by Richard H. Weisler, MD; Amir H. Kalali, MD; Andrew J. Cutler, MD; Thomas D. Gazda, MD; and Lawrence Ginsberg, MD

Dr. Weisler is from Duke University Medical Center, Durham, North Carolina, and University of North Carolina at Chapel Hill Departments of Psychiatry, Raleigh, North Carolina; Dr. Kalali is from Quintiles, Inc. and University of California, San Diego, California; Dr. Cutler is from Department of Psychiatry, University of Florida; Florida Clinical Research Center, LLC, Maitland, Florida; Dr. Gazda is from Banner Behavioral Health Hospital, Scottsdale, Arizona; and Dr. Ginsberg is from Red Oak Psychiatry Associates, Houston, Texas.


This study was supported by funding from Shire Development Inc.

Financial Disclosures

Dr. Weisler has received research support from, has been a speaker for, and/or has been a consultant to Abbott, the Agency for Toxic Substances and Disease Registry/Centers for Disease Control and Prevention, AstraZeneca, Biovail, Bristol-Myers Squibb, Cephalon, CoMentis, Corcept, Eisai, Eli Lilly, Forest, GlaxoSmithKline, Janssen, Johnson and Johnson, Lundbeck, MediciNova, Merck, the National Institute for Mental Health (NIMH), New River, Novartis, Organon, Pfizer, Saegis, Sanofi-Synthelabo, Schwabe, Shire, Solvay, Synaptic, TAP Pharmaceutical Products, UCB Pharma, Vela, and Wyeth, and holds or has held stock in Bristol-Myers Squibb, Merck, and Pfizer; Dr. Kalali is a consultant to Shire Pharmaceuticals; Dr. Cutler has received research grants, is a consultant to, or is a speaker for Abbott Pharmaceuticals, AstraZeneca Pharmaceuticals, Bristol-Myers Squibb Co., Cephalon, Dalnippon Sumitomo Pharma, Eli Lilly & Co., Forest Labs, GlaxoSmithKline, Janssen Pharmaceutica, JDS Pharmaceuticals, Johnson & Johnson PRD, Memory Pharmaceuticals, Novartis Pharmaceuticals, Organon, Otsuka America Pharmaceuticals, Pfizer, Sanofi-Synthelabo, Sanofi-Aventis, Seprecor, Shire Pharmaceuticals, Solvay Pharmaceuticals, Supernus Pharmaceuticals, Vanda Pharmaceuticals, and Wyeth Pharmaceuticals; Dr. Gazda is a consultant to AstraZeneca, Bristol-Myers Squib, Eli Lilly & Co., Janssen, Johnson and Johnson, Pfizer, Sanofi-Aventis, and Shire Pharmaceuticals; and Dr. Ginsberg has received research support from and is a consultant to Shire Pharmaceuticals.


Objective: To compare the efficacy and safety of carbamazepine extended-release capsules (CBZ-ERC) administered twice daily (BID) versus once daily for the treatment of manic symptoms associated with bipolar I disorder in adults.
Design: This was a Phase IIIb, randomized, double-blind, parallel-group, multicenter, 12-week study. Subjects were randomized (1:1) to CBZ-ERC once daily at bedtime (QHS) or BID. Dosing was initiated at CBZ-ERC 200mg/d and titrated to achieve an optimal dose (target dose, 800mg/d; maximum dose, 1600mg/d). The primary efficacy outcome variable was the Young Mania Rating Scale (YMRS). The Hamilton Rating Scale for Depression, 21-item version (HAM-D21), Montgomery-Åsberg Depression Rating Scale (MADRS), Clinical Global Impressions Scale–Bipolar Version (CGI-BP), and time to remission were secondary outcome variables. Safety measures included recording of adverse events, physical examination, vital signs (blood pressure, pulse rate, and weight), and clinical laboratory and electrocardiogram (ECG) parameters. Results: BID and QHS dosing were equally effective in improving symptoms of bipolar disorder, as measured with the YMRS, HAM-D21, MADRS, and CGI-BP. Both BID and QHS dosing significantly improved total scores on the YMRS, HAM-D21, and MADRS at all time points without statistically significant differences between groups. All three components of the CGI-BP improved during the study, and a large percentage of subjects in both groups achieved remission without significant differences between groups. Both CBZ-ERC regimens appeared to be safe and well tolerated. Conclusion: These results suggest QHS dosing may be a safe and effective alternative to BID dosing of CBZ-ERC for treating manic episodes for many adults with bipolar I disorder, although additional studies are needed to confirm this finding.

Key Words

Carbamazepine, extended release, bipolar disorder, manic symptoms, BID dosing, QHS dosing


Bipolar disorder is a chronic condition characterized by episodes of mania, depression, or mixed states (simultaneous manic and depressive symptoms),[1 ]with a lifetime incidence of bipolar I disorder estimated at 3.3 percent in the United States.[2] A recent meta-analysis reported that patients with bipolar disorder typically experience a recurrence rate of 0.66 episodes annually, each lasting approximately three months.[3] Approximately 71 percent of patients with bipolar I disorder have a comorbid substance abuse disorder,[4] 93 percent have an anxiety disorder,[4] 65 percent have a personality disorder,2 and 30 to 50 percent of patients with bipolar disorder will attempt suicide during their lifetime.[5,6]

