NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Butler M, Urosevic S, Desai P, et al. Treatment for Bipolar Disorder in Adults: A Systematic Review [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2018 Aug. (Comparative Effectiveness Review, No. 208.)
We identified 71 publications of 67 unique studies for acute mania that examined 28 separate drugs tested against 14 different comparators. These treatments and their comparators may have been single drug therapies or combination therapies of multiple drugs tested against either placebo monotherapies or other multiple drug therapies. The 67 studies combined into 56 treatment comparisons, 35 of which had only one study contribute information. Only three comparisons had four or more studies contributing. An additional 54 studies were excluded due to attrition higher than 50 percent.
The high attrition studies (greater than 50% were excluded because observed results among patients who complete a trial may not generalize to the entire patient population of interest if systematic differences between patients who do not complete the study and those who do (i.e., attrition is not random and/or likely due to bipolar disorder (BD) or treatment-relevant factors) occur. Moreover, if there are differential rates of attrition across study arms, or even similar rates but a different distribution of reasons for attrition, primary effect estimates and statistical inference can suffer from bias, potentially severe. The Last Observation Carried Forward (LOCF) method—by far the most common method used to address missing outcome data in the included studies—requires an assumption that the health-status of patients who dropped out of the trial would not have changed had future observations been recorded. This is a particularly heroic assumption in the context of withdrawal due to lack of efficacy or adverse events, not uncommon occurrences in the context of pharmaceutical treatment of patients with BD.23 When this assumption is inappropriate, use of LOCF methods can bias effect estimates. Moreover, estimates of standard errors will understate the true uncertainty surrounding effect estimates due to the added uncertainty of having to impute data, leading readers to believe the result is more precise than it actually is and potentially inflating the type-I error rate.24 This potential bias in the estimates of effect is even more problematic in studies with greater than 50 percent attrition that require imputing half or more of the data.
The results in this chapter for treatments for acute mania are organized by general drug category: antipsychotics, mood stabilizers, and drugs otherwise not specified. Within the antipsychotics section, results are presented by specific drug, then broken into single drugs compared to placebo or another drug, then, when appropriate, drug in combination with mood stabilizers compared to placebo or another drug. Likewise, the mood stabilizers and other drugs sections are presented first by single drug results followed by combination therapy results.
Antipsychotic Drugs for Acute Mania
Key Points
- Most antipsychotic drugs had few studies to contribute to findings. Studies for antipsychotics plus mood stabilizers were even more sparse and scattered.
- Low-strength evidence shows improved mania symptoms for all Food and Drug Administration (FDA)-approved antipsychotics, except aripiprazole, when compared to placebo for adults with bipolar I disorder (BD-I). For four of the antipsychotics we were able to provide a point estimate. However, most manic symptom improvements were of modest clinical significance, with values that were less than the minimally important difference (MID) but still large enough that a reasonable proportion of participants likely received a benefit.
- Low-strength evidence showed no statistical differences in acute mania symptoms between olanzapine and divalproex/valproate.
- The ability to draw stronger conclusions for antipsychotics was hindered by high attrition rates.
- Evidence was insufficient to draw conclusions regarding antipsychotic drugs alone compared to placebo or antipsychotic drugs plus mood stabilizers compared with another drug for BD-I for the primary outcomes of interest (response, symptom scores, and function).
- When reported, all comparisons tended to show no differences between groups in serious adverse events. Participants using atypical antipsychotics, except quetiapine, reported experiencing more extrapyramidal symptoms compared to placebo. Participants using haloperidol reported experiencing more extrapyramidal symptoms compared to other antipsychotics. Participants using olanzapine reported experiencing more clinically significant weight gain.
Eligible Studies for Antipsychotics
Eight antipsychotic drugs were examined in 47 publications of 43 unique studies for BD patients experiencing acute manic events. Of these, seven are FDA approved for use in adults with BD experiencing mania: aripiprazole, asenapine, cariprazine, olanzapine, quetiapine, risperidone, and ziprasidone. An additional unpublished study for aripiprazole plus mood stabilizers was also included for metaanalysis. Haloperidol, an antipsychotic treatment available since the late 1950’s and a World Health Organization listed essential medicine, was FDA approved in 1986 for schizophrenia (not for mania in adults with BD). All were examined as single drugs and all but cariprazine were also examined as a treatment combined with mood stabilizers. The populations tested were BD-I patients, which is understandable given the BD-I diagnosis requires history of just one episode of mania. Only one study (for quetiapine) included adults with bipolar II disorder (BD-II) or bipolar disorder not otherwise specificed (BD-NOS). No studies specifically assessed drug effectiveness in treatment of hypomania. The large majority of studies with usable outcomes were measured at 3 weeks duration.
Appendix E provides detailed evidence tables, summary risk of bias assessments, forest plots when appropriate, and assessments of strength of evidence for key comparisons and outcomes. A summary of findings with at least low-strength evidence for drug treatments for acute mania are provided in Table 6. Any intervention and comparison not listed in Table 6, or outcome not listed for an included intervention and comparison, was found to have an evidence base insufficient to draw conclusions.
Table 6
Summary of findings with at least low-strength evidence for antipsychotic drug treatments for acute mania.
Aripiprazole
We identified four unique randomized controlled trials (RCTs) of aripiprazole,47-50 and two eligible publications reporting two unique RCTs of aripiprazole plus mood stabilizers for acute mania with at least 3 weeks followup.51, 52 Three studies were assessed as moderate risk of bias and three were assessed as high. An additional 6 studies were excluded for attrition over 50 percent.53-58 All studies were funded by industry. Three studies compared aripiprazole to placebo48-50 and two compared to haloperidol.47, 48 Sample sizes ranged from 42 to 485 participants, and all participants were BD-I. One unpublished RCT was discovered—clinicaltrials.gov ID: NCT00046384—which compared aripiprazole with placebo, with a total of less than 60 patients. The trial was prematurely closed and allegedly did not produce results.
