Mania associated with antidepressant
treatment: comprehensive meta-analytic
Occurrence of mania or other forms of abnormally
elevated mood (?mood switch?) in association with
antidepressant (AD) treatment, has been recog-
nized since the earliest use of imipramine in the late
1950s (1–6). However, it remains unclear to what
extent such mood switching arises naturally in the
course of major affective illnesses, or is actually
caused by AD treatment (7–13). Mood elevation
Tondo L, Va ´ zquez G, Baldessarini RJ. Mania associated with
antidepressant treatment: comprehensive meta-analytic review.
Objective: To review available data pertaining to risk of mania–
hypomania among bipolar (BPD) and major depressive disorder
(MDD) patients with vs. without exposure to antidepressant drugs
(ADs) and consider effects of mood stabilizers.
Method: Computerized searching yielded 73 reports (109 trials, 114 521
adult patients); 35 were suitable for random effects meta-analysis, and
multivariate-regression modeling included all available trials to test for
effects of trial design, AD type, and mood-stabilizer use.
Results: The overall risk of mania with ⁄without ADs averaged
12.5% ⁄7.5%. The AD-associated mania was more frequent in BPD
than MDD patients, but increased more in MDD cases. Tricyclic
antidepressants were riskier than serotonin-reuptake inhibitors (SRIs);
data for other types of ADs were inconclusive. Mood stabilizers had
minor effects probably confounded by their preferential use in
Conclusion: Use of ADs in adults with BPD or MDD was highly
prevalent and moderately increased the risk of mania overall, with little
protection by mood stabilizers.
L. Tondo1,2, G. V?zquez3,
R. J. Baldessarini1
1Department of Psychiatry and Neuroscience Program,
Harvard Medical School and McLean Division of
Massachusetts General Hospital, Boston, MA, USA,
2Department of Psychology, University of Cagliari, and
Lucio Bini Mood Disorders Research Center, Cagliari,
Sardinia, Italy and3Department of Neurosciences,
University of Palermo, Buenos Aires, Argentina
Key words: antidepressants; bipolar disorder; major
depression; mania; meta-analysis
Leonardo Tondo, MRC 3, McLean Hospital, 115 Mill
Street, Belmont, MA 02178-9106, USA.
Accepted for publication October 27, 2009
Antidepressant drugs (AD) treatment of bipolar disorder (BPD) and major depressive disorder
(MDD) patients approximately doubled moderate spontaneous risk of mania–hypomania.
A small proportion of adults initially diagnosed with MDD emerged as probable BPD cases during
Tricyclic antidepressants carried a higher risk of new mania–hypomania than serotonin reuptake
Mood stabilizers had little preventive effect against mood elevation during AD treatment.
These meta-analytic findings should be viewed cautiously because of heterogeneity of studies
Clinically, we recommend use of mood-stabilizers in BPD patients, especially when exposed to ADs,
despite limited support in this study.
Acta Psychiatr Scand 2010: 121: 404–414
All rights reserved
? 2009 John Wiley & Sons A/S
associated with AD treatment might indicate:
efficacy of the treatment, an adverse pharmacolog-
ical effect, misdiagnosis of previously unrecognized
bipolar disorder (BPD), or actual conversion from
major depressive disorder (MDD) to BPD. It is
also likely that ADs increase occurrence of abnor-
mally elevated mood in some vulnerable patients,
regardless of clinical diagnosis (14–17). The impor-
tance of this topic includes not only inconsistencies
and gaps in previous research, continued uncer-
tainty about the scientific and clinical significance
of manic reactions during AD treatment, and their
legal implications if harm results, but also the
paucity of comprehensive and quantitative analy-
MDD and BPD patients, or assessment of effects
of simultaneously administered mood stabilizers
A widely held assumption is that mood elevation
during AD treatment indicates the presence of
BPD, which sometimes is first recognized during
such treatment, particularly in young depressed
patients (20, 21). Risk factors for mood switching
(as well as diagnostic confusion) during exposure
to an AD may include the presence of previous
mixed manic-depressive states, especially agitation
during a treated depressive episode (21–23), as well
as relatively early age-at-onset (20, 24), and a
cyclothymic (25) or hyperthymic temperament
(26). All ADs have been associated with mania,
although they may vary by type, dose, and latency
to occurrence (10, 27–30).
Aims of the study
To address the several remaining questions about
effects of antidepressant drugs (ADs) on risk of
new psychopathological excited states, we compre-
hensively reviewed all available controlled and
open trials since the dawn of the AD era in the late
1950s, pertaining to the topic or involving mania as
a planned or incidental finding, considering all
reported AD drugs and major mood disorders, as
well as effects of co-administering putative mood-
Material and methods
We considered all reports of original data on AD
treatment of BPD or MDD, and occurrence of new
episodes of mania, mixed-states, or hypomania
(here, all termed ?mania?). Potential studies were
(PubMed?database, using ?mania,? or ?hypoma-
nia,? and ?antidepressant? as search terms) through
December 2008, supplemented with references
cited in publications identified by searching. To
obtain as complete a survey as possible, we
considered blinded, randomized, controlled trials
(RCTs) as well as unblinded and only partially
controlled clinical cohort trials to obtain rates of
mania in both AD-treated and AD-untreated or
placebo-treated patients, in the same trials. For
meta-analysis, we included only non-duplicative
data from studies with ‡2 arms in which at least
eight treated and untreated patients were com-
pared. Data were verified directly with authors
when not clear from publications (31, 32). Since
case–control and case report studies preclude
estimates of actual rates of mania (events ⁄per-
sons-at-risk), they were not included. We also
excluded studies reporting only on episodes or
rates of mania, rather than persons who became
manic among those at-risk. New episodes of mania
were expressed as the proportion (%) of patients
becoming manic during observation, or as rates (%
of patients⁄month). Given the rarity of reporting
precise exposure times per person, we relied on
nominal average treatment exposures (months)
reported for each study treatment arm, or on
reported; average exposure was £24 months in all
We rated study quality by giving one point each
for: i) blinded RCT design with parallel groups; ii)
‡50 subjects ⁄arm; and iii) follow-up observation
for 2–24 months (maximum score = 3.0). For
meta-analysis of studies having a study arm with
no mood-elevation episodes, we used a common
continuity-adjustment method (adding 0.5% to
paired samples with and without AD exposure) in
computations of pooled rate ratios (RR), and also
estimated pooled risk difference [RD, which
tolerates zero values in one study arm and is
(NNH = 1⁄RD)];all
values are reported with their 95% confidence
intervals (CI) (33).Studies
switches, with and without AD, were excluded
from meta-analysis as uninformative. We also
examined the potential influence of unusually
large studies on pooled RRs by serially excluding
them singly from meta-analyses (influence or
In separate meta-analyses, we also compared
studies: i) involving MDD vs. BPD patients
[unspecified forms of major affective disorder
(MAD) were excluded]; ii) using RCT vs. open
designs; iii) with old [tricyclic ADs (TCAs) or
modern ADs [mostly serotonin reuptake inhibitors
(SRIs); occasionally serotonin–norepinephrine
with no mood
reuptake inhibitors (SNRIs) or the atypical agent
bupropion]; and iv) with vs. without concomitant
use of mood stabilizers (lithium, certain anticon-
vulsants, or unspecified agents, with too few
studies of antipsychotic agents with putative
mood-stabilizing effects), including studies with at
least some patients treated with a mood stabilizer.