While the rate of suicidality differs between individuals with bipolar depression (79.3%), depressive-mania (56.3%), and pure mania (2.3%),7 up to 18.9 percent of overall deaths in individuals with manic-depressive illness are due to suicide;[8] this rate is approximately 15 times higher than that of the general population.[9] Average annual suicide rates associated with bipolar disorder are estimated at one percent, a 60-fold higher annual rate of 0.015 percent observed in the international population.[10] Further, it has been shown that the suicide rate increases to between 23 and 26 percent early in the course of bipolar illness,[11] and the rate of premature death and disability among patients with bipolar disorder is second only to depression among neuropsychiatric disorders as a cause of a decreased healthy life worldwide.[12]

The US Food and Drug Administration (FDA) on January 31, 2008, issued an alert for healthcare professionals on suicidality and anticonvulsant agents.[13] This class-related guidance concluded that “all patients who are currently taking or starting on any antiepileptic drug should be closely monitored for notable changes in behavior that could indicate the emergence or worsening of suicidal thoughts or behavior or depression.”[13] It is important to note that for safety reasons the FDA meta-analysis of 199 placebo-controlled trials included all anticonvulsants in the analysis whether they were known to be effective in mood disorders (carbamazepine extended-release capsules [CBZ-ERC], divalproex, or lamotrigine), ineffective, or of unknown efficacy (gabapentin, topiramate, zonisamide, levetiracetam, tiagabine, pregabalin, or oxcarbazepine). Antidepressants frequently used for treatment of bipolar disorder, despite very limited positive data and the potential for mood switches, in addition to atypical neuroleptics with antidepressant properties also carry similar FDA alerts. Thus, when possible, it is important for clinicians to carefully monitor patients with bipolar disorder who are beginning or continuing any type of pharmacotherapy for suicidality.

It is our belief that effective treatment with FDA-approved anticonvulsants for bipolar disorder can ease suffering for many patients, thus possibly reducing the long-term risk of suicidality; however, additional studies are warranted. It is also important for clinicians to remember that lithium treatment has been associated in some monotherapy and augmentation studies with statistically significant reductions in suicidality and completed suicides.[14,15] For this reason, it may be desirable for clinicians to use lithium either alone or adjunctively for many patients where suicide is a concern.

Treatment options for bipolar disorder have recently expanded beyond lithium—once the most widely prescribed mood stabilizer—to include a number of anticonvulsant and antipsychotic medications,[16–18] with a recent survey demonstrating that anticonvulsant prescriptions (17%) now exceed those for lithium (8%) for use as mood stabilizers in the treatment of bipolar disorder.[19] The anticonvulsant agent carbamazepine (CBZ) has long been considered a therapeutic option for bipolar disorder, with early studies in the US[20,21] generating interest in the agent as a mood stabilizer. A meta-analysis of early double-blind trials comparing cbz and lithium reported similar acute antimanic efficacy.[22]

Until recently, most studies evaluating cbz in bipolar disorder were small and used immediate-release preparations that required 3 or 4 daily doses. However, short-term (3-week) multicenter, placebo-controlled trials have now reported significant improvements in manic symptoms in bipolar I patients with beaded CBZ-ERC administered twice daily (BID),[23,24] and a six-month, open-label study comprising participants from two three-week studies demonstrated a low rate of relapse (14.3%, n=11/77) with CBZ-ERC administered BID.[25]

CBZ-ERC received FDA approval for treatment of acute manic and mixed episodes associated with bipolar disorder in December, 2004.[26] Although labeled for BID dosing, anecdotal reports indicate use of once-daily CBZ-ERC regimens, and a retrospective chart review reported similar safety and efficacy with once-daily regimens of CBZ-ERC compared with BID administration.[27] The objectives of this study were to evaluate the efficacy, safety, and tolerability of once daily at bedtime (QHS) versus BID administrations of CBZ-ERC for the treatment of manic symptoms in adults with bipolar I disorder for a duration of up to 12 weeks.


Study design. This Phase IIIb, randomized, double-blind, parallel-group, multicenter study was conducted in accordance with all applicable regulations and the International Conference on Harmonisation (ICH) Good Clinical Practice (GCP) Guideline E6. All protocols, informed consent documents, relevant supporting information, and subject recruitment information were submitted to, and reviewed and approved by, a centralized Institutional Review Board (IRB), BioMed IRB, before site initiation. Each subject provided written informed consent before completing any study-related procedures.

Subjects. Male or non-pregnant female outpatients, aged 18 years or older, who met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV)[1] criteria for bipolar I disorder—most recent episode manic or mixed—were eligible for the study if they met the following inclusion criteria: history of one or more previous manic or mixed episode; screening Young Mania Rating Scale (YMRS)[28] score of 16 or greater; and no comorbid illness that could have affected efficacy, safety, or tolerability of the study drug, or interfered with the subject’s study participation. Females had to agree to adhere to an acceptable contraceptive protocol during the study.

Exclusion criteria included a history of lack of therapeutic response to CBZ for treatment of bipolar I disorder or known or suspected intolerance or hypersensitivity to the study drug; DSM-IV criteria for ultra-rapid cycling (i.e., >6 mood episodes/year); acute risk for suicidal or violent behavior or history of a serious suicide attempt requiring medical intervention; presence of another primary Axis I disorder, borderline or antisocial personality disorder, or nonaffective psychotic disorder; presence of clinically significant hepatic or renal disease that could affect action, absorption, or disposition of the investigational agent or affect clinical or laboratory assessments; history or presence of any serious, severe, or unstable illness that may prevent the subject from completing the study; and history of alcohol or other substance abuse or dependence (except caffeine or nicotine) within three months of screening. Pregnant or lactating females were excluded from study participation.

