treatment of mania.

Over a decade later, numerous case reports of antidepressant discontinuation-induced mania or hypomania have been reported in the literature.13-22 The majority (81%) of reported cases chronicle the development of such symptoms in a unipolar (no history of mania) depressed patient population following TCA discontinuation. Far less frequently described, however, and seemingly more confounded, is mania associated with antidepressant discontinuation in bipolar patients. In a retrospective study, Shriver and colleagues23 report that 12 (15%) of 79 episodes of mania in 39 patients occurred within 2 weeks of antidepressant discontinuation. Of the 12 episodes of mood elevation corresponding with antidepressant discontinuation, 7 (58.3%) were associated with SSRI treatment, 3 (25%) with MAOI treatment, 1 (8.3%) with serotonin-norepinephrine reuptake inhibitor (SNRI) treatment, and 1 (8.3%) with TCA treatment.

To add to the literature and promote further discussion and study of this phenomenon, we present 6 cases of apparent antidepressant discontinuation-related mania in bipolar patients. This cohort was selected from a prospective data set and may be critically distinguished from mania induction by antidepressant treatment, phenomenological misdiagnosis of agitated depression, physiologic antidepressant withdrawal syndrome, and natural course of illness variables.


The study cohort included 73 patients participating in clinical trials at the Biological Psychiatry Branch, National Institute of Mental Health (NIMH). The bipolar diagnosis of each patient was confirmed by the Schedule for Affective Disorders and Schizophrenia24 or the Structured Clinical Interview for DSM-IV.25 Mood, functioning, and medications were prospectively tracked daily in a blinded fashion (in the majority of cases) by both patient and clinician using the NIMH life chart methodology (LCM),26 which has been shown to be a reliable and valid instrument for prospective mood tracking.27 Each patient's retrospective LCM documents previous course of illness with particular emphasis on past episodes, illness pattern, hospitalizations, and treatments.

We primarily identified episodes of mania from the prospective LCM data that emerged during antidepressant taper or within 1 week of antidepressant discontinuation. The manic episodes reported by both clinician and patient on the prospective LCM were then secondarily corroborated by medical record review. Our clinical team next critically evaluated the prospective LCM data in a nonblind fashion to exclude cases in which mood elevation may be alternately attributed to (1) antidepressant induction (the antidepressant appeared to induce the mania and therefore was quickly withdrawn), (2) phenomenological misdiagnosis of agitated depression as dysphoric mania, (3) physiologic antidepressant withdrawal syndrome, or (4) course of illness (i.e., seasonality) and therefore likely not related to antidepressant discontinuation.


As presented in Table 1, all 6 patients were taking lithium for mood stabilization (mean lithium level = 0.95 mmol/L; range, 0.5-1.1 mmol/L). Patients were on antidepressant treatment for a mean of 6.5 months (203 days; range, 35-480 days) prior to the taper. The antidepressant treatment was discontinued owing to nonresponse in 4 cases and euthymia in 2 cases. The mean length of antidepressant taper was 19.7 days (range, 1-43 days), while the mean onset of manic symptoms occurred 13.5 days (range, 1-23 days) into the taper. In one case (patient 6) the antidepressant treatment was abruptly discontinued owing to a drug fever.

Of these 6 episodes of antidepressant discontinuation-related mood elevation, 3 were associated with an SSRI (2 sertraline, 1 fluoxetine), 1 with an SNRI (venlafaxine), and 2 with a TCA (1 desipramine, 1 nortriptyline). Five of the cases were classified as "moderate mania" by clinician ratings on the functional impairment scale of the LCM during the worst period of the episode, and 1 case was a hypomanic episode. Two patients exhibited classic progression from euphoric to dysphoric mania. The mean length of the manic episode was 27.8 days (range, 12-49 days). One patient was hospitalized for a period of 19 days owing to manic symptoms while the others were managed on an outpatient basis. Characteristics of each patient's antidepressant discontinuation-related mania that distinguish it from antidepressant induction, agitated depression, physiologic withdrawal, and course of illness are presented in Table 2.