Aripiprazole Alone
Table 7 summarizes the population and major inclusion and exclusion criteria for each aripiprazole study for acute mania. Appendix E provides further detail.
Table 7
Population and inclusion criteria for studies of aripiprazole alone for acute mania.
Aripiprazole Alone Versus Placebo
Evidence was insufficient for all outcomes from three studies (n=823) to address whether aripiprazole was better than placebo for acute mania in adults with BD-I, due to high study limitations and imprecise data.48-50 Following AHRQ methods, random effect models for pooling data, which allow one to extend the findings to the general population, largely showed no difference between groups in response rates, manic symptom improvement, or withdrawal rates. If fixed effect models are used, which only allows inferences for the specific participants in the specific studies, symptom improvements were seen. However, the improvement may not be clinically meaningful based on values that are less than the MID. The Young Mania Rating Scale (YMRS) mean difference of 3.85 (95% CI 2.27, 5.44) is less than the MID of 6, and the Clinical Global Improvement (CGI) score mean difference of 0.44 (95% CI 0.25, 0.63) is less than the MID of 1. There were no differences between groups for serious adverse events.
Aripiprazole Alone Versus Active Control
Evidence was insufficient for all outcomes from two studies (n=674) to address whether aripiprazole was better than haloperidol for acute mania in adults with BD-I, due to mostly high study limitations and imprecise data.47, 48 Studies reported no differences between groups for response or remission rates and mixed results as to which drug was favored. Akathisia and extrapyramidal symptoms were reported lower in participants using aripiprazole versus haloperidol; other harms or withdrawal outcomes were mixed.
Aripiprazole Plus Mood Stabilizers
Table 8 summarizes the bipolar type and major inclusion and exclusion criteria for each aripiprazole plus mood stabilizers study for acute mania. Appendix E provides further detail.
Table 8
Population and inclusion criteria for aripiprazole plus mood stabilizers studies for acute mania.
Aripiprazole Combination Versus Placebo
Evidence was insufficient for all outcomes from two RCTs (n=752), one published and one unpublished, to address whether aripiprazole with a mood stabilizer (e.g., lithium, valproic acid, or divalproex) was better than placebo for acute mania in adults with BD-I, due to high study limitations, inconsistency, and imprecise data.52 Results were mixed, with the published study reporting aripiprazole plus mood stabilizers showed improvements in response, remission, mania symptoms, and CGI, while the unpublished study reported no difference between groups. Both studies reported no differences between groups in withdrawal rates, serious adverse events, or rates of akathisia.
Aripiprazole Combination Versus Active Control
Evidence was insufficient for all outcomes from one small RCT (n=42) to address whether aripiprazole with mood stabilizers was better than haloperidol with mood stabilizer for acute mania in adults with BD-I, due to high study limitations, unknown consistency, and imprecise data.51 The study reported response rates in the 70 to 90 percent range but no differences between groups for mania symptom (YMRS) or CGI. Participants using aripiprazole reported experiencing more weight gain while participants using haloperidol reported experiencing more extrapyramidal symptoms.
Asenapine
We identified three eligible publications reporting three unique RCTs of asenapine and one RCT examining asenapine with lithium or valproic acid for acute mania with at least 3 weeks followup.25-27 59 Two studies were assessed as low to moderate risk of bias27 59 and two were assessed as high.25, 26 No additional studies were excluded for greater than 50 percent attrition. All were funded by industry. All interventions used a placebo comparator and two also compared to olanzapine.25, 26 Sample sizes ranged from 324 to 489 and all followed participants with BD-I.
Asenapine Alone
Table 9 summarizes the bipolar type and major inclusion and exclusion criteria for each study of asenapine alone for acute mania. Appendix E provides further detail.
Table 9
Population and inclusion criteria for studies of asenapine alone for acute mania.
Asenapine Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from three studies (n=956) showed asenapine improved mania symptoms (YMRS mean difference 4.37, 95% CI 1.27, 7.47) and CGI (mean difference 0.5, 95% CI 0.29, 0.71) compared to placebo after 3 weeks of treatment, although the improvement was about two-thirds the MID.25-27 Response and remission were not significantly different between groups. We found low-strength evidence that asenapine had a lower rate of withdrawal due to lack of efficacy than placebo (moderate study limitations, imprecision). However, low-strength evidence also showed that placebo had a lower rate of withdrawal due to adverse events than asenapine (moderate study limitations, imprecision). Overall withdrawal was less in the asenapine group, and favored asenapine over placebo, but results were not statistically significant. There were no differences between groups for serious adverse events, although participants with asenapine had significantly more extrapyramidal symptoms than those on placebo.
Asenapine Alone Versus Active Control
Evidence was insufficient for all outcomes from two studies (n=763) to address whether asenapine was better than olanzapine for acute mania in adults with BD-I, due to high study limitations and imprecise data.25, 26 Studies reported olanzapine showed a greater response rate but no differences in remissions. Serious adverse events were not different between arms, although participants using asenapine tended to withdraw at higher rates.
Asenapine Plus Mood Stabilizers
Table 10 summarizes the bipolar type and major inclusion and exclusion criteria for each asenapine plus mood stabilizers study for acute mania. Appendix E provides further detail.
Table 10
Population and inclusion criteria for asenapine plus mood stabilizer studies for acute mania.