Study periods. The study consisted of four study periods (screening, double-blind titration, observation [dose maintenance], and follow-up), conducted over a maximum of 18 weeks.

Screening. Eligible subjects provided voluntary written informed consent before completing study-related procedures. Screening was conducted 2 to 10 days before randomization to determine subject eligibility. Diagnosis of bipolar I disorder, mixed or manic, was based on psychiatric evaluation and available medical records. Subjects were screened for inclusion and exclusion criteria, and demographic information was documented. Additional evaluations and assessments conducted at screening and reviewed before randomization included a complete medical, psychiatric, and medication history; physical examination, including vital signs (blood pressure and pulse rate), weight, height, and electrocardiogram (ECG); blood collection for chemistry, hematology, thyroid-stimulating hormone, and serum lithium and valproate levels; urinalysis, urine pregnancy test for females of childbearing potential, and urine drug screen; and baseline efficacy measurements (YMRS, Clinical Global Impressions Scale–Bipolar Version [CGI-BP],[31] Hamilton Rating Scale for Depression, 21-item version [HAM-D21],[29] and Montgomery-Åsberg Depression Rating Scale [MADRS]).[30]

Double-blind titration. Following screening, eligible subjects were randomized (1:1) to one of two treatment schedules (CBZ-ERC administered QHS or BID). To explore a different titration schedule for outpatients than used in the acute inpatient trials, dosing was initiated in all subjects entering the double-blind dose-titration period at CBZ-ERC 200mg/d and, as tolerated and clinically indicated, titrated up by 200mg every 3 to 4 days to achieve an optimal dose (target dose, 800mg/d; maximum dose, 1600mg/d). Study medication was dispensed at clinic visits. Dose reduction was permitted at any time during the treatment period to improve tolerability. Efficacy and safety assessments conducted on Days 1, 7, 14, 21, and 28 during dose titration included the following: YMRS, CGI-BP, HAM-D21, and MADRS; vital signs and weight; and review of adverse events (AEs), drug adherence, and concomitant medications. In addition, ECG; blood draw for chemistry, hematology, and thyroid-stimulating hormone; and urinalysis were evaluated at Day 28.

Observation (dose-maintenance). An eight-week observation period followed dose titration, with efficacy and safety assessments conducted at Weeks 6, 8, and 12, or early termination. YMRS, CGI-BP, HAM-D21, MADRS, vital signs, weight, and review of AEs, concomitant medication, and drug adherence were assessed at Weeks 6, 8, and 12, or at early termination. Blood draw for chemistry, hematology, and thyroid-stimulating hormone and urine pregnancy test in females of childbearing potential were performed at weeks 8 and 12, or at early termination.

Follow-up. During the 30 days after administration of the last study drug dose (Week 12) or after the date of early withdrawal, all subjects were contacted to collect follow-up information regarding ongoing AEs, new serious AEs, or related non-serious AEs.

Study drug. CBZ-ERC is a three-bead (immediate-, extended-, and enteric-release) capsule formulation of carbamazepine, USP (United States Pharmacopeia). All doses were administered using 200mg capsules. Placebo capsules were matched to CBZ-ERC size and color. Patients were instructed to swallow the medication whole, without chewing, dividing, or crushing. The study medication was taken BID, with morning and evening meals, with placebo administered for the morning dose to subjects who were randomized to the QHS-dosage regimen.

A subject was not permitted for study entry if a concomitant mood stabilizer or antipsychotic drug dose was altered within two weeks of the study initiation, or if an alteration was likely to occur during the course of the study. If needed, however, the dosages of concomitant mood stabilizers permitted by protocol—lithium, valproate, olanzapine, risperidone, and quetiapine—could in some instances be adjusted after the optimal CBZ-ERC dose was achieved; adjustments were made if needed according to clinical response, tolerability, and pharmacokinetic and pharmacodynamic reactions. Prohibited concomitant medications included the following: current (?3 months of study) or regular use of medications that could affect the condition being studied, the action, absorption, or disposition of the investigational agent, or clinical or laboratory assessments (except benzodiazepines and the mood stabilizers and antipsychotics permitted by the protocol); mood stabilizers or antipsychotics that had been changed ?2 weeks prior to study initiation or that were likely to require alteration during the study; clozapine ?3 months prior to screening; fluoxetine ?1 month prior to screening; other antidepressants or lamotrigine ?2 weeks prior to screening; antipsychotics (except olanzapine, risperidone, or quetiapine) or any injectable antipsychotic ?1 month prior to screening.

Evaluation criteria. Efficacy assessments. The primary efficacy assessment was the mean change from baseline in the YMRS total score in the intention-to-treat (ITT) population. Designed to assess severity of symptoms associated with the manic state of bipolar disorder, the YMRS is a validated 11-item scale.[28] Total scores, derived from the sum of 11 YMRS questions, range from 0 (least severe) to 60 (most severe).