There was insufficient information to allow
separate analyses of types I vs. II BPD patients
or of specific types of mood elevation (mania,
hypomania, mixed-state). In addition, we consid-
ered whether pooled RR values were altered by
covariates specified below, including use of meta-
regression analysis that considered: publication
year, months of treatment exposure, and quality
rating. We also carried out multivariate linear
regression analyses, with the proportion of patients
becoming manic in each study as the outcome
variable, and independent variables including:
diagnosis (BPD vs. MDD) study design (RCT vs.
cohort-comparison studies), use of a mood stabi-
lizer, type of AD, quality rating, duration of
follow-up, current age, and year of publication.
Timing of mania (months from onset of AD
exposure) was not stated in most studies, obviating
use of survival analysis.
Statistical analyses used commercial microcom-
puter programs (stata 8.0?; Stata Corporation,
College Station, TX, USA: statview-5?; SAS
Institute, Cary, NC, USA). Means are shown
with standard deviation (SD), and other computed
values are shown with 95% CI. Comparisons with
two-tailed P > 0.05, at stated degrees of freedom
(df), were considered not significant.
Studies and population sample
We identified 73 studies [39 (53.4%) RCTs and 34
or 46.6% open trials] with data pertaining to new
episodes of hypomania or mania during 109
treatments with ADs (1–6, 14, 16, 18, 20, 27, 29,
31, 32, 34–92), divided by type of study design,
diagnosis, type of AD, and use of a mood
stabilizer. The 73 studies included a total of
114 521 adult patients (56 212 given ADs, and
58 309 not; 7915 with BPD, 102 501 with MDD,
and 4105 with an unspecified MAD), aged 4l.7
(SD = 7.20) years, and exposed for an average of
5.32 (SD = 6.09) months (Table 1). Of these
studies, 35 provided information suitable for
meta-analysis, in that they reported numbers of
subjects with mania during AD treatment or
without it and total numbers of subjects treated
and not treated.
Risk of manic episodes
Proportions (%) of patients experiencing new
episodes of mania-like excitation during trials are
summarized in Table 1, divided into those exposed
vs. not exposed to an AD, separated by BPD vs.
MDD, diagnoses, and those given a mood stabi-
lizer or not. Overall proportions of manias, with ⁄
without an AD, averaged 12.5% ⁄7.46% (1.7-times
greater with ADs [paired-t (df = 53) = 2.09,
P = 0.04], with an overall RD of 5.04% across
months, or an increase of about 1% ⁄month
(Table 1). As expected, rates of new-mania were
5.32 (SD = 6.09)
Table 1. Characteristics of studies and patients reviewed
Number of reports reviewed
Number of comparisons
Design types: per cent RCTs
Current age, years: mean (SD)
Approximate exposure time, months: mean (SD)
Subjects, n (%)
Major affective disorder (unspecified)
Not given antidepressants
Proportions of patients with mania, % (CI)
BPD: all cases
BPD: no mood stabilizer
BPD: with mood stabilizer
MDD: all cases
MDD: no mood stabilizer
MDD: with mood stabilizer
Not given antidepressants
BPD: all cases
BPD: no mood stabilizer
BPD: with mood stabilizer
MDD: all cases
MDD: no mood stabilizer
MDD: with mood stabilizer
Risk difference: (treated minus untreated)
Rates of new mania (% ⁄month)
4.49 (–1.25; 10.2)
2.38 ()1.58; 6.34)
AD, antidepressant drugs; BPD, bipolar disorder; MDD, major depressive disorder;
RCT, randomized controlled trial.
Patients in an unusually large study by Martin et al. 2004 (20) numbered 81 036;
however, rates of new mania without this study yielded rates of 12.5% with, vs.
7.67% without AD treatment, based on averages per study, and so were very
similar to the reported overall values above. Moreover, follow-up times for BPD
(5.2 months) vs. MDD patients (6.1 months) were similar.
Tondo et al.
much greater among BPD patients [with and
without ADs: 15.3% and 13.8%; ratio = 1.1;
paired-t (df = 26) = 0.20, P = 0.84] than in
1.24%; ratio = 4.8; paired-t (df = 24) = 4.02
P = 0.0005], with much greater relative increases
with ADs among MDD patients. These compari-
sons indicate greater base- and AD-associated risks
among BPD patients, as expected in this mania-
prone subgroup, but a remarkable relative increase
by nearly five times among MDD patients given an
AD. Estimated rates (% ⁄month) of mania with vs.
without AD treatment were 5.0 vs. 1.5, or 3.3-fold
higher with ADs. Among patients diagnosed with
BPD vs. MDD, these rates, averaging 6.23
(SD = 13.6) and 4.01 (SD = 6.86), respectively,
did not differ significantly [F (df = 40) = 0.46;
P = 0.50].