Asenapine Combination Versus Placebo
Evidence from this single study of 324 participants with BD-I was rated as insufficient for all outcomes due to moderate study limitations, unknown consistency, and imprecision. The study reported the asenapine with lithium or valproate group showed improvement in manic symptom (YMRS) and CGI but no differences between groups for response or remission rates.
Cariprazine
We identified three eligible publications reporting three unique RCTs of cariprazine for acute mania with at least 3 weeks followup.28-30 All were assessed as low to moderate risk of bias and used a placebo comparator. No studies were excluded for greater than 50 percent attrition. All were funded by industry. Sample sizes ranged from 238 to 497 and all recruited participants with BD-I. Table 11 summarizes the bipolar type and major inclusion and exclusion criteria for each study. Appendix E provides further detail.
Table 11
Population and inclusion criteria for studies of cariprazine alone for acute mania.
Cariprazine Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from three studies (n=1,047) showed cariprazine improved response (OR 2.14, 95% CI 1.08, 4.23) and remission (OR 1.95, 95% CI 1.45, 2.63) rates, as well as mania symptoms (YMRS mean difference 5.38, 95% CI 1.84, 8.92) and CGI-BP-S (mean difference 0.54, 95% CI 0.35, 0.73), compared to placebo after 3 weeks of treatment.28-30 No differences were seen in withdrawal rates. There were no differences between groups for serious adverse events, although participants using cariprazine had more extrapyramidal symptoms than those using placebo.
Olanzapine
We identified 15 eligible publications reporting 13 unique RCTs of olanzapine and seven eligible publications reporting five unique RCTs of olanzapine with a mood stabilizer for acute mania with at least 3 weeks followup.25, 26, 31, 33, 34, 44-46, 60-72 Five were assessed as low risk of bias, five as moderate, and eight as high. Fourteen studies reported being funded by industry. An additional sixteen studies were excluded for greater than 50 percent attrition.73-88 Nine studies used a placebo comparator, while 14 used active comparators. Two studies of olanzapine with mood stabilizers did not use a placebo in place of olanzapine.46, 69 Sample sizes ranged from 30 to 560 and most reported recruiting participants with BD-I; one study restricted participants to a current DSM-IV mixed episode.69
Olanzapine Alone
Table 12 summarizes the bipolar type and major inclusion and exclusion criteria for each study of olanzapine alone for acute mania. Appendix E provides further detail.
Table 12
Population and inclusion criteria for studies of olanzapine alone for acute mania.
Olanzapine Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from five RCTs (n=1,199) showed olanzapine was better for acute mania than placebo in response (OR 1.99, 95% CI 1.29 to 3.08) and remission (OR 1.75, 95% CI 1.19 to 2.58) rates. Mania symptom improvement was close to a MID (YMRS, mean difference 4.9, 95% CI 2.34 to 7.45).25, 26, 31, 33, 34, 45 CGI trended toward improvement for olanzapine but did not reach significance. Low-strength evidence (moderate study limitations, imprecision) also showed overall withdrawal and withdrawal due to lack of effect were lower for olanzapine. Withdrawal for adverse events did not differ between groups. While serious adverse events did not differ by group, participants using olanzapine reported more extrapyramidal symptoms and weight gain (at least 7 percent increase) than those using placebo.
Olanzapine Alone Versus Active Control
Low-strength evidence (moderate study limitations, imprecision) from four RCTs (n=867) showed no statistically significant difference in outcomes between olanzapine versus divalproex or valproate for acute mania in adults with presumed BD-I.33, 44-46, 63, 66 No differences were noted in response or remission rates (n=635), mania symptoms (YMRS, n=750), CGI (n=578), or withdrawals (n=867). No differences were noted in serious adverse events. However, one study noted participants receiving olanzapine experienced more clinically important weight gain (at least 7%) than those receiving divalproex;33 a trend toward greater weight gain in olanzapine groups was noted in the other studies as well.
Evidence was insufficient for all outcomes from three RCTs (n=210) to address whether olanzapine was better for acute mania than lithium in adults with presumed BD-I, due to moderate study limitations, inconsistency, and imprecision.60, 61, 64 The studies reported mixed results for response, mania symptom improvement (YMRS), or CGI. Withdrawals and adverse events tended to show no differences between groups.
Evidence was also insufficient for all outcomes to address whether olanzapine was better than risperidone (one RCT, n=329),89 or haloperidol (one RCT, n=453), due to single studies of moderate to high study limitations and imprecision.67 The studies reported no differences between groups for response, remission, symptom improvement, function, or withdrawals over 3 weeks. No differences between groups were noted in serious adverse events. However, participants using olanzapine reported more weight gain while participants using haloperidol reported more akathisia.
Results for olanzapine versus asenapine were reported in the asenapine versus active comparator section above (e.g., evidence was insufficient for olanzapine compared to asenapine).
Olanzapine Plus Mood Stabilizers
Table 13 summarizes bipolar type and major inclusion and exclusion criteria for each olanzapine plus mood stabilizers study for acute mania. Appendix E provides further detail.
Table 13
Population and inclusion criteria for olanzapine plus mood stabilizers studies for acute mania.
Olanzapine Combination Versus Placebo/No Placebo
Evidence was insufficient for all outcomes from four RCTs to address whether olanzapine plus mood stabilizers was better for acute mania than mood stabilizers alone for adults with BD-I, due to high study limitations and imprecision. Two studies examined olanzapine plus carbamazepine (n=118)70 or lithium/valproate (n=344).66 The studies showed mixed results for response or remission rates, but both reported olanzapine improved symptoms. Two other studies examined olanzapine plus divalproex (n=202)69 or valproate (n=80)46 compared to the mood stabilizer alone without a placebo present. One study reporting response and remission rates reported results favoring olanzapine, while both reported improvements in mania symptoms. Participants receiving olanzapine reported greater frequency of clinically important weight gain. No other differences between groups were noted in serious adverse events.