Secondary efficacy assessments included the HAM-D21, MADRS, CGI-BP, and time to remission. The HAM-D21 is a clinician-administered, validated instrument that measures the presence and severity of depression in adults.[29] The total score analysis comprised the first 17 items; scores on the remaining four items were assessed but not included in the primary final analysis. The MADRS is a clinician-administered 10-item instrument designed to detect the presence and severity of depression in adults.[30] Based on clinical interviews, items were rated on a scale of 0 to 6. The CGI-BP is a modification of the CGI, specifically designed to assess global illness severity at baseline (CGI-BP-S) and change (CGI-BP-C) following treatment for manic, depressive and overall symptoms in patients with bipolar disorder.31 In this study, severity of illness was measured during the assessment period with each item on the CGI-BP-S scored on a scale of 1 (normal, not ill) to 7 (very severely ill). Change in illness was measured in comparison with the baseline visit, graded on a scale of 1 (very much improved) to 7 (very much worse) on the CGI-BP-C. Time to remission was defined as the first post-baseline visit YMRS total score ?12.

Safety assessments. The safety assessments included AEs, physical examination, laboratory parameters, vital signs, and ECG. The medication history included both medications that were taken during the course of the study and those taken three months before screening. An AE was defined as any unfavorable or unintended sign, symptom, disease, or exacerbation of a preexisting condition associated temporally with use of the investigational agent, but that did not necessarily imply a causal relationship with the investigational agent. AEs (serious, nonserious, related, and unrelated) were recorded from the time of informed consent through the end-of-treatment exposure, and during the 30 days after the last dose of study drug. Serious AEs were those that resulted in death; were life threatening; required inpatient hospitalization or prolongation of existing hospitalization; resulted in persistent or significant disability or incapacity; or resulted in a congenital abnormality or birth defect. AEs were categorized by the investigator according to intensity, outcome, and relationship to the investigational agent. A treatment-emergent adverse event (TEAE) was an AE that occurred on or after the first day of double-blind dose of study medication. A treatment-related AE was a TEAE that was considered by the investigator to be causally related to the study drug.

Samples for chemistry, hematology, urinalysis, and thyroid-stimulating hormone were obtained at screening, Weeks 6 and 8, and the final visit or study termination. Chemistry parameters included sodium, potassium, calcium, urea, creatinine, albumin, total protein, lactic dehydrogenase (LDH), aspartate amino transferase (AST), alanine amino transferase (ALT), alkaline phosphatase (ALP), gamma glutamyl transferase (GGT), total bilirubin, glucose, and total cholesterol. It is important to note that fasting was not required before obtaining chemistry samples. Hematology parameters included hemoglobin, hematocrit, red blood cell count, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, and total and differential white blood cell and platelet counts. Urinalysis parameters included glucose, blood, protein, and pH, as well as a drug screen. Serum lithium and valproate levels were assessed at screening and monitored as clinically indicated throughout the study, but they may not have always been trough levels. The significance of any laboratory changes from baseline was determined by the investigator.

A physical examination was performed at screening and follow-up, with significant changes from baseline recorded as AEs. Height was recorded at baseline, and weight was recorded at each study visit. Vital signs (blood pressure after 5 minutes in sitting position and pulse rate) were measured at all study visits, and clinically significant deviations from screening values were recorded as AEs. A 12-lead ECG was performed at screening, Day 28, and the final study visit; and a urine pregnancy test was performed on all females of childbearing potential at screening, baseline, Days 28 and 56, and the final study visit.

Statistical methods. The safety population comprised all subjects who received one or more doses of study medication. All randomized subjects who received one or more doses of study medication, had a baseline YMRS evaluation, and greater than or equal to one post-randomization YMRS comprised the ITT population. Safety assessments were performed on the safety population and efficacy assessments on the ITT population. Sample size estimates were based on the YMRS total score, with 51 subjects per group (102 total) determined to be sufficient to ensure a power of 80 percent to reject the null hypothesis that QHS and BID groups were not equivalent.

Efficacy assessments. The last observation carried forward (LOCF) method was used for the ITT analysis of the primary and secondary efficacy outcome variables. For the total YMRS scores—the primary efficacy variable—summary statistics were presented for each time point and the change from baseline to each time point for each treatment group. Differences in the change from baseline between and within treatment groups were analyzed using a two-way analysis of covariance (ANCOVA) model, including treatment (QHS vs. BID) and site as main factors, and the baseline YMRS total score as a covariate. A 95-percent confidence interval (CI) was constructed around the difference in the change in YMRS total score between patients receiving QHS versus BID dosing.

Before analysis of the change on the CGI-BP was determined, scores were grouped into two categories: “Very Much Improved” and “Much Improved” responses were grouped into an “Improved” category, and the remaining responses were grouped as “Not Improved.” To illustrate the difference in the proportion of patients improved at endpoint between QHS and BID dosing in the ITT population, the Improved and Not Improved groups were analyzed using the Cochran-Mantel-Haenszel (CMH) test. A 95-percent CI was constructed around the difference in the proportion of patients in the QHS vs. BID groups who were in the Improved category.

Summary statistics at baseline and endpoint, as well as the change from baseline to endpoint, were presented for each treatment group for the HAM-D21 (total score on the first 17 items) and MADRS (total score). Differences in the changes from baseline between and within treatment groups were analyzed using a two-way ANCOVA model, with factors of treatment (QHS vs. BID), site, and corresponding baseline score as covariates.

Time to remission was defined as the number of days from first treatment to the first post-baseline occurrence of remission (YMRS score ?12) or premature study withdrawal. The median time to initial remission was estimated using the Kaplan-Meier method, and a 95-percent CI was presented around the median. A two-sided log-rank p value <0.05 was considered significant.