We tested selected variables for association with
risk for mania in a multivariable, linear-regression
model (Table 2). Among patients given ADs, rates
of new mania were independently and significantly
higher (in descending order of significance) among:
i) patients with BPD vs. MDD; ii) longer reported
average follow-up; iii) younger average current
age; and iv) clinical cohorts vs. RCTs. Without
ADs, factors associated with mania were: i) diag-
nosis [BPD > MDD (Fig. 1)]; ii) longer follow-up;
and iii) younger age (Table 2). Other factors not
associated with risk of mania (with or without
ADs) in such multivariate modeling were: year of
publication, study design (RCT vs. clinical) and
quality rating, type of AD given, and use of a
Mania vs. antidepressant type
Crude proportions (%) of patients becoming
manic during exposure to particular types of ADs
ranked: TCAs > modern agents [mainly SRIs plus
a few trials involving an SNRI (duloxetine, venla-
faxine) or bupropion] >MAOIs (Table 3). How-
ever, TCAs did not differ from modern ADs
[SRIs + SNRIs + bupropion;
2.43; P = 0.123], nor did MAOIs differ significantly
from modern ADs: [(df = 39) F = 1.57; P =
(df = 71),F =
Table 2. Factors associated with percent of patients experiencing mania or
hypomania with vs. without antidepressants
b-Coefficient (95% CI)t-scoreP-value
BPD vs. MDD
Younger current age
Open trials vs. RCTs
BPD vs. MDD
Younger current age
BPD, bipolar disorder; MDD, major depressive disorder; RCT, randomized controlled
Based on multivariate, linear regression modeling. Factors found to be unrelated to
the outcome in both groups, included study quality (based on RCT design, ‡50
subjects in each study-arm, and ‡2 months of exposure to treatment), year of
publication, type of antidepressant given, use of a mood stabilizer; and design
(open vs. RCT) without AD treatment.
Fig. 1. Risk (%) of mania or hypomania (with SEM) associ-
ated with and without antidepressant drug treatment for
bipolar disorder (n = 29 studies) and major depressive disor-
der patients (n = 25 studies).
Table 3. Antidepressant type and risk of mania or hypomania
Patients with new
mania (% CI)
AD, antidepressant drugs; MAOI, monoamine oxidase inhibitors; TCA, tricyclic ADs;
SRI, serotonin reuptake inhibitors; SNRI, serotonin–norepinephrine reuptake
Statistical comparisons were by paired tests comparing mania risk with vs. without
ADs for each AD type. ?Mixed? drugs are not defined in reports, other to note that
?ADs? were given. For comparison (see Table 1), untreated patients with presum-
ably spontaneous mania were included in 45 trials involving 58 309 subjects,
yielding a rate of 7.46% (95% CI: 7.36–7.56) mania or hypomania.
0.217], although TCAs were associated with more
risk than MAOIs [(df = 58) F = 6.38; P =
Meta-analytic modeling of factors associated with mania
Based on random effects meta-analyses, there was
a highly significant difference of mania risk with vs.
without AD exposure (RR = 1.76, P < 0.0001,
with an estimated NNH of 27 persons; Table 4).
Risk assessed with meta-analyses also was associ-
ated significantly and similarly with: i) AD treat-
ment in studies involving both RCTs and clinical
cohort designs; ii) in MDD but not BPD; iii) with
TCAs but not modern ADs or MAOIs (Fig. 2);
and iv) with little difference in apparent risk with
vs. without mood stabilizers; Table 4).
In patients given ADs, risk with ⁄without a mood
stabilizer was: for BPD, 15.9% ⁄13.8%; for MDD,
4.49%⁄6.16%; without an AD, the corresponding
pooled ratios of risks were: BPD: 11.2% ⁄16.5%,
MDD: 2.38% ⁄1.03%, and none of these contrasts
with ⁄without a mood stabilizer in diagnostic or
treatment subgroups was statistically significant
(Table 1). The preceding findings, in contrast to
computations of crude risks of mania (Table 3),
excluded studies with no mania events with or
without AD treatment (notably with MAOI stud-
ies, 5 ⁄6 of which found no cases of new mania),
possibly accounting for somewhat different impres-
sions, notably including a lack of important
differences in risk with particular types of ADs,
based on meta-analyses (Table 4).
Applying an influence or sensitivity analysis to
the meta-analytic results (eliminating one study at
a time) indicated that only one study appeared to
be an outlier; this was an unusually large trial by
Martin and colleagues reported in 2004 (20) that
involved AD treated vs. untreated MDD patients.
Even excluding this study from meta-analytic
modeling, the association of AD treatment with
(RR = 1.65, and NNH = 30, P = 0.005). In
found that no factor (trial design, study quality,
diagnosis, AD type, mood-stabilizer use, year of
publication, or duration of AD exposure) was
independently and significantly associated with risk
of mania (Table 4).
Among mood-disordered subjects overall, we
found an increase of risk (RD) of pathological
mania like mood elevation with AD exposure of
5.04% (12.5% with, 7.46% without ADs, or 1.68-
fold). This risk was much higher among BPD vs.
MDD patients, without AD-treatment (13.8% ⁄
1.24%, or 11.1-fold), as well as with such treatment
(15.3 ⁄5.97%, or2.6-fold; Table 1), again as
expected. Notably, in addition to the expected
and substantial risks of new mania in BPD, some
patients initially diagnosed with unipolar MDD
also carried a finite risk of mania, hypomania, or
mixed states (?mania?), which averaged 5.97% with,
and 1.24% without AD exposure (Table 1). Over-
all, the present findings indicate a gradient of rates
of new mania, ranking: BPD with ADs > BPD
without ADs > MDD with ADs > MDD without
ADs (Table 1), in agreement with clinical experi-
ence. Diagnosis (BPD vs. MDD) also was a highly
significant and independent risk factor for new
mania in multivariate modeling (Table 2). For
comparison, reported rates of new manias for BPD
patients are 15–50%, and among patients diag-
nosed with MDD, 3–10% (11, 93, 94). However,
the relative risk due to AD exposure vs. spontane-
ous risk in BPD (only 1.11-fold) was far less than
in MDD cases (4.81-fold). This conclusion is
supported not only in crude estimates (Table 1),
Table 4. Summary of results of meta-analyses for
proportion of patients with new mania ⁄hypomania
Measure ContrastStudies (n) RR (95%CI)zP-valueNNH (95%CI)
SRIs + SNRIs
Data are relative risk (RR) with ⁄without AD treatment, and its 95% confidence interval (CI), and number needed to
harm (NNH) estimates based on reciprocals of pooled mean differences in risks with vs. without antidepressant (AD)
treatment. Moreover, in separate meta-regression analysis, these factors plus the publication year and months of
treatment exposure were not independently associated with occurrence of mania.
