Antipsychotic agents in the treatment of bipolar mania


  • MT is a former employee of Eli Lilly & Co. (2008) and his spouse is a current Eli Lilly & Co. employee and stockholder; and he has been a consultant to or received honoraria from AstraZeneca, Bristol-Myers Squibb, Eli Lilly & Co., GlaxoSmithKline, Johnson & Johnson and Wyeth. EV has received grants or honoraria as a consultant from Almirall, AstraZeneca, Bristol-Myers Squibb, Eli Lilly & Co., Esteve, Forest, GlaxoSmithKline, Janssen-Cilag, Jazz, Lundbeck, Novartis, Pfizer, Sanofi, Servier, Shering-Plough, and UBC; and is supported in part by grants from the Instituto Carlos III (Fondos de Investigacion Sanitaria and CIBER-SAM), and by the EU 7th Framework Programme (ENBREC).

Corresponding author:
Mauricio Tohen, MD, Dr PH, MBA Division of Mood and Anxiety Disorders The University of Texas Health Science Center at San Antonio 7703 Floyd Curl Drive, MC 7792 San Antonio, TX 78229, USA Fax: 210-567-3759 e-mail:


Objectives:  Antipsychotics have been widely used in the treatment of bipolar mania. The purpose of this manuscript was to briefly review the evidence of typical and atypical antipsychotics for the treatment of bipolar mania.

Methods:  A detailed literature review was conducted on the use of typical and atypical antipsychotics in the treatment of bipolar mania using standard search engines. A summary of the published literature on each agent is described followed by a discussion on the overall comparison of the different agents.

Results:  For typical antipsychotics, up until recently, there was a paucity of published evidence on their strengths and limitations in the treatment of bipolar mania. Recent studies have demonstrated clear evidence on the efficacy of haloperidol on the treatment of acute mania. The literature suggests a faster onset of action of haloperidol as compared to either lithium or atypical antipsychotics. A limitation of typical antipsychotics however, is the risk of tardive dyskinesia, extrapyramidal side effects and a possible increased risk of non-adherence. Evidence on the efficacy for atypical antipsychotics has been demonstrated for aripiprazole, clozapine, olanzapine, quetiapine, risperidone and ziprasidone. Limitations as regards the use of atypical antipsychotics include the risk of weight gain and dyslipidemia. Comparison among different atypical antipsychotics agents are difficult to determine as there are no conclusive head to head studies. There is also a paucity of studies comparing atypical antipsychotics with lithium.

Conclusions:  Evidence exists on the efficacy of both typical and atypical antipsychotics on the treatment of acute mania such that they are now clearly first-line along with lithium. An important limitation of the published literature is that most of the studies were designed to obtain regulatory approval for the different agents therefore the generalizability of the findings to clinical practice remains unclear.

The year 1949 marked the beginning of new hope for patients suffering from bipolar disorders. Cade’s findings on the use of lithium as a new treatment for bipolar mania are no doubt one of the most impactful events in the history of psychopharmacology. However, the use of lithium in clinical practice both in North America and Europe did not happen until the early 1970s (1). In fact, the clinical use of antipsychotic agents in the treatment of mania predates the use of lithium. In this article, we will review the available published evidence on the use of typical and atypical antipsychotic agents in the treatment of bipolar mania.