Safety assessments. AEs that occurred on or after the first double-blind dose of study medication (TEAEs) were summarized by numbers and percentages of patients reporting the occurrence. Investigators’ terms used to identify the occurrences were coded using the Medical Dictionary for Regulatory Activities (version 7.0) and tabulated by severity and their relationship to the study drug for each treatment group. Descriptive statistics for baseline, analysis time point, and change from baseline in vital signs were summarized, and changes considered to be of potential clinical importance were flagged. Results of the 12-lead ECG performed at screening, Visit 6, and study endpoint were summarized as normal, abnormal (not clinically significant), or abnormal (clinically significant), and a shift table comparing the baseline results to those from Week 4 and the study endpoint was constructed. Screening and final visit findings on the physical examination were recorded for each subject. Descriptive statistics for hematology, serum chemistry, and urinary laboratory values and changes from baseline were displayed by analysis time point, and values were evaluated to identify trends within treatment groups. Inferential testing was not performed on any safety measure.


This study was conducted at 12 sites in the United States between January 18 and October 4, 2005.
Subject disposition. Among the 111 subjects randomized to study medication (safety population), 47 (42.3%) completed the study and 64 (57.7%) withdrew early (Table 1). The most common reasons (reported by >10% of subjects) for early termination were: lost to follow-up (20.7%), AEs (14.4%), and consent withdrawal (13.5%). A higher percentage of subjects in the BID (24.5%) vs. the QHS group (17.2%) were lost to follow-up. A slightly higher percentage of subjects in the QHS versus BID group withdrew early due to AEs (17.2% vs. 11.3%) and consent withdrawal (17.2% vs. 9.4%). The ITT population for efficacy analysis comprised 105 subjects.

Subject demographics. Baseline demographic and disease severity characteristics are summarized in Table 2.
Baseline demographic characteristics were similar between treatment groups. Most subjects were female (62.2%) and white (76.6%); the mean age was 37 years. Baseline disease severity was similar between treatment groups (e.g., YMRS total score [standard deviation (SD)] in the BID group was 22.8 [7.32] and in the QHS group was 20.6 [6.61]); however, the percentage of subjects taking concomitant medications was higher in the BID group (83.0%) compared with the QHS group (69.0%). While the percentage of subjects on concomitant lithium was similar between groups (BID, 5.7%; QHS, 6.9%), more subjects in the QHS group (10.3%) compared with the BID group (3.8%) received quetiapine (Table 3).

Efficacy assessments. YMRS. There was a significant mean reduction (improvement) from baseline in the YMRS total score with both BID and QHS dosing at all time points in the LOCF analysis for subjects in the ITT population. The least squares (LS) mean decrease from baseline was -5.3 at Week 1 and -11.4 at Week 12 in the BID group, and -5.0 at Week 1 and -10.1 at Week 12 in the QHS group. There were no statistically significant differences between treatment groups at any time point (Figure 1).

HAM-D. Similarly, there was a significant mean reduction (improvement) from baseline at all time points in the first 17 items on the HAM-D scale, a secondary efficacy variable, for subjects in the ITT population in both treatment groups. The LS mean decrease from baseline was -2.9 at Week 1 and -4.2 at Week 12 in the BID group, and
-2.2 at Week 1 and -3.0 at Week 12 in the QHS group. The treatment effect was not statistically significant at any time point (Figure 2).

MADRS. There was also a significant mean reduction (improvement) from baseline in the MADRS total score at all time points in both treatment groups. The LS mean decrease from baseline was -3.7 at Week 1 and -9.3 at Week 12 in the BID group, and -4.3 at Week 1 and -8.6 at Week 12 in the QHS group. Similar to the other efficacy outcome variables, the treatment effect was not statistically significant at any time point (Figure 3).

CGI-BP. For each of the three components of the CGI-BP-C (mania, depression, and overall bipolar disorder), the percentage of subjects rated as “very much improved,” “much improved,” and “minimally improved” increased, while the percentage rated as “very much worse,” “much worse,” or “minimally worse” decreased in both treatment groups, indicating improvement in severity of illness. For the LOCF analysis, the CGI-BP-C scores at each postbaseline visit were reduced at Week 12 compared with Week 1 in both treatment groups (Table 4). Figure 4 illustrates improvement in the overall bipolar scores at each postbaseline visit.

Time to initial remission. Initial remission (YMRS ?12) was achieved by 84.6 percent of subjects in the BID group and in 81.1 percent of those in the QHS group, a nonsignificant difference. The median number of days to remission was 16 in the BID group compared with 15 in the QHS group; however, the difference between treatment groups was not statistically significant.

Safety results. Study drug exposure. The mean daily doses of CBZ-ERC were 690.54mg overall, with 727.33mg in the BID group and 656.38mg in the QHS group. Most subjects in the BID group (67.9%) and the QHS group (62.1%) were exposed to study drug for >4 weeks, and 30.2 percent in the BID group and 19.0 percent in the QHS group were exposed for >12 weeks. Overall exposure to CBZ-ERC was slightly greater in the BID group (8.15 patient-years) compared with the QHS group (7.41 patient-years), and mean CBZ-ERC exposure was slightly longer in the BID group (8.18 weeks) compared with the QHS group (6.91 weeks).