Tondo et al.
but also by meta-analysis (the relative increase in
mania with ADs in BPD ⁄MDD was 1.13 ⁄3.76),
and was statistically significant only among the
MDD patients (Table 4). This seemingly paradox-
ical outcome reflects the relatively high rate of
spontaneous mania among BPD patients that was
little increased (by only 1.13%) during AD expo-
sures averaging 5 months.
The surprisingly high relative increase of new
mania-like states associated with AD treatment
among MDD patients (ratio of drug-associated vs.
spontaneous risk) raises several possibilities. AD
treatment itself may induce a manic, hypomanic, or
mixed state in some patients previously considered
to have MDD (14, 15, 95), with a relatively large
gain in risk of nearly five-fold among MDD treated
vs. not treated with ADs (Table 1). This inter-
pretation is further supported by the low rate
(1.0–2.3%) of spontaneous states of abnormal
excitation among MDD patients without AD
treatment (96, 97), again reflecting clinical experi-
ence. However, induction of mania in MDD
patients may indicate current limitations to differ-
entially diagnosing BPD vs. MDD, and uncertain-
ties about how to classify presumed MDD patients
who later become manic, especially only with
exposure to ADs. Patients considered to have
MDD, but who become pathologically excited
when exposed to an AD, sometimes are considered
?misdiagnosed BPD? patients. Others (sometimes
Fig. 2. Forest plots [for individual trials (gray boxes reflecting study size and measurement variance) vs. computed rate ratios (RR)
with 95% CI, with pooled RR and CI values (vertical dotted lines and black diamonds)] for a total of 45 trials yielding pooled RR
values (with 95% CIs) computed by random effects meta-analyses for three major antidepressant drug groups, with null value of
RR = 0.0 (as vertical solid lines) and computed pooled RR in descending order of apparent effect on manifestation of new mania,
for: (a) monoamine oxidase inhibitors [MAOIs, six trials; pooled RR = 2.83 (CI: 0.79–10.2); z = 1.60, P = 0.110], (b) tricyclic
antidepressants [TCAs, 30 trials; pooled RR = 1.92 (CI: 1.13–3.30); z = 2.41, P = 0.016), and (c) serotonin-reuptake inhibitors
[SRIs, nine trials; pooled RR = 1.70 (CI: 0.87–3.32); z = 1.56, P = 0.12], indicating that TCAs alone, were probably riskier than
SRIs, but with overlapping CIs. In addition, we estimated number needed to harm [NNH as reciprocals of random effects meta-
analyses of risk differences (RD), as: (a) MAOIs: NNH = 31.3 (CI: 14.6–infinity); (b) TCAs: NNH = 18.5 (CI: 8.65–infinity)];
(c) SRIs: NNH = 90.9 (CI: 17.0–infinity); these results suggest that MAOIs and TCAs may carry indistinguishable risks, and when
results from RCTs of these first generation antidepressants pooled, NNH = 21.7 (CI: 10.3–infinity), a value that is 5.4-times lower
(greater risk) with older agents than with modern SRIs, whose NNH = 117 (CI: 15.7–infinity). Note that many of these results are
marginally or not statistically significant, and only 4 ⁄45 studies considered showed significantly higher risk with ADs. References to
studies cited are provided in parentheses (some studies included more than one treatment condition).
considered ?type-III? BPD (98), or as ?BPD not
otherwise specified? in DSM-IV) may truly be cases
of MDD, vulnerable to transient, pharmacologi-
cally induced excitement during AD treatment
without evidence of spontaneous mania or hypo-
mania. Yet another hypothetical possibility is that
AD treatment itself may induce BPD in formerly
MDD patients (14, 15, 93). Differentiating among
these possibilities can be facilitated by following
such patients clinically after mood elevation in
association with AD treatment. Many who later
develop spontaneous mania, hypomania, or mixed
states would be considered as falling within a
proposed ?BP spectrum? (95–97, 99–101). Aside
from the uncertain significance of mania associated
with AD treatment, the research literature on this
topic is large, complex, and leaves unanswered
questions, including medico-legal uncertainties
pertaining to harm associated with unexpected,
drug-associated mania, all of which encouraged the
In accordance with current clinical opinions,
TCA treatment was more likely to be associated
with manic or other excited states than modern
ADs, mainly represented by SRIs (12, 68, 89). This
difference confirms findings from a previous,
smaller meta-analytic study of effects of ADs in
BPD patients only (10) and from a recent review of
RCTs in bipolar depression (102). The present
review indicated minor differences in risk among
modern ADs as a group, including SRIs and very
few studies of SNRIs or of the atypical agent
bupropion (Table 3). There are insufficient num-
bers of trials involving specific ADs to support
meta-analyses comparing risks between AD types.
However, like the TCAs, venlafaxine, especially
when given at high doses, has a reputation for
producing more mania than other modern ADs
(28, 72). The present data pertaining to this agent
or duloxetine, another SNRI, showed no difference
in rates of new manias with the other modern ADs
but were too limited to support separate meta-
Risk factors associated with newly emerging
included currently younger age, at least among
BPD patients (Table 2). We also found that study
design was significantly associated with risk of
mania, in that clinical cohort trials yielded higher
risks of new manic episodes than RCTs. This
observation may reflect the relatively brief AD
exposures in most RCTs (4.76 vs. 6.62 months in
RCTs vs. open trials; Table 1), exclusion of obvi-
ously mania-prone subjects entering RCTs involv-
ing risk of being given an AD, and the lack of
blinded treatment in the cohort studies, that may
risk inflating estimates of new mania. Even the
observed average exposure of 5.3 months may be
insufficient to capture episodes of emerging, espe-
cially spontaneous, mania and too brief to docu-
ment cycle acceleration (100), but too long to
exclude spontaneous mania. That is, new mania
arising during a relatively brief RCT may be
particularly indicative of a causal association with
AD treatment. Instead, prolonged observation,
with or without AD treatment, probably increases
chances of observing spontaneous mania that may
confound estimates of risks associated with AD
treatment. Longer follow-up was associated with
greater risk of mania in AD-treated patients in a
study from the Stanley Bipolar Treatment Net-
work, with nearly twice-higher rates of mania or
hypomania when follow-up involved a year vs.
only 10 weeks, but that study could not distinguish
spontaneous from AD-induced mania (101). In
addition, Angst and his colleagues reported that
risk of new episodes of mania, or conversion from
diagnoses of unipolar MDD to BPD rose contin-
uously at about 1% ⁄year for 50 years, during
prolonged follow-up (96).