Typical antipsychotics

In 1952, Delay and Deniker (2) published their results on the use of chlorpromazine in agitated manic states. During the early/mid-1950s, typical antipsychotics began to be used in acute manic states (1). The unclear diagnostic boundaries between schizophrenia and bipolar disorder both in the United States and Europe during the three decades following Delay and Deniker’s findings makes it difficult to determine the full impact that antipsychotic agents had in the treatment of bipolar mania (3, 4). During the time that only typical antipsychotics (neuroleptics) were available, the advantages and disadvantages of neuroleptics compared to lithium were unclear (5). Some of the earlier comparative trials in the treatment of acute mania suggested superiority of neuroleptics over lithium, but a meta-analysis published in 1992 (6) found lithium to be more effective but with the important limitation of having a slower onset of action. Studies also suggested that neuroleptics have the advantage of controlling behavioral symptoms, a documented limitation of lithium (7). Before the atypical antipsychotics were available, neuroleptics were widely used in the treatment of acute mania. Chlorpromazine was the first antipsychotic to be approved by the U.S. Food and Drug Administration (FDA) for the treatment of bipolar mania. The first treatment guidelines for bipolar disorders published by the American Psychiatric Association in 1994, however, cautioned their use, especially in long-term treatment (8). In spite of this recommendation, neuroleptics were widely used; a meta-analysis that included publications through the year 2000 (9) found that 91% of inpatients and 65% of outpatients with bipolar mania were treated with typical antipsychotics.

Typical antipsychotics have for the most part been combined with lithium to treat mania (10). A small number of studies compared typical antipsychotics with each other but found no clear differences (11–13). Intermediate potency neuroleptics were frequently chosen (14), perhaps due to the perception that patients suffering form mood disorders were more sensitive to extrapyramidal adverse events.

Chlorpromazine and haloperidol are the best-studied typical antipsychotics in the treatment of bipolar mania (11–13, 15–23). Haloperidol has been compared as monotherapy to placebo, risperidone, olanzapine, quetiapine, ziprasidone, and aripiprazole (16–22), and as add-on treatment to lithium (10). The literature suggests that haloperidol has antimanic properties (16, 17) and a faster onset of antimanic action compared to atypical antipsychotics (19). Chlorpromazine has been studied only in one small, placebo-controlled trial (15) and a few comparative, randomized studies versus lithium, haloperidol, pimozide and clozapine (11–13, 23). Regarding dose, Baldessarini et al. (24) determined that for schizophrenia the therapeutic dose lies between 100–700 mg in chlorpromazine equivalents with a mean of 350 mg/day. For acute mania, a similar systematic review has not been conducted; however, Rifkin et al. (25) studied haloperidol dosing in a double-blind, fixed-dose study. Three haloperidol doses (10, 30, or 80 mg/day) were compared in a six-week study. The authors’ conclusion was that haloperidol doses greater than 10 mg/day did not provide additional advantages.

Adverse effects of typical antipsychotics

Clinicians have been concerned about the use of typical antipsychotics in bipolar disorder due to their adverse-event profile and the risk of noncompliance. An adverse event that has been considered unique to bipolar patients is the potential risk of inducing depression. Zarate and Tohen (14) reported an increased risk of relapse into depression and higher rates of discontinuation in patients with mania who received mood stabilizers combined with a typical antipsychotic compared with mood stabilizers alone. Other concerns about the safety of typical antipsychotics specific to mood disorders have been reported, specifically extrapyramidal symptoms (EPS) and tardive dyskinesia (26, 27), although no controlled studies have been designed to determine if indeed the risk of tardive dyskinesia is higher in mood disorder patients compared to patients with schizophrenia. A comprehensive review by Kane and Smith (27) estimated the risk of tardive dyskinesia to be 4% per year of treatment in patients with schizophrenia. The same group estimated that the risk in bipolar disorder may be twice as high (28). For neuroleptic-induced dystonia, one study reported the risk to be 25% for patients with mania compared to 5.9% for patients with schizophrenia (26). It is not known whether this differential vulnerability holds true for treatment with atypical antipsychotics. Cavazzoni et al. (29) conducted a meta-analysis of all olanzapine controlled trials, published and unpublished, conducted by the manufacturer until July 2001 comparing olanzapine, haloperidol, and placebo in the acute treatment of schizophrenia and bipolar mania. The investigators identified 18 studies, 13 of them comparing olanzapine versus haloperidol in patients with schizophrenia, 3 comparing olanzapine versus placebo in schizophrenia, 1 comparing haloperidol versus olanzapine in mania, and 2 comparing olanzapine versus placebo in mania. The authors compared the data for incidence of treatment-emergent EPS in patients with either schizophrenia or bipolar disorder (n = 4,417) and reported that haloperidol-treated patients with bipolar disorder appeared to be more vulnerable to the development of EPS than those with schizophrenia. However, olanzapine-treated patients suffering from bipolar disorder were no more likely to develop EPS than those suffering from schizophrenia. These findings suggest that it is possible that, compared to patients with schizophrenia, patients with bipolar disorder may be more sensitive to EPS when exposed to typical antipsychotics, but not to atypical antipsychotics.