Adverse events. Slightly more subjects in the BID group (90.6%) compared with the QHS group (86.2%) reported TEAEs (Table 5). Most TEAEs in the BID and QHS groups were mild or moderate in severity, with 70 percent of events in both groups graded as mild in intensity. Ten severe AEs were reported in the BID group, and 14 were reported in the QHS group. In addition, several TEAEs occurred among fewer subjects in the QHS group compared with the BID group. Notable differences occurred in the incidence of nausea (QHS 24.1%; BID 30.2%), fatigue (QHS 6.9%; BID 17.0%), increased appetite (QHS 3.4%; BID 11.3%), dizziness (QHS 17.2%; BID 24.5%), and somnolence (QHS 8.6%; BID 15.1%).

Treatment-related AEs were reported in 81.1 percent of subjects in the BID group vs. in 84.5 percent of subjects in the QHS group; most events were considered mild in intensity. The most common treatment-related AEs in the BID group were nausea and dizziness (24.5% each), fatigue and headache (17% each), sedation and somnolence (15.1% each), and dry mouth and increased appetite (11.3% each); headache (1.9%) was the only severe treatment-related AE in the BID group. The most common treatment-related AEs in the QHS group were nausea (22.4%), headache (20.7%), dry mouth (17.2%), dizziness (15.5%), sedation and vomiting (12.1% each), and blurred vision (10.3%). The incidence of rash in both groups was low (<10%). Severe treatment-related AEs in the QHS group included fatigue, hypersensitivity, weight increase, increased appetite, dizziness, headache, syncope, and night sweats (1.7% each). The single death that occurred during the study (BID group) was due to a fatal overdose of illicit methadone and was considered accidental and unrelated to the study drug by the medical examiner.

Overall, 14.4 percent of subjects discontinued CBZ-ERC because of AEs or serious AEs: six (11.3%) subjects in the BID group and 10 (17.2%) subjects in the QHS group (p=NS). AEs that were considered study drug-related and led to CBZ-ERC discontinuation were reported for three subjects in the BID group (1 with pruritus and rash; 1 with increased GGT; and 1 with headache and rash), and eight subjects in the QHS group (1 with worsening of bipolar disorder; 1 with hypersensitivity; 1 with cognitive disorder; 1 with vertigo and pruritus; 2 with drug hypersensitivity; 1 with pruritus; and 1 with blurred vision and headache).

Other safety results. For the majority of subjects in both groups, hematology parameters were within normal reference range at Week 12, and there were no notable differences observed between groups from normal at screening to abnormal at Week 12. In addition, except for greater mean changes in platelet counts in the QHS group
(-28.3×103/µL) compared with the BID group (-7.2×103/µL), and greater mean changes in the neutrophil counts in the BID group (-0.627×103/µL) compared with the QHS group (-0.266×103/µL), overall, similar trends (slight decreases) were observed in mean hematology changes between treatment groups. No subjects discontinued treatment because of hematologic abnormalities during the study.

In general, similar trends in mean serum chemistry and electrolyte parameter changes were also observed between treatment groups, and no clinically significant laboratory abnormalities were considered to be serious AEs. Slightly more subjects in the QHS group (8) compared with the BID group (3) had abnormally high glucose results at Week 12. In addition, the mean baseline glucose (94.9mg/dL [SD, 29.59]) was also slightly higher in the QHS versus the BID group (89.0mg/dL [SD, 11.21]), and the mean change from screening was much higher in the QHS (14.6mg/dL [SD, 40.78]) versus the BID group (1.8mg/dL [SD, 20.18]). In contrast, mean cholesterol increases were higher in the BID group (18.1mg/dL) compared with the QHS group (12.5mg/dL). While elevated liver function tests were reported in both groups (normal to high ALP [BID 7.5%; QHS 6.9%]; normal to high GGT [BID 26.4%; QHS 20.7%]; and normal to low bilirubin [BID 13.2%; QHS 12.1%]), most patients with an elevated GGT after the first dose of study drug also had an elevated GGT at screening.

Mean changes in urinalysis parameters from screening to endpoint were negligible in both the BID group (<0.1) and the QHS group (-0.2). There were few changes in physical examination results from screening to endpoint; however, rash was reported for three subjects in each treatment group and urticaria was reported in one subject in the BID group. At Week 12, mean changes from baseline in vital sign parameters, including weight gain—QHS group, 1.1 pounds and BID group, 1.7 pounds—were clinically insignificant and similar between groups. No trends were observed between groups in overall qualitative ECG interpretation, and no subjects with normal ECG results at screening had clinically significant abnormal ECG results at Week 12.


In this study, CBZ-ERC efficacy in the treatment of bipolar disorder—evaluated using standard measures used in clinical studies in bipolar disorder[28–31]—did not differ significantly between the QHS and BID dosing regimens. Both QHS and BID dosing were significantly effective in reducing manic symptoms (as evidenced by decreases in the YMRS total score from baseline) as well as depressive symptoms (as evidenced by decreases in the HAM-D21 and MADRS total scores) with no statistically significant differences observed between treatment groups at any time point with any measure. All three CGI-BP components also improved from baseline, indicating improvements in severity of mania, depression, and overall illness. Remission was achieved by a final YMRS score ?12 in >80 percent of subjects in both groups. In addition, CBZ-ERC was generally well tolerated, with most AEs mild or moderate in severity in both groups. There were no significant changes from baseline in hematology parameters for the majority of patients in either dosing group, the trends in mean serum chemistry and electrolyte parameters were generally similar between groups, and there were no notable changes in vital signs, ECG parameters, or weight.