Olanzapine Combination Versus Active Control
Evidence was insufficient for all outcomes from one RCT (n=83) to address whether olanzapine plus lithium was better for acute mania than chlorpromazine plus lithium for adults with BD-I, due to a single study and imprecision.68 The study reported no difference between groups for all outcomes including symptomatic recovery, response, remission, depressive symptoms, and CGI. No differences were noted in serious adverse events or clinically significant weight gain.
Quetiapine
We identified six eligible publications reporting six unique RCTs of quetiapine and two eligible publications reporting two unique RCTs of quetiapine plus mood stabilizers for acute mania with at least 3 weeks followup.35-39, 90-92 Two studies were assessed as low risk of bias, three as moderate, and three as high risk. Three additional studies were excluded for greater than 50 percent attrition.93, 94, 95 All studies were funded by industry. Five studies used placebos and three used active comparators. Sample sizes ranged from 39 to 493. All enrolled participants with BD-I; one study restricted participants to a current episode of mania but a history of manic or mixed episodes in the last 5 years. Another small study enrolled participants with mild to moderate hypomania or mild mania regardless of type of BD.39
Quetiapine Alone
Table 14 summarizes the bipolar type and major inclusion and exclusion criteria for each study of quetiapine alone for acute mania. Appendix E provides further detail.
Table 14
Population and inclusion criteria for studies of quetiapine alone for acute mania.
Quetiapine Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from four RCTs showed improved response rates (OR 2.07, 95% CI 1.39 to 3.09, n=1,007) and manic symptom improvement close to the MID (YMRS, mean difference 4.92, 95% CI 0.31 to 9.53, n=699) for participants receiving quetiapine.35-38 Evidence was insufficient to address remission rates (n=699) due to fewer studies of higher risk of bias contributing to the outcome. Low-strength evidence (moderate study limitations, imprecision) showed CGI improved for participants using quetiapine, but the improvement was about half the MID (mean difference 0.54, 95% CI 0.35 to 0.74, n=806). Withdrawal due to lack of efficacy was lower for quetiapine but overall withdrawal and withdrawal due to adverse events did not differ between groups (low-strength, n=1,007). Most studies reported no serious adverse events and no differences between groups for extrapyramidal symptoms. One small study (n=41) enrolling patients with mild to moderate hypomania or mild mania also found participants using quetiapine showed improvements in manic symptoms (YMRS) and CGI.39 Weight gain greater than 7 percent was infrequently reported but tended to be more common in participants using quetiapine.
Quetiapine Alone Versus Active Control
Evidence was insufficient for all outcomes from one RCT (n=199) to address whether quetiapine was better for acute mania than haloperidol in adults with BD-I, due to a single study and imprecision.37 The study reported no differences between groups for response or remission rates, manic symptom improvement, CGI, or withdrawals. Participants using haloperidol reported more extrapyramidal symptoms; otherwise, no differences in serious adverse events were noted.
Evidence was insufficient from two RCTs (n=456) to address whether quetiapine was better for acute mania than lithium for in adults with BD-I, due to high study limitations, inconsistency, and imprecision.36, 90 Both studies reported response and remission rates, and change in manic symptoms; one trial reported benefit with quetiapine90 and one reported no difference.36 One trial reported no difference in CGI.36 Both reported no difference in withdrawals or serious adverse events.
Quetiapine Plus Mood Stabilizers
Table 15 summarizes bipolar type and major inclusion and exclusion criteria for each quetiapine plus mood stabilizers study for acute mania. Appendix E provides further detail.
Table 15
Population and inclusion criteria for quetiapine plus mood stabilizer studies for acute mania.
Quetiapine Combination Versus Placebo
Evidence was insufficient for all outcomes from two RCTs (n=570) to address whether quetipine plus mood stabilizers was better for acute mania than mood stabilizer alone in adults with BD-I, due to high study limitations and imprecise data.91, 92 The studies reported quetiapine added to mood stabilizers improved response and remission rates, manic symptoms (YMRS), and CGI score. Both studies reported no differences between groups in withdrawal rates and serious adverse events, and results for extrapyramidal symptoms were mixed.
Risperidone
We identified three eligible publications reporting three unique RCTs of risperidone and one RCT for risperidone plus mood stabilizers for acute mania with at least 3 weeks followup.40, 41, 96 97 One study was assessed as low risk of bias and three studies were assessed as moderate. All were funded by industry. Four additional studies were excluded for greater than 50 percent attrition.84, 94, 98, 99 Three studies used placebo comparators and two studies used active comparators. Sample sizes ranged from 45 to 438 and enrolled BD-I participants.
Risperidone Alone
Table 16 summarizes the bipolar type and major inclusion and exclusion criteria for each study of risperidone alone for acute mania. Appendix E provides further detail.
Table 16
Population and inclusion criteria for studies of risperidone alone for acute mania.
Risperidone Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from two studies (n=584) showed risperidone was better for acute mania than placebo for adults with BD-I.40, 41 Although we were unable to conduct a meta-analysis based on the two studies, the finding in favor of risperidone was consistent across the studies for response rate, manic symptom improvement (YMRS), and CGI. No serious adverse events were reported. However, participants using risperidone experienced more extrapyramidal symptoms than those using placebo.