Surprisingly, mood stabilizers did not have an
appreciable mania-limiting effect in the studies
reviewed, with or without co-treatment with an
AD (Table 1). Moreover, in a multivariate regres-
sion analysis, with or without ADs, mood-stabi-
lizing treatment was not specifically associated
with decreased risk of mania (Table 2). These
unexpected observations may be confounded by
bias toward giving mood stabilizers to known
cases of BPD or generally sicker patients with
higher rates of spontaneous or AD-induced mania,
and by short average observation times (5 months
of AD exposure) that may limit potentially bene-
ficial, long-term effects of mood stabilizers. The
lack of clear evidence of reduced risk of mania
when ADs were given to BPD patients treated with
a mood stabilizer, as found in this analysis accords
with some recent controlled studies (10, 72, 88, 91,
103, 104), but not others, especially when lithium
was used (30, 66, 102, 105). Interpretation of
effects of combining mood stabilizers and ADs for
BPD patients should also considered that a need
for adjunctive AD treatment, itself, may reflect
lack of effectiveness of prophylaxis with a mood
stabilizer alone. Whatever its interpretation, our
findings concerning mood-stabilizer co-treatment
with an AD does not support the prevalent clinical
impression that these drugs effectively limit risk of
mania, at least among type I BPD patients,
including many who are exposed to ADs (12,
103, 106, 107). In addition, evidence from animal
studies indicates that mood stabilizers do not
Tondo et al.
prevent behavioral sensitization to ADs or other
agents that may model drug-induced mania (108).
Finally, modern pharmacogenomic and other
biological research promise to characterize, and
support prediction of patients likely to become
manic when given an AD, and should be pursued
In conclusion, in this extensive review, risk of
mania or hypomania in MAD patients exposed to
ADs (?switch? risk) averaged 12% overall. Despite
uncertainties about interpretation of new episodes
of mania or hypomania during AD treatment, we
support the practice of treatment with a mood
stabilizer when an AD is prescribed for BPD
patients as clinically prudent, pending further
This study was supported in part by awards from NARSAD
and the Centro Bini Private Donors Mood Research Fund
(LT), by grants from the Bruce J. Anderson Foundation and
by the McLean Private Donors Neuropsychopharmacology &
Bipolar Disorder Research Fund (RJB). The authors thank
Roy Perlis, M.D. of Massachusetts General Hospital and
Nassir Ghaemi, M.D. of Tufts University School of Medicine
for valuable advice, and Christian Teter, Pharm.D., and Aime ´ e
Mertz, pre-Pharm.D. of the Bouve ´ College of Health Sciences,
Northeastern University and McLean Hospital for biblio-
graphic advice and assistance.
Declaration of interest
Dr Tondo has collaborated in investigator-initiated research
with Janssen and Eli Lilly. Dr Va ´ zquez is a consultant with
AstraZeneca, Glaxo-SmithKline, and Eli Lilly Corporations.
Dr Baldessarini has been a consultant or research collaborator
with: AstraZeneca, Auritec, Biotrofix, Janssen, JDS-Noven,
Lilly, Luitpold, NeuroHealing, Novartis, Pfizer, and SK-
BioPharmaceutical Corporations. No author is a member of
pharmaceutical speakers? bureaus, nor do they or any family
member hold equity positions in biomedical or pharmaceutical
1. Ball JR, Kiloh LG. Controlled trial of imipramine in
treatment of depressive states. Br Med J 1959;2:1052–
2. Keup W, Apolito A, Olinger L, Schwartz M, Yachnes E.
Inpatient treatment of depressive states with tofranil
(imipramine hydrochloride). Am J Psychiatry 1959;
3. Leyberg JT, Denmark JC. The treatment of depressive
states with imipramine hydrochloride (Tofranil). J Ment
4. Kiloh LG, Child JP, Latner G. A controlled trial of ipro-
niazid in the treatment of endogenous depression. J Ment
5. Miller A, Baker EF, Lewis D, Jones A. Imipramine, a
clinical evaluation in a variety of settings. Can Psychiatr
Assoc J 1960;5:150–160.
6. Ho ¨hn R, Gross GM, Gross M, Lasagna L. A doubleblind
comparison of placebo and imipramine in the treatment
of depressed patients in a state hospital. J Psychiatr Res
7. Mo ¨ller HJ, Grunze H. Have some guidelines for the
treatment of acute bipolar depression gone too far in the
restriction of antidepressants? Eur Arch Psychiatry Clin
8. Angst J, Gamma A. A new bipolar spectrum concept: a
brief review. Bipolar Disord 2002;4:11–14.
9. Ghaemi SN, Hsu DJ, Soldani F, Goodwin FK. Antidepres-
sants in bipolar disorder: the case for caution. Bipolar
10. Gijsman HJ, Geddes JR, Rendell JM, Nolen WA, Goodwin
GM. Antidepressants for bipolar depression: a systematic
review of randomized, controlled trials. Am J Psychiatry
11. Visser HM, van der Mast RC. Bipolar disorder, antide-
pressants and induction of hypomania or mania: systemtic
review. World J Biol Psychiatry 2005;6:231–241.
12. Goodwin GM, Anderson I, Arango C et al. ECNP consen-
sus meeting. Bipolar depression. Nice, March 2007. Eur
13. Angst J. Do antidepressant drugs include hypoma-
nia ⁄mania? Proceedings of the Annual Congress of the
Italian Society of Psychopathology (SOPSI), Rome, Italy,
14. Wehr TA, Goodwin FK. Can antidepressants cause mania
and worsen the course of affective illness? Am J Psychiatry
15. Kukopulos A, Reginaldi D, Laddomada P, Floris G, Serra G,
Tondo L. Course of the manic-depressive cycle and chan-
ges caused by treatment. Pharmakopsychiatr Neuropsy-
16. Ghaemi SN, Rosenquist KJ, Ko JY, Baldessano CF, Kontos
NJ, Baldesssarini RJ. Antidepressant treatment in bipolar
vs. unipolar depression. Am J Psychiatry 2004;161:163–
17. Goodwin FK, Jamison KR, eds. Manic-depressive illness,
2nd edn. New York: Oxford University Press, 2007:751–
18. Lewis JL, Winokur G. The induction of mania: a natural
history with controls. Psychopharmacol Bull 1987;23:74–
19. McElroy SL, Kotwal R, Kaneria R, Keck PE Jr. Antide-
pressants and suicidal behavior in bipolar disorder.