Atypical antipsychotics

Prior to 2008, the FDA had approved five atypical antipsychotics for the treatment of acute bipolar mania: aripiprazole, olanzapine, quetiapine, risperidone, and ziprasidone.


Although not approved by the FDA, clozapine was the first atypical antipsychotic to be studied in the treatment of acute bipolar mania. To date, however, no double-blind, placebo-controlled clinical trials with clozapine in acute mania have been conducted. There are number of retrospective reviews (30–33) and prospective open-label studies with small numbers of patients that suggest that clozapine may have antimanic effects (32, 33). Prospective open-label trials also suggest that clozapine may be effective in patients with dysphonic mania (34) and in those who do not respond to lithium or valproate (35, 36). There is also a double-blind comparison study against chlorpromazine that suggested a faster onset of action for clozapine (23).


Aripiprazole is a partial agonist of dopamine D2/D3 and serotonin 5-HT1A receptors, an antagonist of 5-HT2A and histamine H1 receptors, and a moderate serotonin reuptake inhibitor.

In a three-week, double-blind study, patients (n = 262) with an acute manic or mixed episode were randomized either to aripiprazole or placebo (37). Aripiprazole had greater improved Young Mania Rating Scale (YMRS) scores (−8.2 versus −3.4 for placebo; p < 0.01). Response rate was also significantly greater (40% versus 19%; p < 0.01). The percentage of aripiprazole-treated patients achieving response was significantly greater than placebo-treated patients, and as early as Day 4 (14% versus 5%; p < 0.05). Akathisia was significantly higher with aripiprazole when compared to placebo (38, 39). Importantly, aripiprazole is the only atypical antipsychotic agent found to have superior response rate compared to haloperidol (50% versus 28.4%) in a 12-week comparative trial (22).

A recent study found aripiprazole as adjunctive treatment to mood stabilizers (lithium or valproate) to be more effective than mood stabilizers alone (22). Combination treatment demonstrated greater decrease in YMRS total scores compared to mood-stabilizer monotherapy (−13.1 versus −10.7, respectively; p < 0.01). Akathisia occurred significantly more frequent in the combination group (18.6% versus 5.4%, respectively; p < 0.001), but there was no significant difference in weight gain (+0.55 kg versus +0.23 kg, respectively).


Two short-term trials (three and four weeks) of olanzapine monotherapy compared to placebo in acute mania have been reported (40, 41). In both trials, patients treated with olanzapine showed significantly greater mean improvements in YMRS total score compared with placebo-treated patients [olanzapine versus placebo: −10.3 versus −4.9; p < 0.02 (40); and −14.8 versus −8.1; p < 0.001 (41)]. Adverse events that occurred significantly more often in olanzapine-treated patients than in placebo-treated patients included somnolence, dry mouth, dizziness, and weight gain (40, 41).

A study by Tohen et al. (19) compared olanzapine- to haloperidol-treated patients. The primary outcome remission rate, defined as YMRS scores ≤ 12 and Hamilton Depression Rating Scale (HAMD-21) scores ≤ 8 at Week 6 of treatment, were similar in both treatment groups (olanzapine, 52%; haloperidol, 46%); however, YMRS decrease in the first six weeks statistically favored haloperidol-treated patients over olanzapine-treated patients (−23.5 versus −21.3, respectively; p = 0.03), but there was no statistical difference at 12 weeks (−26.5 versus −26.8, respectively). Of note, for a subgroup of patients whose index episode did not present with psychotic features, rates of remission were significantly greater for olanzapine-treated patients compared with the haloperidol group (57% versus 42%; p = 0.04). Also of interest, in patients who entered the study without depression but became depressed during the course of the trial, haloperidol treatment was associated with a significantly shorter time to switch into depression (p = 0.04). Adverse events that were significantly more common in olanzapine-treated patients than in haloperidol-treated patients included somnolence, weight gain, infection, dizziness, and fever. On all measures assessing EPS, both patient-reported and scale-defined, significantly more haloperidol-treated patients experienced a worsening of symptoms.