This case series presents data consistent with an antidepressant discontinuation-related mania phenomenon. The very nature of bipolar disorder makes it difficult to attribute causality of mood episodes with certainty. We therefore considered 4 possible confounds to antidepressant discontinuation-related mania.

The literature describes antidepressant-induced mania28 and antidepressant-induced cycle acceleration2 as typically occurring in the same 4- to 8-week time period associated with antidepressant response. By this criterion, causality by antidepressant induction seems very unlikely in all but one case (patient 6, 4.5 weeks exposure), since the mean drug exposure time was 6 months prior to antidepressant discontinuation. Furthermore, in 3 of the 6 patients (2, 4, and 6), the mania in question was more severe and dense than that patient's other manic episodes (both by prospective and retrospective LCM review), which stands in contrast to findings by Stoll et al.29 in which antidepressant-induced mania was found to be milder and briefer than spontaneous mania. Finally, Wehr and Goodwin30 reported that in suspected cases of cycling induction, discontinuation of antidepressants may be associated with periods of remission. Not only are these 6 cases noncycling, but none of the cases described here experienced remission thereafter. Thus, the phenomenon of antidepressant discontinuation-related mania appears to be distinguished from antidepressant-induced mania and cycle acceleration. Of additional interest, and perhaps suggestive of a mood-stabilizing effect of antidepressants, is the partial mania resolution after reinstitution of the antidepressant agent (venlafaxine) in patient 4.

Upon retrospective LCM review, 5 of the 6 patients have no history of antidepressant-induced mania or cycle acceleration. Patient 5 experienced one previous nortriptyline-induced manic episode within 2 weeks of drug initiation. On further follow-up, 2 patients have since experienced 2 episodes each of mood elevation that appear to be associated with antidepressant discontinuation.

Phenomenologically, agitated depression can be difficult to distinguish from dysphoric mania. Although not conclusive, we were able to differentiate the 2 states when the dysphoria was preceded by euphoria (1 and 2) or when core manic symptoms were clearly identified (patients 3, 4, 5, and 6).

Again, there can be overlap between the psychiatric symptoms of a physiologic withdrawal syndrome and mania. Seventy-two percent (31/43) of the symptoms on the Discontinuation-Emergent Signs and Symptoms checklist10 are primarily somatic, yet 5 of the cases presented here experienced prominent mood symptoms with little to no report of bothersome physical symptoms. Additionally, if this were a strictly physiologic withdrawal phenomenon, symptom abatement would be expected on drug elimination, yet the length of ensuing manic episode (mean = 21 days) does not support this explanation. Clearly, this period allows adequate time for metabolic clearance of the primary and parental compound, thus rendering the phenomenon we describe unlikely attributable to physiologic withdrawal alone.

The most challenging differential to antidepressant discontinuation-related mania in bipolar patients is mania occurring as part of the patient's expected course of illness. The retrospective assessment of prospective LCM data do allow, however, critical assessment of seasonal liability, manic severity, and previous postdepression manic episodes. Of the 6 cases of antidepressant discontinuation-related mania presented, each appears distinct in terms of seasonal liability, severity, and prior pattern of illness.

This case series is limited by the small number of cases presented, the absence of a control group, and the lack of large retrospective or prospective literature to guide in our assessment. Additionally, while the data for this study were collected prospectively without intent of identifying the present phenomenon, its retrospective interpretation is not immune to bias. The authors have attempted to minimize any bias by considering other possible perspectives (see Table 2).

The study of this phenomenon may be further complicated by the lack of a true animal model for mania, whereby the effect of antidepressant treatments could subsequently be evaluated. The low reported incidence of this phenomenon may be accounted for by episode misattribution and/or the current trend toward use of multiple mood-stabilizing agents, thus serving as prophylaxis against discontinuation-related mania in the midst of antidepressant taper.