Bipolar Disord 2006;8:596–617.
20. Martin A, Young C, Leckman JF et al. Age effects on
antidepressant-induced manic conversion. Arch Pediatr
Adolesc Med 2004;158:773–780.
21. Baldessarini RJ, Faedda GL, Hennen J. Risk of mania with
serotonin reuptake inhibitors vs. tricyclic antidepressants
in children, adolescents and young adults. Arch Pediatr
Adolesc Med 2005;159:298–299.
22. Koukopoulos A, Koukopoulos AE. Agitated depression as a
mixed state and the problem of melancholia. Psychiatr
Clin North Am 1999;22:547–564.
23. Bottlender R, Sato T, Kleindienst N, Strauss A, Mo ¨ller
HJ. Mixed depressive features predict maniform switch
during treatment of depression in bipolar I disorder.
J Affect Disord 2004;78:149–155.
24. Lewis DA, Nasrallah HA. Mania associated with electro-
convulsive therapy. J Clin Psychiatry 1986;47:366–367.
25. Akiskal HS, Djenderedjian AM, Rosenthal RH, Khani MK.
Cyclothymic disorder: validating criteria for inclusion in
the bipolar affective group. Am J Psychiatry 1977;
26. Kukopulos A, Reginaldi D, Laddomada P, Floris G, Serra G,
Tondo L. Rapid cyclers, temperament, and antidepres-
sants. Compr Psychiatry 1983;24:249–258.
27. Peet M. Induction of mania with selective serotonin re-
uptake inhibitors and tricyclic antidepressants. Br J Psy-
28. Stoll AL, Mayer PV, Kolbrener M et al. Antidepressant-
associated mania: a controlled comparison with sponta-
neous mania. Am J Psychiatry 1994;15:1642–1645.
29. Sachs GS, Lafer B, Stoll AL et al. A double-blind trial of
bupropion versus desipramine for bipolar depression.
J Clin Psychiatry 1994;55:391–393.
30. Leverich GS, Altshuler LL, Frye MA et al. Risk of
switch in mood polarity to hypomania or mania in pa-
tients with bipolar depression during acute and contin-
uation trials of venlafaxine, sertraline, and bupropion as
adjuncts to mood stabilizers. Am J Psychiatry 2006;
31. Akiskal HS, Hantouche EG, Allilaire JF et al. Validating
antidepressant-associated hypomania (bipolar III): a sys-
tematic comparison with spontaneous hypomania (bipo-
lar II). J Affect Disord 2003;3:65–74.
32. Serretti A, Artioli P, Zanardi R, Rossini D. Clinical fea-
tures of antidepressant associated manic and hypomanic
switches in bipolar disorder. Prog Neuropsychopharma-
col Biol Psychiatry 2003;27:751–757.
33. DerSimonian R, Laird NM. Meta-analysis in clinical trials.
Control Clin Trials 1986;7:177–188.
34. Rees L, Davies B. A controlled trial of phenelzine (‘‘Nar-
dil’’) in the treatment of severe depressive illness. J Ment
35. Bartholomew AA. An evaluation of tranylcypromine
(‘‘Parnate’’) in the treatment of depression. Med J Aust
36. Sharbek A. Trial of amitriptyline in chronic depression.
Dis Nerv Syst 1963;24:115–119.
37. Browne MW, Kreeger LC, Kazamias NG. A clinical trial of
amitriptyline in depressive patients. Br J Psychiatry
38. Gottfries CG. Clinical trial with the monoamido oxidase
inhibitor tranylcypromine of a psychiatric clientele. Acta
Psychiatr Scand 1963;39:463–472.
39. Vahia NS, Bagadia VCN, Doongaji DR. Value of some
antidepressants. Hospital Reports by the Medical and
Professional Staff of the Jamsetji Jijibhai Hosp Grant
Med Coll 1964;9:199–203.
40. Klein DF. Chlorpromazine-procyclidine
imipramine and placebo in depressive disorders. Can
Psychiatr Assoc J 1966;11(Suppl.):146–149.
41. Prien RF, Klett CJ, Caffey EM. Lithium carbonate and
imipramine in prevention of affective episodes. A com-
parison in recurrent affective illness. Arch Gen Psychiatry
42. Bunney WE Jr. Psychopharmacology of the switch process
in affective illness. In: Lipton M, DiMascio A, Killam K,
eds. Psychopharmacology: a generation of progress. New
York: Raven Press, 1978:1249–1259.
43. Van Scheyen JD, van Kammen DP. Clomipramine-induced
mania in unipolar depression. Arch Gen Psychiatry
44. Quitkin FM, Kane J, Rifkin A, Ramos-Lorenzi JR, Nayak
DV. Prophylactic lithium carbonate with and without
imipramine for bipolar 1 patients. A double-blind study.
Arch Gen Psychiatry 1981;38:902–907.
45. Tondo L, Laddomada P, Serra G, Minnai G, Kukopulos A.
Rapid cyclers and antidepressants. Int Pharmacopsychi-
46. Kane JM, Quitkin FM, Rifkin A, Ramos-Lorenzi JR, Nayak
DD, Howard A. Lithium carbonate and imipramine in the
prophylaxis of unipolar and bipolar II illness: a prospec-
tive, placebo-controlled comparison. Arch Gen Psychiatry
47. Jann MW, Bitar AH, Rao A. Lithium prophylaxis of tri-
cyclic-antidepressant-induced mania in bipolar patients.
Am J Psychiatry 1982;139:683–684.
48. Himmelhoch JM, Fuchs CZ, Symons BJ. A double-blind
study of tranylcypromine treatment of major anergic
depression. J Nerv Ment Dis 1982;170:628–634.
49. De Wilde JE, Doogan DP. Fluvoxamine and chlorimipr-
amine in endogenous depression. J Affect Disord 1982;
50. Nasrallah HA, Lyskowski J, Schroedere D. TCA-induced
mania: differences between switchers and nonswitchers.
Biol Psychiatry 1982;17:271–274.