Risperidone Alone Versus Active Control
Evidence was insufficient for all outcomes to address whether risperidone performed better than an active comparator for acute mania, due to moderate study limitations, inconsistency, and imprecision. Findings were mixed from two studies comparing risperidone to haloperidol in adults with BD-I.41, 96 Those who received risperidone also had lower total scores on the extrapyramidal symptom rating scale at 3 weeks compared to those who received haloperidol. Withdrawal rates were similar between groups. One study compared risperidone to lithium, finding no difference between groups in bipolar outcomes and extrapyramidal symptoms at 4 weeks.96 Withdrawal rates were not reported.
Results for risperidone versus olanzapine were reported in the olanzapine versus active comparator section above and were determined to yield insufficient evidence.
Risperidone Plus Mood Stabilizers
Table 17 summarizes bipolar type and major inclusion and exclusion criteria for each risperidone plus mood stabilizers study for acute mania. Appendix E provides further detail.
Table 17
Population and inclusion criteria for risperidone plus mood stabilizer studies for acute mania.
Risperidone Combination Versus Placebo
Evidence from this single study of 150 participants with BD-I was rated as insufficient due to unknown consistency and imprecision. The study reported that adding risperidone to mood stabilizers improved response rates and CGI and a trend toward reduced manic symptoms (YMRS). No differences were reported in adverse events; however, participants using risperidone experienced more extrapyramidal symptoms.
Ziprasidone
We identified two eligible publications reporting two unique RCTs of ziprasidone and one RCT of ziprasidone plus mood stabilizers for acute mania with at least 3 weeks followup.42, 43, 100 Two studies were assessed as moderate risk of bias, and one was high. All were funded by industry. Two additional studies were excluded for greater than 50 percent attrition.101, 102 Sample sizes ranged from 197 to 680.
Ziprasidone Alone
Table 18 summarizes the bipolar type and major inclusion and exclusion criteria for each study of ziprasidone alone for acute mania. Appendix E provides further detail.
Table 18
Population and inclusion criteria for studies of ziprasidone alone for acute mania.
Ziprasidone Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from two studies (n=402) showed ziprasidone was better for acute mania than placebo in adults with BD-I.42, 43 Although we were unable to conduct a meta-analysis, the finding in favor of ziprasidone was consistent across the studies for response rate, manic symptom improvement (YMRS), and CGI. Withdrawal due to lack of effect also was lower for the ziprasidone group, while no differences were seen for overall withdrawal or adverse events. Serious adverse events were reported in one study, with no difference between groups. However, in one study participants using ziprasidone experienced more extrapyramidal symptoms than those using placebo.42
Ziprasidone Plus Mood Stabilizers
Table 19 summarizes bipolar type and major inclusion and exclusion criteria for each ziprasidone plus mood stabilizers study for acute mania. Appendix E provides further detail.
Table 19
Population and inclusion criteria for ziprasidone plus mood stabilizer studies for acute mania.
Ziprasidone Combination Versus Placebo
Evidence from a single study of 680 participants with BD-I was rated as insufficient due to a single high risk of bias study and imprecision. The study reported no differences between groups in manic symptom (YMRS) and CGI. No differences were reported in adverse events, however participants using high dose ziprasidone experienced more extrapyramidal symptoms.
Haloperidol
We identified two eligible publications reporting two unique RCTs of haloperidol for acute mania with at least 3 weeks followup.37, 41 One study was assessed as moderate risk of bias41 and one was assessed as high.37 Both were funded by industry. One additional study was excluded for greater than 50 percent attrition.73 Both studies used a placebo comparator. Sample sizes ranged from 299 to 438 and recruited participants with BD-I.
Haloperidol Alone
Table 20 summarizes the bipolar type and major inclusion and exclusion criteria for each study. Appendix E provides further detail.
Table 20
Population and inclusion criteria for studies of haloperidol alone for acute mania.
Haloperidol Versus Placebo
Evidence was insufficient for all outcomes from two studies (n=483) to address whether haloperidol was better for acute mania than placebo in participants with BD-I, due to high study limitations and imprecision.37, 41 Studies reported results generally favored haloperidol. Neither study reported serious adverse events.
Haloperidol Versus Active Control
Results for haloperidol versus aripiprazole were reported in the aripiprazole versus active comparator section above and yielded insufficient evidence.
Mood Stabilizers for Acute Mania
Key Points
- Studies for mood stabilizers were sparse and scattered.
- Evidence was largely insufficient to draw conclusions regarding mood stabilizers compared to placebo or other drugs for BD-I for the primary outcomes of interest (response, symptom scores, and function).
- Low-strength evidence showed lithium increased response and remission rates and manic symptom improvement in BD-I participants with acute mania compared to placebo.
- When reported, all comparisons tended to show no differences between groups in serious adverse events. Participants using carbamazepine reported experiencing more severe rash and adverse events compared to placebo.
- The ability to draw stronger conclusions was hindered by high attrition rates.
Eligible Studies for Mood Stabilizers
Four mood stabilizers, all FDA approved for use in patients with bipolar disorder experiencing mania, were examined in 12 publications of 12 unique studies for BD patients with acute mania. All were tested as single drugs: carbamazepine, divalproex/valproate, lamotrigine, and lithium. All studies enrolled adults with BD-I. Only one study (for lithium) also included adults with BD-II. There were no studies assessing drug effectiveness in treatment of hypomania. The large majority of studies with usable outcomes were measured at 3 weeks duration.