Although there was a nonsignificant trend favoring BID over QHS dosing—with a slightly greater reduction in mean total scores in the YMRS, HAM-D21, and MADRS, an improvement in depression and overall CGI-BP-C scores at Week 12, and a slightly higher percentage achieving total remission in the BID group—the two dosing regimens had a similar treatment effect at all time points during the study. A more consistent cbz serum level may have been achieved with BID compared with QHS dosing and higher daily dosing was also achieved in the BID group, perhaps contributing to this difference. In addition, although a slightly lower incidence of several common TEAEs was reported in the QHS group compared with the BID group (i.e., nausea, fatigue, increased appetite, dizziness, and somnolence), the incidence of severe AEs and discontinuation because of AEs was higher with QHS compared with BID dosing, which may again be partially accounted for by the fluctuating serum levels and/or higher peak levels present with once-daily dosing.

There are certain safety concerns with anticonvulsants, including cbz, such as the potential to cause increased cholesterol and weight gain and to precipitate rash, including severe and life-threatening rash syndromes (i.e., Stevens-Johnson syndrome, toxic epidermal necrolysis).[32] While a low incidence of rash and no serious rashes were reported in this study, hypersensitivity, pruritus, and rash were reported as a reason for discontinuation by five subjects in the QHS group compared with two subjects in the BID group; no subject was receiving concomitant antipsychotics or other medications frequently associated with a rash.

These reasons for study discontinuation may have been due to chance. They may also have been related to dosing; there were probably greater peak-trough changes in the QHS group due to the single dose administration. It is also possible that the differences are an artifact, since seven percent more subjects in the BID group were lost to follow-up; perhaps these subjects were experiencing similar AEs for which they would have eventually discontinued the study, thus reducing the observed differences between the treatment groups. No subject discontinued the study due to suicidality or suicide.

Overweight and obesity are commonly encountered in patients with bipolar disorder,33 making the potential for weight gain a concern in the choice of treatment.[34] Consistent with results of other studies with CBZ-ERC,[23–25] clinically significant weight gain was not reported in this study. Along with weight gain, increases in lipids and glucose levels can contribute to the development of the metabolic syndrome,[35] which is now recognized to occur at a rate at least as high,[36] if not higher,[35] in patients with bipolar disorder compared with the overall population. A normal-to-high shift in values from screening to endpoint in total cholesterol levels was reported in 11.3 percent of the BID group and 13.8 percent of the QHS group, and a normal-to-high shift in glucose levels was reported in 5.7 percent of the BID group and 13.8 percent of the QHS group, although fasting was not required for this study, and no other lipids were measured. Many of the subjects were also on atypical neuroleptics that may also have impacted lipids and glucose readings. The long-term clinical significance of these findings and what would happen to these values over time are both presently unknown. In addition to total cholesterol, measurement of other lipid parameters, such as low-density lipoproteins, high-density lipoproteins, and triglycerides, would allow a more accurate assessment of the metabolic risk associated with CBZ-ERC.

Efficacy results in the present study corroborate those from two previously published short-term studies of hospitalized patients with bipolar mania,[23,24] with CBZ-ERC treatment leading to significant improvements in symptoms of bipolar disorder in adults, and the percentage of YMRS responders (84.6% of subjects in the BID group and 81.1% of subjects in the QHS group) comparable to that of a six-month, open-label extension study comprising subjects from two short-term studies (78.4% of prior CBZ-ERC- and 72.5% of prior placebo-treated subjects).[25] Further, although the present study demonstrated equivalent efficacy of two different dosing regimens, the overall mean daily dose (690.54mg) was similar to that of the previously published short[-term studies (756.44mg[23] and 642.6mg[24]), and lower than that in the long-term study (938mg).25 It should be noted, however, that the inclusion criteria in this study dictated a lower baseline total YMRS score (?16) compared with that in previously published short-term studies (?20)[37] and some concomitant medications were permitted. The subjects also had a lower baseline mean YMRS score (27.6)[37] in those studies compared with that in the present study (21.6), suggesting the subjects were less symptomatic at baseline in the present study.[37]

While the overall incidence of TEAEs in the present study (BID 90.6%; QHS 86.2%) was similar to previous studies (88.1%;23 91.8%[24]), the incidence of somnolence (BID 15.1%; QHS 8.6%) and dizziness (BID 24.5%; QHS 17.2%) in each treatment group was substantially lower compared with previous studies (somnolence 32.7%23 and 27.9%;[24] dizziness 48.5%23 and 41.8%[24]), which used a much more aggressive dose-titration schedule. These differences may be attributed to the lower starting dose and slower CBZ-ERC titration schedule used in the present study, chosen in part to explore such a titration schedule in less ill outpatients. While elevated serum glucose levels in the present study were not reported in previous CBZ-ERC studies, several factors could have confounded these parameters in the present study. Unlike the present study, previously published short-term studies included a washout period, and no concomitant mood stabilizers or atypical antipsychotic agents[23,24]—which could have altered the serum glucose levels—were permitted. Further, patient fasting was not required for clinical chemistry samples, limiting the utility of total cholesterol and glucose values in the present study.