51. Prien RF, Kupfer DJ, Mansky PA et al. Drug therapy in the
prevention of recurrences in unipolar and bipolar affective
disorders. Report of the NIMH Collaborative Study
Group comparing lithium carbonate, imipramine, and a
lithium carbonate-imipramine combination. Arch Gen
52. Casacchia M, Carolei A, Barba C et al. A placebo-
controlled study of the antidepressant activity of moclo-
bemide, a new MAO-A inhibitor. Pharmacopsychiatry
53. Angst J. Switch from depression to mania, or from mania
to depression: role of psychotropic drugs. Psychophar-
macol Bull 1987;23:66–67.
54. Kupfer DJ, Carpenter LL, Frank E. Possible role of an-
tidepressants in precipitating mania and hypomania in
recurrent depression. Am J Psychiatry 1988;145:804–
55. Cohn JB, Collins G, Ashbrook E, Wernicke JF. A com-
parison of fluoxetine imipramine and placebo in patients
with bipolar depressive disorder. Int Clin Psychophar-
56. Solomon RL, Rich CL, Darko DF. Antidepressant treat-
ment and the occurrence of mania in bipolar patients
admitted for depression. J Affect Disord 1990;18:253–
57. Johnstone EC, Owens DG, Lambert MT, Crow TJ, Frith CD,
Done DJ. Combination tricyclic antidepressant and lith-
ium maintenance medication in unipolar and bipolar
depressed patients. J Affect Disord 1990;20:225–233.
58. Himmelhoch JM, Thase ME, Mallinger AG, Houck P.
depression. Am J Psychiatry 1991;148:910–916.
59. Thase ME, Mallinger AG, McKnight D, Himmelhoch JM.
Treatment of imipramine-resistant recurrent depression,
IV: A double-blind crossover study of tranylcypromine
for anergic bipolar depression. Am J Psychiatry 1992;
60. Bocchetta A, Bernardi F, Burrai C, Pedditzi M, Del Zompo
M. A double-blind study of L-sulpiride versus amitripty-
line in lithium-maintained bipolar depressives. Acta Psy-
chiatr Scand 1993;88:434–439.
61. Altshuler LL, Post RM, Leverich GS, Mikalauskas K,
Rosoff K. Antidepressant-induced mania and cycle accel-
eration: a controversy revisited. Am J Psychiatry 1995;
62. Howland RH. Induction of mania with serotonin reuptake
inhibitors. J Clin Psychopharmacol 1996;16:425–427.
63. Amsterdam J. Efficacy and safety of venlafaxine in the
treatment of bipolar II major depressive episode. J Clin
Tondo et al.
64. Grossman F, Potter WZ, Brown EA, Maislin G. A double-
blind study comparing idazoxan and bupropion in bipolar
depressed patients. J Affect Disord 1999;56:237–243.
65. Young LT, Joffe RT, Robb JC, MacQueen GM, Marriott M,
Patells-Siotis I. Double-blind comparison of addition of a
second mood stabilizer versus an antidepressant to an
initial mood stabilizer for treatment of patients with
bipolar depression. Am J Psychiatry 2000;157:124–126.
66. Bottlender R, Rudolf D, Strauß A, Mo ¨ller HJ. Mood-
stabilizers reduce the risk of developing antidepressant-
induced maniform states in acute treatment of bipolar I
depressed patients. J Affect Disord 2001;63:79–83.
67. Kupfer DJ, Chengappa KN, Gelenberg AJ et al. Citalopram
as adjunctive therapy in bipolar depression. J Clin Psy-
68. Silverstone T. Moclobemide vs. imipramine in bipolar
depression: a multicentre double-blind clinical trial. Acta
Psychiatr Scand 2001;104:104–109.
69. Henry C, Sorbara F, Lacoste J, Gindre C, Leboyer M.
Antidepressant-induced mania in bipolar patients: identi-
fication of risk factors. J Clin Psychiatry 2001;62:249–255.
70. Altshuler L, Kiriakos L, Calcagno J et al. The impact of
continuation on 1-year risk for relapse of bipolar
depression: a retrospective chart review. J Clin Psychiatry
71. Nemeroff CB, Evans DL, Gyulai L et al. Double-blind,
placebo-controlled comparison of imipramine and par-
oxetine in the treatment of bipolar depression. Am J
72. Vieta E, Martinez-Ara ´n A, Goikolea JM et al. A ran-
domized trial comparing paroxetine and venlafaxine in
the treatment of bipolar depressed patients taking mood
stabilizers. J Clin Psychiatry 2002;63:508–512.
73. Goldberg JF, Whiteside JE. The association between sub-
stance abuse and antidepressant-induced mania in bipolar
disorder: a preliminary study. J Clin Psychiatry 2002;
74. Altshuler L, Suppes T, Black DO et al. Impact of antide-
pressant discontinuation after acute bipolar depression
remission on rates of depressive relapse at 1-year follow-
up. Am J Psychiatry 2003;160:1252–1262.
75. Tohen M, Vieta E, Calabrese J et al. Efficacy of olanzapine
and olanzapine-fluoxetine combination in the treatment
of bipolar I depression. Arch Gen Psychiatry 2003;
76. Amsterdam JD, Shults J. Fluoxetine monotherapy of
bipolar type II and bipolar NOS major depression: a
double-blind, placebo-substitution, continuation study.
Int Clin Psychopharmacol 2005;20:257–264.
77. Ghaemi SN, Goodwin FK. Antidepressants for bipolar
depression. Am J Psychiatry 2005;162:1545–1547.
78. Joffe RT, MacQueen GM, Marriott M, Young LT. One-
year outcome with antidepressant-treatment of bipolar
depression. Acta Psychiatr Scand 2005;112:105–109.
79. Kupka RW, Luckenbaugh DA, Post RM et al. Comparison
of rapid-cycling and non-rapid-cycling bipolar disorder
based on prospective mood ratings in 539 outpatients. Am
J Psychiatry 2005;162:1273–1280.
80. Dunner DL, D?Souza DN, Kajdasz DK, Detke MJ, Russell
JM. Is treatment-associated hypomania rare with dul-
oxetine: secondary analysis of controlled trials in non-
bipolar depression. J Affect Disord 2005;87:115–119.
81. Schaffer A, Zuker P, Levitt A. Randomized, double-blind
pilot trial comparing lamotrigine versus citalopram for the
treatment of bipolar depression. J Affect Disord 2006;
82. Wada K, Sasaki T, Jitsuki H et al. Manic ⁄hypomanic
switch during acute antidepressant treatment for unipolar
depression. J Clin Psychopharmacol 2006;26:512–515.
