Does acute mania occur with antidepressant treatment, and if so, how frequent does it? To assess whether it occurs, one would ideally need
RCTs designed to answer the question. However, RCTs of antidepres-sants in bipolar depression are not designed to assess side effects like acute mania induction; rather, they are designed to assess efficacy. This leads to the common mistake of saying that there is no risk for a side effect like mania induction because of the lack of statistical significance for such risk. However, absence of evidence is not evidence of absence—such studies would need to include about 10,000 patients to demonstrate statistical differences between antidepressant and placebo, given the low rates seen in RCTs.
The rates of acute mania seen in RCTs of acute bipolar depression tend to hover around 5% for placebo, compared with 0%-50% for anti-depressants, depending on the study. The overall numbers tend to be lower than reported in observational studies. A possible reason for the lower rates is that patients who enter RCTs are a highly selected population: these persons typically are not severely ill, do not have psychiatric or medical comorbidities, and are highly compliant and motivated. Such persons may not possess some of the risk factors for antidepres-sant-induced mania (such as substance abuse comorbidity). Therefore, one should not place great emphasis on the actual frequency of acute mania seen in RCTs because this is likely an underestimate of the frequency in the real world. Furthermore, one should not pay attention to statistical significance (or P values) because such RCTs are woefully underpowered to assess acute mania. The main question is whether, given the limitations of the RCT design, antidepressants cause acute mania more frequently than placebo does.
Regarding antidepressant-induced mania, the meta-analysis described above reported no such risk (Gijsman et al. 2004). However, 58% of the subjects in this meta-analysis were derived from only one study that involved an olanzapine-fluoxetine combination (OFC), which single-handedly drove the results. Since "placebo" in the analysis was actually olanzapine or placebo, one can conclude only that antidepressants may not cause mania in the setting of antimanic agents. Two studies reported no mania at all, with placebo or antidepressant, suggesting possible measurement bias (inadequate assessment of manic symptoms). Another found an unfavorable relative risk of manic switch with imi-pramine versus paroxetine, but the two groups were combined for the meta-analysis, thus washing away the likely tricyclic-related increased risk. In that study, Nemeroff et al. (2001) found that switch into mania occurred more frequently at lithium levels less than 0.8 mEq/L. At these lower levels, 11% of imipramine-treated subjects and 5% of mood-stabilizer alone-treated subjects (spontaneous rate) became manic or hypomanic. At lithium levels greater than 0.8 mEq/L, none of the mood-stabilizer-alone subjects switched, whereas 8% of the imipramine-treated subjects experienced a switch. None of the 33 paroxetine-treated subjects experienced a switch into mania or hypomania (Nemeroff et al. 2001).
Thus, the available placebo-controlled RCTs do not shed much light on the question of which method of treatment is more likely to induce mania one way or the other, with the exception that TCAs appear to have a higher risk than other antidepressants.
How common is antidepressant-induced mania in the real world? To answer this question, one needs to look at well-designed observational studies. In fact, observational studies are often more useful than RCTs in identifying safety risks in medications. A number of studies have been conducted, but one of the best designed, in terms of prospective assessment of this issue with mood rating scales, found about a 20%-25% manic switch rate with SRIs, which was similar to TCAs in this study (Henry et al. 2001). Other studies have reported somewhat higher rates with TCAs, in the 30%-60% range (Goodwin and Jamison 1990). In another observational study, about 50% of patients with bipolar disorder experienced antidepressant-induced acute mania at some point in their lives (Ghaemi et al. 2004). Thus, it is likely that 20%-60% of persons with bipolar disorder type I will develop antide-pressant-induced acute mania at least once and even more frequently with TCAs than SRIs. Frequency of antidepressant-induced mania in bipolar disorder type II may be lower, although this topic has not been carefully studied.
The rate of antidepressant-induced mania in unipolar depression, in contrast, is much smaller than in bipolar disorder. For instance, in the study by Amsterdam et al. (1998), the manic switch rate in bipolar II disorder, though low, was four times higher than in unipolar depression (4% vs. 1%). Other observational studies tend to report no or very little antidepressant-induced mania in those for whom unipolar depression was appropriately diagnosed (Ghaemi et al. 2004).
Although definitive risk factors are not established, a number of studies suggest some likely factors for antidepressant-induced mania (Goldberg and Truman 2003). Antidepressant-induced manias have a significant irritability component, often resembling mixed rather than pure manic episodes (Stoll et al. 1994). This feature may be an important component of the potential for increased suicidality in some persons treated with antidepressants, because mixed episodes are associated with elevated suicidality (Dilsaver et al. 1994). Additionally, subjects with cyclothymia may convert to a type II illness when given antide-pressants (Akiskal et al. 1977). Hyperthymic personality (a chronic baseline hypomanic state) appears to increase the risk of antidepressant-induced mania (Henry et al. 2001). Importantly, current or past substance abuse appears to be a major predictor of antidepressant-induced manic switch (Goldberg and Whiteside 2002; Manwani et al. 2005). Young age may also be a risk factor. In one study (Biederman et al. 2000), the SRIs were associated with a high probability of inducing manic symptoms (hazards ratio = 3.0 [1.2-7.8], P = 0.02).
The mood state of the individual when he or she is given antidepres-sants may be an important determinant in the induction of mania. El-Mallakh (2001) reported a case of type I bipolar patient whose depression improved with the addition of bupropion without any adverse consequence, but who developed a mania when he took the same medication for smoking cessation while euthymic.
Lastly, treatment with antidepressant monotherapy, in the absence of a mood stabilizer, likely increases the risk of induction of acute mania (Ghaemi et al. 2004). This is most likely to occur in the setting of the mis-diagnosis of bipolar disorder as unipolar depression, which has been shown to occur in about 40% of persons with bipolar disorder (Hirschfeld et al. 2003). Due to misdiagnosis, there is a frequent delay of over 5 years for the diagnosis of type I illness and over a decade for the diagnosis of a type II illness after the initial contact is made with a mental health practitioner; by the time of correct bipolar diagnosis, 78% have already been prescribed antidepressants, often in monotherapy (Ghaemi et al. 2000).
Was this article helpful?
If you're wanting to learn about bipolar disorder... Then this may be the most important letter you'll ever read! You are about to take an in-depth look at bipolar disorder. It's all you need to know about bipolar disorder to help you or a loved one lead a normal life. It doesn't matter if you or a loved one have been recently diagnosed or been struggling with bipolar disorder for years - This guide will tell you everything you need to know, without spending too much brainpower!