While the precipitation of a manic episode from the discontinuation of an antidepressant agent appears paradoxical, some preliminary speculation has been put forth in the literature attempting to explain the biological mechanism whereby this phenomenon may occur. For example, Dilsaver and colleagues15,31 have proposed that withdrawal-induced cholinergic overdrive can mobilize monoaminergic systems, resulting in manic symptomatology. Zajecka et al.9 proposed that the presynaptic reuptake blockade of serotonin by antidepressants over time results in desensitization of postsynaptic serotonin and norepinephrine receptors. Antidepressant discontinuation after chronic treatment may result in acute enhancement of reuptake, thus depleting serotonin and resulting in a hyposerotonergic state of mania. Noradrenergic and dopaminergic receptor sensitization may also occur and in similar fashion may explain why TCA as well as MAOI discontinuation-related manias have been reported.23,31

Clinical experience shows that many psychotropic medications are successfully utilized as treatments outside of our conventionally accepted definition, including the tricyclically configured anticonvulsant carbamazepine for mood stabilization32,33; the anticonvulsant valproic acid for panic disorder34; nefazodone, an SNRI, for treatment of dysphoric mania35; the typical neuroleptics haloperidol and thiothixene for bipolar depression36; the TCAs trimipramine37 and amoxapine38 for primary psychotic disorders; and clozapine, an atypical neuroleptic, for mood stabilization.39 Along these lines, perhaps reconceptualizing antidepressants as mood stabilizers for some bipolar patients may be warranted based on the premise that treatment for mood episodes may be less about directionality of mood (mania vs. depression) than amplitude of change from euthymia and normal brain function.

Given the case reports, further research into antidepressant discontinuation-related mania is encouraged, including its mechanism of action, prevalence, drug-related differences, and implications in the treatment of bipolar disorder.

Drug names: amitriptyline (Elavil and others), amoxapine (Asendin and others), carbamazepine (Tegretol and others), clozapine (Clozaril), desipramine (Norpramin and others), fluoxetine (Prozac), haloperidol (Haldol and others), nefazodone (Serzone), nortriptyline (Pamelor and others), paroxetine (Paxil), perphenazine (Trilafon and others), sertraline (Zoloft), thioridazine (Mellaril and others), thiothixene (Navane), trimipramine (Surmontil), valproic acid (Depakene), venlafaxine (Effexor).

Disclosure of off-label usage: The authors of this article have determined that, to the best of their knowledge, no investigational information about pharmaceutical agents has been presented herein that is outside Food and Drug Administration-approved labeling.


1. American Psychiatric Association. Practice Guideline for the Treatment of Patients With Bipolar Disorder. Am J Psychiatry 1995;151(suppl 12):1-36

2. Altshuler LL, Post RM, Leverich GS, et al. Antidepressant-induced mania and cycle acceleration: a controversy revisited. Am J Psychiatry 1995;152:1130-1138

3. Kupfer D, Frank E, Perel J, et al. Five-year outcome for maintenance therapies in recurrent depression. Arch Gen Psychiatry 1992;49:769-773

4. Suppes T, Baldessarini R, Faedda G, et al. Risk of recurrence following discontinuation of lithium treatment in bipolar disorder. Arch Gen Psychiatry 1991;48:1082-1088

5. Post RM, Uhde TW, Ballenger JC, et al. Prophylactic efficacy of carbamazepine in manic-depressive illness. Am J Psychiatry 1983;140:1602-1604

6. Charney DS, Heninger GR, Sternberg DE, et al. Abrupt discontinuation of tricyclic antidepressant drugs: evidence for noradrenergic hyperactivity. Br J Psychiatry 1982;141:377-386

7. Dilsaver SC, Greden JF. Antidepressant withdrawal phenomena. Biol Psychiatry 1984;19:237-254

8. Wolfe RM. Antidepressant withdrawal reactions. Am Fam Physician 1997;56:455-462

9. Zajecka J, Tracy KA, Mitchell S. Discontinuation symptoms after treatment with serotonin reuptake inhibitors: a literature review. J Clin Psychiatry 1997;58:291-297

10. Rosenbaum JF, Fava M, Hoog SL, et al. Selective serotonin inhibitor discontinuation syndrome: a randomized clinical trial. Biol Psychiatry 1998;44:77-87

11. Akimoto H. Ten years of psychopharmacology: critical assessment of the present and future. In: Bradley P, Flugel F, Hoch P, eds. Neuropsychopharmacology. Amsterdam, the Netherlands: Elsevier Science; 1964:552-555