83. Fonseca M, Soares JC, Hatch JP, Santin AP, Kapczinski F.
Open trial of adjunctive escitalopram in bipolar depres-
sion. J Clin Psychiatry 2006;67:81–86.
84. Mundo E, Cattaneo E, Russo M, Altamura AC. Clinical
variables related to antidepressant-induced mania in
bipolar disorder. J Affect Disord 2006;92:227–230.
85. Perlis RH, Ostacher MJ, Patel JK et al. Predictors of
recurrence in bipolar disorder: primary outcomes from
the Systematic Treatment Enhancement Program for
Bipolar Disorder (STEP-BD). Am J Psychiatry 2006;
86. Koukopoulos A, Sani G, Koukopoulos AE, Albert MJ,
Girardi P, Tatarelli R. Endogenous and exogenous
cyclicity and temperament in bipolar disorder: review,
new data and hypotheses. J Affect Disord 2006;96:165–
87. Post RM, Altshuler LL, Leverich GS et al. Mood switch in
bipolar depression: comparison of adjunctive venlafaxine,
bupropion and sertraline. Br J Psychiatry 2006;189:124–
88. Truman CJ, Goldberg JF, Ghaemi SN et al. Self-reported
history of manic ⁄hypomanic switch associated with anti-
depressant use: data from the Systematic Treatment
Enhancement Program for Bipolar Disorder (STEP-BD).
J Clin Psychiatry 2007;68:1472–1479.
89. Sachs GS, Nierenberg AA, Calabrese JR et al. Effectiveness
depression. N Engl J Med 2007;356:1711–1722.
90. Carlson GA, Finch SJ, Fochtmann LJ et al. Antidepressant-
associated switches from depression to mania in severe
bipolar disorder. Bipolar Disord 2007;9:851–859.
91. Amsterdam JD, Shults J. Comparison of short-term ven-
lafaxine versus lithium monotherapy for bipolar II major
depressive episode. J Clin Psychopharmacol 2008;28:171–
92. McElroy S, Olausson B, Chang W, et al. A double-blind,
placebo-controlled study with acute and continuation
phase of quetiapine in adults with bipolar depression
(EMBOLDEN II). Proc 3rd Biennial Conference of the
International Society for Bipolar Disorders. Delhi, India,
Jan 27–28, 2008.
93. Goldberg JF, Truman CJ. Antidepressant-induced mania:
an overview of current controversies. Bipolar Disord
94. Benvenuti A, Rucci P, Miniati M et al. Treatment-emergent
mania ⁄hypomania in unipolar patients. Bipolar Disord
95. Akiskal HS, Akiskal KK, Lancrenon S et al. Validating the
bipolar spectrum in the French National EPIDEP Study:
overview of the phenomenology and relative prevalence of
its clinical prototypes. J Affect Disord 2006;96:197–205.
96. Angst J, Sellaro R, Stassen HH, Gamma A. Diagnostic
conversion from depression to bipolar disorders: results of
a long-term prospective study of hospital admissions.
J Affect Disord 2005;84:149–157.
97. Akiskal HS, Maser JD, Zeller P et al. Switching from
‘‘unipolar’’ to bipolar II: 11-year prospective study of
clinical and temperamental predictors in 559 patients.
Arch Gen Psychiatry 1995;52:114–123.
98. Klerman GL. The spectrum of mania. Compr Psychiatry
99. Akiskal HS, Pinto O. The evolving bipolar spectrum:
prototypes I, II, III, IV. Psychiatr Clin North Am
100. Post RM, Leverich GS, Nolen WA et al. Re-evaluation of
the role of antidepressants in the treatment of bipolar
depression: data from the Stanley Foundation Bipolar
Network. Bipolar Disord 2003;5:396–406.
101. Licht RW, Gijsman H, Nolen W, Angst J. Are antidepres-
sants safe in the treatment of bipolar depression? A crit-
ical evaluation of their potential risk to induce switch into
102. Salvi V, Fagiolini A, Swartz HA, Maina G, Frank E. Use of
antidepressants in bipolar disorder. J Clin Psychiatry
103. Reginaldi D, Tondo L, Floris G, Pignatelli A, Kukopolos A.
Poor prophylactic lithium response due to antidepres-
sants. Int Pharmacopsychiatry 1981;16:124–128.
104. Baldessarini RJ, Tarazi FI. Pharmacotherapy of psychosis
and mania. Chapt 18 In: Brunton LL, Lazo JS, Parker KL,
eds. Goodman and Gilman?s the pharmacological basis of
therapeutics, 11th edn. Cambridge: McGraw-Hill Press,
105. Henry C, Demotes-Mainard J. Avoiding drug-induced
switching in patients with bipolar depression. Drug Saf
106. Baldessarini RJ, Leahy L, Arcona S, Gause D, Zhang W,
Hennen J. Patterns of psychotropic drug prescription for
U.S. patients with diagnoses of bipolar disorders. Psy-
chiatr Serv 2007;58:85–91.
107. Baldessarini RJ, Henk HJ, Sklar AR, Chang J, Leahy LF.
Use of psychotropic medicines in bipolar disorder patients
in the United States. Psychiatr Serv 2008;59:1175–1183.
108. D?Aquila PS, Fanin F, Serra G. Chronic valproate fails to
prevent imipramine-induced behavioral sensitization to
the dopamine D2-like receptor agonist quinpirole. Eur J
109. DeLuca V, Mundo E, Trakalo J, Wong GW, Kennedy JL.
Investigation of polymorphism in the MDR1 gene and
antidepressant-induced mania. Pharmacogenomics J 2003;
110. Barr M, Rose D. The great ambivalence: factors likely to
affect service user and public acceptability of the phar-
Health Illn 2008;30:944–958.
111. Eppel AB. Antidepressants in the treatment of bipolar
disorder: decoding contradictory evidence and opinion.
Harv Rev Psychiatry 2008;16:205–209.
112. Lin E, Chen PS. Pharmacogenomics with antidepressants in
the STAR*D study. Pharmacogenomics 2008;9:935–946.
113. Yadid G, Friedman A. Dynamics of the dopaminergic
system as a key component to the understanding of
depression. Progr Brain Res 2008;172:265–286.
Tondo et al.