Three trials have compared olanzapine with divalproex. The first trial (42, 43) evaluated patients hospitalized for acute bipolar manic or mixed episodes and consisted of a three-week acute phase (42), followed by a 44-week double-blind phase to assess maintenance of response (43). The mean change in the YMRS total score was −13.4 points for olanzapine-treated patients and −10.4 points for divalproex-treated patients (p = 0.03). The following adverse events were observed significantly more often with olanzapine- than divalproex-treated patients: dry mouth, increased appetite, weight gain, somnolence, tremor, speech impairment, neck rigidity, and tongue edema. A second randomized, 12-week, double-blind study compared olanzapine with valproate in patients with bipolar mania (44). In this study, the efficacy outcome baseline to endpoint (21 days) change utilizing the YMRS total score showed no statistical difference between treatment groups. Olanzapine-treated patients experienced more somnolence, weight gain, rhinitis, edema, and slurred speech than divalproex-treated patients. More recently, a placebo-controlled study compared olanzapine versus divalproex at three and 12 weeks in patients with mild to moderate nonpsychotic mania (45). The placebo arm was limited to three weeks. In patients who received olanzapine, the decrease in YMRS scores was significantly greater than in those who received placebo; no statistical difference was found between divalproex-treated patients compared to placebo-treated patients. At three weeks, there was no statistical difference between olanzapine and divalproex; however, at 12 weeks the decrease in YMRS scores was significantly greater for olanzapine compared to divalproex (−13.3 versus −10.7, respectively; p = 0.004). The authors indicated that a limitation of this study was the relatively low dose of divalproex received by patients (mean 864 mg/day), which may explain the lack of separation of divalproex from placebo. Olanzapine-treated patients had statistically significant increases in weight gain, glucose, triglycerides, cholesterol, uric acid, and prolactin than divalproex-treated patients.

Another recent study (46) compared olanzapine to lithium. Olanzapine-treated patients had a greater decrease in YMRS total scores (−24.63 versus −20.15, respectively; p = 0.013). A significantly greater proportion of olanzapine-treated patients achieved response (50% decrease in YMRS score) compared to lithium (odds ratio 2.71, 95% confidence interval: 1.07–6.87; p = 0.035), but no statistical difference was observed in remission rates (olanzapine 87%, lithium 73.2%; p = 0.073). However, significantly more olanzapine- than lithium-treated patients experienced at least one adverse event related to the study drug and significantly more olanzapine-treated patients had a clinically significant weight increase (≥ 7% of baseline weight) compared to lithium-treated patients.

Efficacy of olanzapine in combination with valproate or lithium for manic or mixed acute episodes was examined in a six-week study (47). The study included 344 patients with bipolar disorder (manic or mixed type) who had been inadequately responsive to at least two weeks of lithium or valproate treatment monotherapy. Patients were randomly assigned to receive adjunctive olanzapine or mood-stabilizer monotherapy. Patients who received adjunctive olanzapine had significantly greater mean improvements in YMRS total scores than those who received mood-stabilizer monotherapy (−13.1 versus −9.1, respectively; p = 0.003). Rates of clinical response, defined as a ≥ 50% improvement in YMRS total scores from baseline to end point, were also significantly higher with combination treatment (68% versus 45%, p < 0.001). Regarding adverse events, cotherapy patients were more likely than monotherapy patients to discontinue due to adverse events (cotherapy, 11%; monotherapy, 2%). Adverse events that occurred significantly more often in olanzapine cotherapy patients than in monotherapy mood-stabilizer patients included somnolence, dry mouth, weight gain, increased appetite, tremor, speech disorder, and elevated prolactin. Somnolence and weight gain occurred with sufficient severity to cause treatment discontinuation in the olanzapine cotherapy group. A recent trial, however, failed to prove any further benefit of the addition of olanzapine to carbamazepine compared to carbamazepine alone (48). Adverse events, however, such as somnolence, weight gain, and increased cholesterol were more common with combination treatment. Furthermore, the authors pointed out that the degree of increased cholesterol appeared to be higher in the olanzapine-carbamazepine group than that reported in other olanzapine monotherapy trials (48).