Appendix F provides detailed evidence tables, summary risk of bias assessments, forest plots when appropriate, and assessments of strength of evidence for key comparisons and outcomes. A summary of findings with at least low-strength evidence for mood stabilizers for acute mania are provided in Table 21. Any intervention and comparison not listed in Table 21, or outcome not listed for an included intervention and comparison, was found to have an evidence base insufficient to draw conclusions.
Table 21
Summary of findings with at least low-strength evidence for mood stabilizers for acute mania.
Carbamazepine
We identified four RCTs examining carbamazepine for acute mania with at least 3 weeks followup.104-107 One study was moderate risk of bias and three were high. Three additional studies were excluded for greater than 50 percent attrition.108-110 Three studies were funded at least in part by industry. One study used placebo104 and three used active comparators.105-107 Sample sizes ranged from 30 to 443. All trials recruited participants with BD-I. Table 22 summarizes the bipolar type and major inclusion and exclusion criteria for each study. Appendix F provides further detail.
Table 22
Population and inclusion criteria for carbamazepine for acute mania.
Carbamazepine Alone Versus Placebo
Evidence was insufficient for all outcomes from one pooled analysis (n=443) of two high risk of bias trials to address whether carbamazepine was better for acute mania than placebo in adults with BD-I, due to high study limitations and imprecision.104 The study reported improvements in participants receiving carbamazepine in response rate, manic symptoms (YMRS) and CGI. Withdrawal for lack of efficacy and adverse events was lower for carbamazepine, but not overall withdrawals. Participants using carbamazepine experienced more frequent severe rash.
Carbamazepine Alone Versus Active Control
Evidence was insufficient for all outcomes from two small RCTs (n=82) for carbamazepine compared to lithium106, 107 and one small RCT (n=30) for carbamazepine versus valproate for acute mania in adults with BD-I, due to high study limitations and imprecision.105 When reported, the studies generally reported no differences between groups for response rates, manic symptoms, CGI, or withdrawal rates. Participants receiving carbamazepine reported more adverse events. Evidence was also insufficient from one small RCT (n=30) for carbamazepine compared to valproate due to single study and imprecision.
Divalproex/Valproate
We identified two RCTs examining divalproex for acute mania with at least 3 weeks followup.33, 111 One study was low risk of bias and one was high. Both studies were funded by industry. Two additional studies were excluded for greater than 50 percent attrition.112, 113 Both studies used placebo and one used active comparators. Sample sizes ranged from 364 to 521. One small study examining valproate versus no placebo was also included.46 Seven additional valproate studies were excluded for greater than 50 percent attrition.114-120 Table 23 summarizes the bipolar type and major inclusion and exclusion criteria for each study. Appendix F provides details.
Table 23
Population and inclusion criteria for divalproex/valproate for acute mania.
Divalproex Alone Versus Placebo
Evidence was insufficient for all outcomes from two RCTs (n=670) to address whether divalproex sodium was better for acute mania than placebo in adults with BD-I, due to moderate to high study limitations, inconsistency, and imprecision.33, 111 Results were mixed for response, remission, and symptoms at 3 weeks. Both studies reported no difference in CGI or function (Global Assessment Score (GAS)), withdrawal, or serious adverse events.
Evidence was also insufficient for all outcomes from one small study (n=79) whether valproate plus olanzapine was better for acute mania than olanzapine alone in adults with BD-I, due to single study and imprecision. The study reported improvement in manic symptoms (YMRS) and CGI.
Divalproex Alone Versus Active Control
Results for divalproex versus olanzapine were reported in the olanzapine versus active comparator subsection of the antipsychotic section above (e.g., low-strength evidence for no difference in remission or response rates, or improvements in manic symptoms or function).
Lamotrigine
We identified a single small, industry-funded, moderate risk of bias RCT examining lamotrigine for acute mania with at least 3 week followup.121 Seven additional studies were excluded for attrition over 50 percent.53, 114, 115, 122-125 Table 24 summarizes the bipolar type and major inclusion and exclusion criteria. Appendix F provides details.
Table 24
Population and inclusion criteria for lamotrigine for acute mania.
Lamotrigine Alone Versus Active Control
Evidence was insufficient for all outcomes to address whether lamotrigine was better for acute mania than lithium in adults with BD-I, due to single study and imprecision. The study reported no differences between group in bipolar symptoms or response. No serious adverse events were reported and withdrawal rates were similar between groups.
Lithium
We identified three RCTs36, 96, 126 and one meta-analysis that pooled individual patient data from four RCTs103 examining lithium for acute mania with at least 3 weeks followup. One study was low risk of bias, two moderate, and one high. Five additional studies were excluded for greater than 50 percent attrition.53, 58, 117, 118, 127 All studies were funded by industry. Two studies used placebo and all used active comparators. Sample sizes ranged from 45 to 876. Table 25 summarizes the bipolar type and major inclusion and exclusion criteria for each study. Appendix F provides detail.
Table 25
Population and inclusion criteria for lithium for acute mania.
Lithium Alone Versus Placebo
Low-strength evidence (moderate study limitations, imprecision) from one RCT and one meta-analysis of independent data from 4 RCTs (n=847) showed lithium increased response and remission rates in BD-I participants compared to placebo for acute mania.36, 103 Using data available to pool from one RCT and two individual RCTs reported in the meta-analysis, lithium improved manic symptoms essentially to the level of MID (YMRS, MD 5.81, 95% CI 2.21, 9.4; n=643). Withdrawal rates did not differ by group. Serious adverse events were inconsistently reported and showed mixed results.
Lithium Alone Versus Active Control
Evidence was insufficient for all outcomes from one RCT (n=270) to address whether lithium was better for acute mania than divalproex in adults with BD-I, due to a single study and imprecise data.126 The study reported response rate, symptoms (YMRS), CGI, and withdrawals did not differ between groups. One of two measures of remission showed benefit for divalproex. No differences in frequency of serious adverse events between groups were noted.