12. Klein DF. Importance of psychiatric diagnosis in prediction of clinical drug effects. Arch Gen Psychiatry 1967;16:118-126

13. Mirin SM, Schatzberg AF, Creasey DE. Hypomania and mania after withdrawal of tricyclic antidepressants. Am J Psychiatry 1981;138:87-89

14. Nelson JC, Schottenfeld RS, Conrad CD. Hypomania after desipramine withdrawal. Am J Psychiatry 1983;140:624-625

15. Dilsaver S, Kronfol Z, Sackellares J, et al. Antidepressant withdrawal syndromes: evidence supporting the cholinergic overdrive hypothesis. J Clin Psychopharmacol 1983;3:157-164

16. Theilman SB, Christenbury MM. Hypomania following withdrawal of trazodone [letter]. Am J Psychiatry 1986;143:1482-1483

17. Corral M, Sivertz K, Jones BD. Transient mood elevation associated with antidepressant drug decrease. Can J Psychiatry 1987;32:764-767

18. Hartmann PM. Mania or hypomania after withdrawal from antidepressants. J Fam Pract 1990;30:471-472

19. Ceccherini-Nelli A, Bardellini L, Cur A, et al. Antidepressant withdrawal: prospective findings [letter]. Am J Psychiatry 1993;150:165

20. McGrath P, Stewart JW, Tricamo E, et al. Paradoxical mood shifts to euthymia or hypomania upon withdrawal of antidepressant agents [letter]. J Clin Psychopharmacol 1993;13:224-225

21. Galynker II, Rosenthal RN, Perkel C, et al. Doxepin withdrawal mania [letter]. J Clin Psychiatry 1995;56:122-123

22. Landry P, Roy L. Withdrawal hypomania associated with paroxetine [letter]. J Clin Psychopharmacol 1997;17:60-61

23. Shriver AE, Sachs GS, Baldassano CF. Mania and hypomania following antidepressant discontinuation. In: New Research Program and Abstracts of the 151st Annual Meeting of the American Psychiatric Association; June 1, 1998; Toronto, Ontario, Canada. Abstract NR161:111

24. Mannuzza S, Fyer AJ, Klein DF, et al. Schedule for Affective Disorders and Schizophrenia-Lifetime Version modified for the study of anxiety disorders (SADS-LA): rationale and conceptual development. J Psychiatr Res 1986;20:317-325

25. First MB, Spitzer RL, Gibbon M, et al. Structured Clinical Interview for DSM-IV Axis I Disorders-Patient Edition (SCID-I/P, Version 2.0). New York, NY: Biometric Research, New York State Psychiatric Institute; 1995

26. Leverich GS, Post RM. Life charting the course of bipolar disorder. Curr Rev Mood Anxiety Disord 1996;1:48-61

27. Denicoff KD, Smith-Jackson E, Disney ER, et al. Preliminary evidence of the reliability and validity of the prospective life chart methodology (LCM-P). J Psychiatry Res 1997;5:593-603

28. Wehr TA, Goodwin FK. Rapid cycling between mania and depression caused by maintenance tricyclics. Psychopharmacol Bull 1979;15:17-19

29. Stoll A, Mayer P, Kolbrener M, et al. Antidepressant-associated mania: a controlled comparison with spontaneous mania. Am J Psychiatry 1994;151:1642-1645

30. Wehr T, Goodwin F. Can antidepressants cause mania and worsen the course of affective illness? Am J Psychiatry 1987;144:1403-1411

31. Dilsaver SC, Greden JF. Antidepressant withdrawal-induced activation (hypomania and mania): mechanism and theoretical significance. Brain Res Rev 1984;7:29-48

32. Post RM, Ketter TA, Denicoff K, et al. The place of anticonvulsant therapy in bipolar illness. Psychopharmacology (Berl) 1996;128:115-129

33. Arana GW, Hyman SE. Handbook of Psychiatric Drug Therapy. 2nd ed. Boston, Mass: Little, Brown & Co; 1991

34. Keck PE, McElroy SL, Tugrul K, et al. Antiepileptic drugs for the treatment of panic disorder. Neuropsychobiology 1993;27:150-153