Quetiapine has been studied in acute mania in two monotherapy and two adjunctive therapy randomized, placebo-controlled trials (17, 49–52). A total of 302 patients with an acute manic episode participated in a double-blind trial randomized to quetiapine, haloperidol, or placebo (17). At Day 21, both quetiapine and haloperidol had significantly greater improved YMRS score (−12.29 versus −15.71, respectively) compared to placebo [−8.32; p < 0.01 (versus quetiapine) and p < 0.001 (versus haloperidol)]. At Day 84, difference from placebo was also significant for both treatments (−17.52 versus −18.92, respectively) compared to placebo (−9.48; p < 0.001) (17).

In a separate double-blind trial, quetiapine, lithium, and placebo were randomly administered to patients with bipolar mania. Quetiapine was significantly superior to placebo in reducing the YMRS score and similar to lithium (52). Somnolence, hypotension, and weight gain were the main adverse events reported to be more frequent with quetiapine compared to placebo (49).

Two randomized, double-blind, placebo-controlled studies (51, 52) were conducted to evaluate the efficacy and tolerability of quetiapine adjunct to lithium or divalproex in the treatment of acute mania. In the first study (51), the quetiapine-plus-mood-stabilizer group had a significantly greater reduction in the YMRS score when compared to the placebo-plus-mood-stabilizer group (−13.76 versus −9.93; p = 0.021). The response rate (reduction of at least 50% of the YMRS score) was significantly higher in the quetiapine-plus-mood-stabilizer group than in the placebo-plus-mood-stabilizer group (54.3% versus 32.6%; p = 0.005). Clinical remission (YMRS score < 12) was also significantly greater (45.7% versus 25.8%; p = 0.007) in the cotherapy group (51). In the second study, quetiapine plus mood stabilizer did not separate from mood stabilizer alone at study endpoint (52).


The first open-label prospective study suggesting antimanic effects of risperidone was published in 1996 (53). There are three published controlled studies assessing the antimanic effects of risperidone as monotherapy (16, 54, 55). In a three-week, multicenter, double-blind, placebo-controlled study (n = 259), risperidone significantly improved both YMRS and Clinical Global Impressions (CGI) scores (54). Improvement was significant from the third day of treatment onward (p < 0.01 versus placebo). Another three-week trial (n = 290) reported that those randomized to risperidone improved significantly starting at Day 3 compared to placebo (55). Response to treatment, defined as at least 50% decrease in YMRS score, was achieved in 73% and 36% of those randomized to risperidone and placebo, respectively (p < 0.001). The main adverse events of risperidone were dose-related EPS and hyperprolactinemia (55). In the third study (16), Smulevich et al. reported a three-week, controlled trial where patients were randomized to risperidone, haloperidol, or placebo, followed by a nine-week, double-blind trial of risperidone and haloperidol. Risperidone and haloperidol were similarly effective in the treatment of acute mania and both had statistically significant greater YMRS decrease compared to placebo (16).