Evidence was insufficient for all outcomes from one small RCT (n=30) to address whether lithium was better than haloperidol, due to a single study and imprecise data.96 The study reported no differences between groups in manic symptoms (YMRS) and CGI. Serious adverse events were not reported.
Risperidone versus lithium and risperidone versus haloperidol comparisons are discussed in the risperidone subsection of the antipsychotics section above. Also discussed above in the antipsychotics section are olanzapine versus lithium and quetiapine versus lithium. Overall, while all comparisons were assessed as having insufficient evidence, studies generally reported no differences between the antipsychotic drug and lithium.
(The topiramate comparisons will be discussed in the following section.)
Drugs Not Approved by FDA for Acute Mania in Bipolar Disorder
Key Points
- Ten drugs were examined for acute mania in BD: allopurinol, celecoxib, donepezil, dipyridamole, endoxifen, gabapentin, paliperidone, tamoxifen, topiramate, and oxcarbazepine, some in combination with mood stabilizers.
- Low-strength evidence showed paliperidone improved manic symptoms over placebo in adults with BD-I, although the improvement was not a clinically important difference (n=763). Participants using 12 mg paliperidone reported more common akathisia and dystonia.
- Low-strength evidence showed topiramate was not significantly different from placebo for symptom improvement, and participants using placebo withdrew less for adverse events (n=876) in adults with BD-I. In addition, low-strength evidence showed lithium significantly improved manic symptoms compared to topiramate (n=453) in adults with BD-I, although participants receiving lithium withdrew more for adverse events.
- Low-strength evidence showed allopurinol plus mood stabilizers/other psychotropic medications did not differ significantly from mood stabilizers alone for manic symptom or CGI improvement or overall withdrawals (n=355) in adults with BD-I.
- Evidence was largely insufficient to draw conclusions for all other nonapproved FDA drugs for BD-I for the primary outcomes of interest (response, symptom scores, and function).
Eligible Studies for Drugs Not Approved by FDA
Sixteen unique studies examined nine other drugs for patients experiencing manic events.35, 103, 128-141 Four studies were assessed as low risk of bias,103, 129, 130, 135 ten were moderate,128, 132, 134, 136, 137, 139-141 and four were assessed as high.35, 131, 133, 138 Three additional studies were excluded for greater than 50 percent attrition.142-144 Eight studies were funded or assisted by industry.35, 103, 128, 135, 138-141 All but three studies128, 138, 139 used a placebo comparator. Sample sizes ranged from 27 to 876. Appendix G provides detailed evidence tables, a summary of risk of bias assessments, and assessments of strength of evidence for key comparisons and outcomes. A summary of findings with at least low-strength evidence for other drugs not approved by FDA for acute mania are provided in Table 26. Any intervention and comparison not listed in Table 26, or outcome not listed for an included intervention and comparison, was found to have an evidence base insufficient to draw conclusions.
Table 26
Summary of findings with at least low-strength evidence for drugs not approved by FDA for acute mania.
Table 27 summarizes the bipolar type and major inclusion and exclusion criteria for each study.
Table 27
Population and inclusion criteria for drugs not approved by FDA for acute mania.
Drugs Not Approved by FDA Versus Placebo
Twelve unique trials35, 129-137, 140, 141 and one pooled analysis of a further four trials103 examined nine drugs versus placebo. Five studies examined the drugs as a single drug,35, 103, 129, 134, 140 while eight were added to mood stabilizers or other current psychiatric medications.130-133, 135-137, 141 Studies ranged from 3 to 12 weeks long. All enrolled BD-I participants.
Low-strength evidence (moderate study limitations, imprecision) from two studies (n=763) showed paliperidone was better than placebo for improvement in mania symptoms (YMRS).35, 140 However, for the highest dose of 12 mg in the moderate risk of bias study, the improvement may not be clinically meaningful based on values that are less than the MID. The study reported YMRS mean difference of 3.4 (p=0.025), which is less than the MID of 6. While a dose response was suggested, authors stated results were driven largely by participants in India, who comprised only 10 percent of the analysis set. Low-strength evidence (moderate study limitations, imprecision) showed no statistically significant differences between groups for withdrawal for lack of efficacy. Evidence for CGI and response and remission rates was insufficient due to moderate study limitations, inconsistency, and imprecision. With the exception of more common akathisia and dystonia EPS symptoms for 12 mg paliperidone versus placebo, no differences in serious adverse events were noted Appendix G provides further detail.
Topiramate versus placebo was examined in a pooled analysis of four trials (n=876).103 Low-strength evidence (high imprecision) showed no differences between topiramate and placebo for manic symptoms (YMRS) or withdrawal due to lack of efficacy for adults with BD-I. Additionally, overall withdrawals and withdrawals due to adverse events were lower in the placebo group (low-strength evidence, high imprecision). No differences in severe adverse events between groups were reported. Appendix G provides further detail.
Evidence was insufficient for all outcomes for the two drugs celecoxib129 and tamoxifen,134 examined as single drug versus placebo for acute mania, due to single studies and imprecision.
For adjunctive medications, low-strength evidence (moderate study limitations, imprecision) from four RCTs (n=355) showed no significant differences between allopurinol plus mood stabilizers compared to mood stabilizers alone in manic symptoms (YMRS), CGI, or overall withdrawals.130-132, 136 Evidence was insufficient for response (high study limitations, imprecision)131, 136 and remission (moderate study limitations, inconsistent, imprecision) rates.130, 136 No serious adverse event were reported. Appendix G provides further detail.