35. Worthington JJ, Pollack MH. Treatment of dysphoric mania with nefazodone. Am J Psychiatry 1996;153:732-733

36. Hendrik V, Altshuler LL, Szuba MP. Is there a role for neuroleptics in bipolar depression? J Clin Psychiatry 1994;55:533-535

37. Berger M, Gastpar M. Trimipramine: a challenge to current concepts on antidepressives. Eur Arch Psychiatry Clin Neurosci 1996;246:235-239

38. Kapur S, Cho R, Jones C, et al. Is amoxapine an atypical antipsychotic? supportive PET evidence [abstract]. Biol Psychiatry 1998;43:18S

39. Frye MA, Ketter TA, Altshuler LL, et al. Clozapine in affective illness: implications for other atypical antipsychotics in the treatment of bipolar disorder. J Affect Disord 1998;48:91-104

Psychiatrists may receive 1 hour of Category 1 credit toward the American Medical Association Physician's Recognition Award by reading the preceding article and correctly answering at least 70% of the questions in the posttest that follows.

1. Read each question carefully and circle the correct corresponding answer on the Registration form.

2. Type or print your full name, address, Social Security, phone, and fax numbers in the spaces provided.

3. Mail the Registration form along with your Net Society username and password to: Physicians Postgraduate Press, Office of CME, P.O. Box 752870, Memphis, TN 38175-2870.

4. For credit to be received, answers must be postmarked by the deadline shown on the CME Registration form. After that date, correct answers to the posttest will be printed in the next issue of the Journal.

All replies and results are confidential. Answer sheets, once graded, will not be returned. Unanswered questions will be considered incorrect and so scored. Your exact score can be ascertained by comparing your answers with the correct answers to the posttest, which will be printed in the Journal issue after the submission deadline. The Physicians Postgraduate Press Office of Continuing Medical Education will keep only a record of participation, which indicates the completion of the activity and the designated number of Category 1 credit hours that have been awarded.
1. Which of the following is not a hallmark symptom in the physiologic withdrawal syndrome referred to as the "serotonin reuptake inhibitor discontinuation syndrome"?

a. Dizziness c. Nausea

b. Anxiety d. Tinnitus

2. Antidepressant-induced mania typically occurs:

a. Within 3-5 days of drug initiation

b. In the same 4- to 8-week period associated with drug response

c. At any given time on the drug

d. 6 months into drug treatment

3. According to the cases presented, antidepressant discontinuation-related mania appears to be:

a. Less severe than spontaneous mania

b. No different from spontaneous mania

c. More severe than spontaneous mania

d. No different from antidepressant-induced mania

4. An argument against this phenomenon being purely physiologic includes:

a. The 6 cases presented all had mood symptoms in addition to their prominent somatic symptoms

b. Only 4 of the 6 experienced dizziness and nausea

c. None of the 6 experienced any physiologic disturbance

d. Mood symptoms did not clear when the antidepressant was eliminated from the system

5. Which of the following is a possible explanation for the low reported incidence of antidepressant discontinuation-related mania?

a. The common use of mood stabilizers in combination with antidepressants

b. Minimal use of antidepressants with bipolar patients

c. The practice of gradual antidepressant discontinuation

d. Abrupt antidepressant discontinuation

6. In the study, mood elevations were attributed to induction by antidepressant discontinuation only when which of the following was ruled out?

a. Previous history of antidepressant-induced mania

b. Previous history of antidepressant use

c. Antidepressant induction, misdiagnosis, physiologic syndrome, and course of illness

d. History of antidepressant-discontinuation induction, physiologic syndrome, and seasonal pattern to mania

7. What theoretical explanation do the authors offer for the observed phenomenon?

a. Treatment recommendations may be based on change from normal brain function

b. Treatment recommendations may be based on polarity of mood

c. Physiologic withdrawal syndrome was mistaken for mania

d. The manic episodes described were simply course of illness

Answers to the February 1999 CME posttest

1. d 2. c 3. e 4. b 5. a 6. b 7. e