Risperidone has also been studied as adjunct treatment to mood stabilizers (lithium, valproate, or carbamazepine). A three-week, double-blind, randomised, controlled trial studied mood stabilizers plus risperidone or placebo in the treatment of acute mania (56). At the study endpoint, YMRS scores improved by −14.5 and −10.3 in the risperidone plus mood stabilizer and mood-stabilizer monotherapy groups, respectively, but the difference was not statistically significant (p < 0.089), probably due to the effects of carbamazepine on risperidone’s plasma levels through hepatic enzyme induction. When risperidone plus lithium or valproate semisodium were compared to placebo plus lithium or valproate semisodium, YMRS scores improved, respectively, by −15.2 and −9.8 (p < 0.047). In another double-blind, placebo-controlled trial with adjunct haloperidol, risperidone, or placebo to mood stabilizers compared to mood stabilizers alone (18), both adjunct haloperidol and risperidone achieved significantly greater reductions in YMRS scores compared to the mood-stabilizer monotherapy group.


A three-week, double-blind trial randomized 210 patients with a manic or mixed episode either to ziprasidone or to placebo (57). Ziprasidone improved relative to baseline and placebo on the primary and most secondary efficacy measures at endpoint. Measures included were CGI (severity and improvement), Positive and Negative Syndrome Scale (PANSS), and Schedule for Affective Disorders and Schizophrenia-Change Mania Rating Scale (SADS-C MRS). Responders to treatment (at least 50% improvement on MRS) were 50% of the ziprasidone group and 35% of the placebo group (p < 0.05). Another three-week trial also found ziprasidone to be more effective than placebo (58). Somnolence and EPS were the most reported adverse events (58, 59). A third monotherapy, placebo-controlled trial, which also included a haloperidol arm, found significant superiority of ziprasidone over placebo, but lower efficacy versus haloperidol (up to 30 mg/day) at the three-week and 12-week endpoints (58). Ziprasidone was, however, better tolerated than haloperidol.

A three-week, double-blind trial (n = 205) compared ziprasidone as add-on treatment to lithium compared to lithium alone. This trial found no difference between the two treatments (60). Somnolence, EPS, dizziness, and agitation were more frequent in the group receiving ziprasidone and lithium (60).

Other antipsychotics

Paliperidone has shown efficacy in the treatment of mania in three recent placebo-controlled trials (61). Few studies with amisulpride in the treatment of mania have been published. A multicenter, open-label, randomized trial compared amisulpride with haloperidol in patients with mania taking valproate (62). Amisulpride was not superior to haloperidol as adjunctive treatment to valproate (62). Zotepine has only been studied in open-label, noncontrolled studies (63). Asenapine is not yet available for clinical use, but it has been studied in two placebo-controlled trials and one adjunctive study, with overall positive results (64).

Comparisons between pharmacological agents

Studies have shown that aripiprazole, quetiapine, olanzapine, risperidone, and ziprasidone are all superior to placebo. The only randomized, controlled study comparing two atypical agents showed no difference in the primary outcome between olanzapine and risperidone (65). Two outstanding systematic reviews on treatment for acute mania have been recently published (66, 67). Smith et al. (66) conducted a meta-analysis of all pharmacological agents licensed in the United States or United Kingdom for the treatment of acute mania compared against placebo. The authors included 13 studies (n = 3,089) comparing lithium, anticonvulsants (carbamazepine and valproate), haloperidol, and atypical antipsychotics (aripiprazole, quetiapine, olanzapine, and risperidone). All treatments were more effective than placebo. Regarding response (50% improvement), all treatments showed similar relative risks as shown in Table 1, with one small outlier study that compared valproate against placebo (68). The investigators pooled the results from studies of the atypical antipsychotics and estimated the reduction of the YMRS to be 70% better, and for all mood stabilizers to be twice as good compared to placebo. The investigators concluded that the effect sizes compared to placebo were similar for antipsychotics and mood stabilizers. Regarding safety and tolerability, the investigators concluded that carbamazepine and lithium may be more poorly tolerated. Scherk et al. (67) conducted a thoughtful, systematic review focusing on second-generation antipsychotic agents in the treatment of acute mania. The authors reviewed 24 studies (n = 6,187), both published and unpublished, that studied aripiprazole, amisulpride, clozapine, olanzapine, quetiapine, risperidone, ziprasidone, and zotepine, either monotherapy or in combination with mood stabilizers. These authors also came to the conclusion that atypical antipsychotics are more effective than placebo and that atypical antipsychotics are comparable to mood stabilizers. Importantly, based on their review, the authors concluded that adding an antipsychotic to a mood stabilizer was the most efficacious treatment for acute mania. Regarding tolerability, the authors identified EPS and somnolence as the main adverse events for atypical antipsychotics. Weight gain has also been a concern regarding the use of atypical antipsychotics, especially for clozapine and olanzapine. Torrent and colleagues (69) recently conducted a systematic review of weight gain associated with pharmacological treatments for bipolar disorder. The authors concluded that weight gain was highest for clozapine and olanzapine, but also present with the use of quetiapine, risperidone, valproate, and lithium, as shown in Table 2. In contrast aripiprazole, carbamazepine, lamotrigine, and ziprasidone did not appear to be associated with weight gain. The authors recommended careful monitoring of weight before and after treatment in patients with bipolar disorder. Related adverse events include increases in cholesterol, triglycerides, and glucose, which have been associated with the use of atypical antipsychotics. Recent guidelines by the International Society for Bipolar Disorders (70) recommend careful monitoring of lipids for all bipolar patients receiving an atypical agent.