Evidence was insufficient for all outcomes for dipyridamole,136 donepezil,137 or gabapentin133 plus lithium versus placebo largely due to single studies and imprecision. Evidence was also insufficient for all outcomes for one study of topiramate plus mood stabilizers versus mood stabilizers alone, although the general finding of no significant differences between groups was similar to the findings for topiramate as single drug.135 Likewise, one study of paliperidone plus mood stabilizers, while in itself providing insufficient evidence, repeated the general finding of no significant differences between groups observed in comparison of paliperidone as monotherapy versus placebo.141
Drugs Not Approved by FDA Versus Active Control
Six trials examined drugs versus active comparators in BD-I participants, each a unique comparison.35, 103, 136, 138, 139 Study sizes ranged from 30 to 388 and ran from 3 to 12 weeks.
Low-strength evidence (high imprecision) from a pooled analysis of individual patient data from two trials (n=453) of topiramate versus lithium for adults with BD-I with acute mania showed manic symptoms (YMRS) improved more with lithium and the difference was at the MID level (6.14, 95% CI 3.94, 8.34).103 Overall withdrawals and withdrawals due to lack of efficacy did not differ between groups (low-strength evidence). However, less participants receiving topiramate withdrew due to adverse events (7% vs. 3%). There were no differences in severe adverse events between lithium and topiramate groups.
Evidence was insufficient for all outcomes to address if endoxifen was better for acute mania in adults with BD-I than divalproex (unknown consistency, imprecise),128 or if paliperidone extended release was better than olanzapine139 or quetiapine (high study limitations, unknown consistency, imprecise).35
Evidence was insufficient for all outcomes to address if allopurinol plus lithium was better for acute mania in adults with BD-I than dipyridamole plus lithium (moderate study limitations, unknown consistency, imprecise), or if topiramate plus risperidone was better than divalproex plus risperidone (high study limitations, unknown consistency, imprecise).138
Interpreting the Findings for Drugs for Acute Mania
All FDA-approved antipsychotics, except aripiprazole, when compared to placebo improved mania symptoms for adults with BD-I (low-strength evidence). For four of the antipsychotics we were able to provide a point estimate. Lithium also reached low-strength evidence for improving mania symptoms, however, studies for carbamazepine, divalproex/valproate, and lamotrigine failed to reach sufficient evidence due to too few studies and imprecise results. Likewise, evidence was insufficient to draw conclusions for the efficacy of antipsychotics added to mood stabilizers.
Except for the finding that lithium improved mania symptoms better than topiramate (low-strength evidence), evidence from studies of drugs compared to other drugs, whether as single drug or drug combinations, for treatment of acute mania was also insufficient to draw conclusions. Our ability to draw conclusions was hampered by the small number of studies and sample sizes to allow confidence in findings of no differences between groups. Study designs generally tested for superiority of one drug over the other, rather than noninferiority of the two drugs. With noninferiority tests, if the relative equivalence of the performance of two drugs is not demonstrated strongly enough, nonequivalence cannot be ruled out; that is, the treatment effects of the two drugs are too different.
Only two small studies attempted to address efficacy and harms for specific populations of interest, pregnant women with BD (lamotrigine), and BD patients with hypomania (quetiapine). Unfortunately, results for the effect of quetiapine treatment for patients with hypomania were not reported separately from patient with mild mania, thus no conclusions can be made. Similarly, the single observational study for pregnant women provided insufficient evidence to address whether lamotrigine provided benefits. Because of the weak evidence, there was little to be gained from the very few studies that did attempt post-hoc analysis of subgroups. Post-hoc analyses cannot reach the same level of strength of evidence due to the inherent higher study limitations from studies that generated low-strength evidence for main findings would. Given the generally high levels of attrition observed in the included studies, results of any subgroup analysis of such a restricted set are even more suspect.
Adverse events were somewhat consistently reported for extrapyramidal symptoms, and clinically significant weight gain of greater than 7 percent, but otherwise variably reported. The harms findings from the included placebo-controlled studies were consistent with information currently reported by FDA labels. While most studies reported no differences between groups in studies comparing drugs to drugs, we noted a general pattern of participants receiving atypical antipsychotics experiencing fewer extrapyramidal symptoms than participants receiving other medications.
The seventeen studies examining efficacy and comparing drugs to drugs of ten other medications, either as single drug or added to other psychiatric medications, largely yielded insufficient evidence due to a single study for each specific comparison, small sample sizes, and/or inconsistent findings.
There were a few exceptions, such as a low-strength evidence that lithium improved manic symptoms more than topiramate, although topiramate had lower rates of withdrawal due to adverse events than lithium. There was also low-strength evidence for no group differences in examined outcomes for topiramate versus placebo and allopurinol plus mood stabilizers/lithium/other psychiatric medications versus these other medications alone. Low-strength evidence supported that paliperidone improved manic symptoms more than placebo, although the improvement was not clinically significant since it did not reach the MID.
Several issues impact the applicability of the studies. Over three quarters of the studies also excluded participants experiencing a first manic episode and most enrolled participants were 30 to 50 years of age. Moreover, given the inclusion criteria and actual participant characteristics, it is not clear if the current findings extend to populations with first manic episodes, current comorbid substance use, or pregnant or nursing women with BD I, or older adults with BD-I.
- Drug Treatments for Acute Mania - Treatment for Bipolar Disorder in Adults: A Sy...Drug Treatments for Acute Mania - Treatment for Bipolar Disorder in Adults: A Systematic Review
Your browsing activity is empty.
Activity recording is turned off.
See more...