Table 1.   Comparisons between atypical antipsychotics and placebo in acute mania
 Relative risk (95% confidence interval)
  1. Modified from Smith et al. (66).

 Brecher & Huizar 2003 (76)1.60 (1.16–2.21)
 Smulevich et al. 2005 (16)1.44 (1.07–1.93)
 Tohen et al. 1999 (40)2.00 (1.23–3.27)
 Tohen et al. 2000 (41)1.51 (1.05–2.17)
 Brecher & Huizar 2003 (76)1.22 (0.86–1.73)
 Paulsson & Huizar 2003 (77)1.91 (1.33–2.76)
 Vieta et al. 2002 (78)2.04 (1.61–2.59)
 Smulevich et al. 2005 (16)1.46 (1.10–1.95)
 Hirschfeld et al. 2004 (54)1.78 (1.22–2.58)
 Keck et al. 2003 (37)2.16 (1.40–3.33)
 Sachs et al. 2006 (39)1.63 (1.21–2.19)
 Bowden et al. 1994 (79)1.94 (1.15–3.29)
 Paulsson & Huizar 2003 (77)1.87 (1.29–2.71)
Valproate semisodium
 Pope et al. 1991 (68)5.18 (1.26–21.20)
 Bowden et al. 1994 (79)1.91 (1.19–3.06)
 Weisler et al. 2004 (80)1.85 (1.19–2.88)
 Weisler et al. 2005 (81)2.09 (1.52–2.87)
Table 2.   Weight change for atypical antipsychotics
AuthorsPopulationTreatmentDurationMean weight gain (kg)p value
  1. Modified from Torrent et al. (69). QUE = quetiapine; PBO = placebo; Li = lithium; VPA = valproate; OLZ = olanzapine; BP-I = bipolar I disorder; FLUOX = fluoxetine; HAL = haloperidol; RIS = risperidone; LMG = lamotrigine; ARI = aripiprazole; LOCF = last observation carried forward; NR = not recorded; NS = not significant; OFC = olanzapine-fluoxetine combination; NA = not applicable.

Bowden et al. 2005 (52)Phase bipolar mania (n = 302) patients randomized; QUE (n = 107); PBO (n = 97); Li (n = 98) QUE vs Li vs PBO12 weeksQUE: +2.6 (LOCF) and +3.3 (observed cases); PBO: −0.0008 (LOCF) and +0.3 (observed cases); Li: +0.7 (LOCF) and +1 (observed cases)<0.001 in the Li group; NS from PBO
Tohen et al. 2003 (43)VPA (n = 126); OLZ (n = 125)OLZ vs VPA47 weeksOLZ: +2.7; VPA: +1.2<0.001
DelBello et al. 2006 (82)Pediatric BP-I patients (n = 50) with a manic or mixed episodeQUE vs VPA4 weeksQUE: 4.4 ± 5.0; VPA: 3.6 ± 6.00.2
Tohen et al. 2000 (41) Manic or mixed were OLZ (n = 55) or PBO (n = 60)OLZ vs PBO4 weeksOLZ: +2.1 ± 2.8; PBO: 0.45 ± 2.30.002
Tohen et al. 2003 (83) Bipolar depression PBO (n = 377); OLZ (n = 370); OLZ + FLUOX (n = 86)OLZ vs FLUOX vs PBO8 weeksNROLZ versus PBO <0.001; OFC versus PBO >0.99
Tohen et al. 2002 (47) Patients achieving remission after treatment with OLZ + Li or VPA received Li or VPA plus either OLZ or PBO (n = 99)OLZ + Li/VPA vs Li or VPA18 monthsCombination: +2; Monotherapy: −1.8 NA
Tohen et al. 2003 (19) Comparison flexible dosing of OLZ (n = 234) and HAL (n = 219)OLZ vs HAL18 monthsOLZ: +2.82; HAL: +0.02<0.001
Biederman et al. 2005 (84) Preschool-age children with bipolar disorder (n = 31)RIS vs OLZ8 weeksRIS: 2.2 ± 0.4; OLZ: 3.2 ± 0.7NS
Brown et al. 2006 (85) BP-I patients, depressed (n = 410) OLZ + FLUOX vs LMG 7 weeksNR0.001
Perlis et al. 2006 (65) BP-I with non-psychotic acute manic or mixed episodes treated with OLZ (n = 165) or RIS (n = 164) OLZ vs RIS3 weeksOLZ: +2.46; RIS: +1.60.005
Guille et al. 2000 (86) BP-I patients (n = 42) Clozapine vs OLZ vs RIS12 weeksOLZ: +7.2 ± 6.2; RIS: +3.5 ± 5; Clozapine: no statistical power0.03
Keck et al. 2003 (57) 76 bipolar patients recently hospitalized and treated for a manic or mixed episodeARI vs PBO26 weeksARI: 1.7 ± 0.8; PBO: 0.5 ± 0.80.02
DelBello et al. 2006 (82)Pediatric patients (n = 15) QUE vs VPA6 weeksQUE: 4.2 ± 3.2; VPA: 2.5 ± 2.1NS
Khanna et al. 2005 (55) 290 BP-I patients with a current manic or mixed episodeRIS vs PBO3 weeksRIS: 0.06; PBO: 0.07NR
Vieta et al. 2005 (21) Bipolar mania (n = 347) ARI vs HAL12 weeksARI: +0.27; HAL: −0.10NS
Vieta et al. 2005 (87)Manic patients (n = 20)Amisulpride6 weeks0.4 ± 0.9NS
Tohen et al. 2005 (88)Bipolar patients (manic/mixed) (n = 431)OLZ vs Li52 weeksOLZ: 1.8; Li: −1.4<0.001


In this article, we have reported the evidence available for antipsychotic agents in the treatment of acute bipolar manic or mixed episodes. Importantly, most of the published studies were designed in order to obtain regulatory approval. The populations included in regulatory-oriented, placebo-controlled studies may not be representative of patients commonly treated in clinical settings, so the generalizability of their findings may be limited. Understanding the results of controlled trials with placebo or active control are no doubt needed to determine the overall efficacy and safety of pharmacological compounds, but they can only aid in the treatment of individual patients. In fact, despite the available number of treatment options, bipolar mania remains a difficult to treat condition (71, 72), even at first episode (73). Finally, nonpharmacological treatments including psychotherapy (74, 75) remain a core tool in the treatment of bipolar disorder, especially in the maintenance phase, once acute manic or mixed symptoms have abated and the patient is in better shape to benefit from strategies such as psychoeducation.