Poor medication adherence is a common problem in bipolar disorder, with approximately 30 percent of patients with bipolar disorder in a recent study reporting incomplete adherence to their drug treatment.[38] However, it has long been recognized that adherence improves with less frequent dosing schedules.[39,40] Because these studies suggest similar efficacy and safety of BID dosing compared with QHS dosing of CBZ-ERC, one practical interpretation of this study (not included in the package labeling) is that patients can be started on a BID regimen of CBZ-ERC, and then confidently be switched to QHS dosing once stabilized. Patients with bipolar disorder require chronic medication and are often treated with combinations of medications, which can negatively affect compliance; it is possible that a once-daily dosing regimen may improve outcomes in some patients with bipolar disorder who are poorly compliant.

This study has several limitations. First, although the sample size was sufficient to evaluate efficacy and commonly reported AEs associated with CBZ-ERC, it was insufficient to detect more rare, and potentially serious, AEs. Second, the study design did not include a placebo control, allowing subjects and clinicians to be aware of the medication being administered, and the comparators for both safety and efficacy were administered at similar doses (BID mean daily dose, 727.33mg; QHS mean daily dose, 656.38mg) on different dosing regimens. Further, the use of concomitant mood stabilizers, atypical antipsychotics, and benzodiazepines could potentially confound the safety and efficacy results. Finally, as mentioned above, the safety results (e.g., glucose and cholesterol values) were also potentially compromised by not requiring fasting sample collections.


Up to 12 weeks of CBZ-ERC in doses of up to 1600mg/d effectively improved manic symptoms and was generally safe and well tolerated in adults with bipolar I disorder, with no significant differences observed between QHS and BID dosing. Tolerability with once-daily dosing may be improved compared with BID dosing for some patients, as indicated by a lower incidence of several commonly reported TEAEs in the QHS group like somnolence, dizziness, and increased appetite. However, the increased number of early discontinuations in the QHS group due to hypersensitivity, pruritus, and rash requires further study to rule out explanations other than a chance association. Because compliance has been shown to improve with decreased dosing frequency, additional studies specifically evaluating compliance with QHS versus BID dosing regimens of CBZ-ERC as well as other dosing strategies (e.g., starting BID and switching to QHS), in adult patients with bipolar disorder are needed.


1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Text Revision: DSM-IV-TR. Washington, DC: American Psychiatric Association, 2000.
2. Grant BF, Stinson FS, Hasin DS, et al. Prevalence, correlates, and comorbidity of bipolar I disorder and axis I and II disorders: Results from the National Epidemiologic Survey on Alcohol and Related Conditions. J Clin Psychiatry 2005;66:1205–15.
3. Angst J, Sellaro R. Historical perspectives and natural history of bipolar disorder. Biol Psychiatry 2000;48:445–57.
4. Kessler RC, Rubinow DR, Holmes C, et al. The epidemiology of DSM-III-R bipolar I disorder in a general population survey. Psychol Med 1997;27:1079–89.
5. Suppes T, Leverich GS, Keck PE, et al. The Stanley Foundation Bipolar Treatment Outcome Network. II. Demographics and illness characteristics of the first 261 patients. J Affect Disord 2001;67:45–59.
6. Valtonen H, Suominen K, Mantere O, et al. Suicidal ideation and attempts in bipolar I and II disorders. J Clin Psychiatry 2005;66:1456–62.
7. Dilsaver SC, Chen YW, Swann AC, et al. Suicidality, panic disorder and psychosis in bipolar depression, depressive-mania and pure-mania. Psychiatry Res 1997;73:47–56.
8. Goodwin FK, Jamison KR. Suicide. Manic-Depressive Illness: Bipolar Disorders and Recurrent Depression. New York, NY: Oxford University Press, 2007:247–69.
9. Harris EC, Barraclough B. Suicide as an outcome for mental disorders. A meta-analysis. Br J Psychiatry 1997;170:205–28.
10. Baldessarini RJ, Pompili M, Tondo L. Suicide in bipolar disorder: Risks and management. CNS Spectr 2006;11:465–71.
11. Inskip HM, Harris EC, Barraclough B. Lifetime risk of suicide for affective disorder, alcoholism and schizophrenia. Br J Psychiatry 1998;172:35–7.
12. Murray CJ, Lopez AD. Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet 1997;349:1436–42.
13. US Food and Drug Administration, Center For Drug Evaluation and Research. Information for healthcare professionals: Suicidality and antiepileptic drugs. Website. Available at: Access date: 2/11/08.
14. Kessing LV, Sondergard L, Kvist K, et al. Suicide risk in patients treated with lithium. Arch Gen Psychiatry 2005;62:860–6.
15. Sondergard L, Lopez AG, Andersen PK, et al. Mood-stabilizing pharmacological treatment in bipolar disorders and risk of suicide. Bipolar Disord 2008;10:87–94.
16. Mitchell PB, Malhi GS. The expanding pharmacopoeia for bipolar disorder. Ann Rev Med 2002;53:173–88.
17. Weisler RH, Cutler AJ, Ballenger JC, et al. The use of antiepileptic drugs in bipolar disorders: A review based on evidence from controlled trials. CNS Spectr 2006;11:788–99.
18. El Mallakh R, Weisler RH, Townsend MH, et al. Bipolar II disorder: current and future treatment options. Ann Clin Psychiatry 2006;18:259–66.
19. Baldessarini RJ, Leahy L, Arcona S, et al. Patterns of psychotropic drug prescription for U.S. patients with diagnoses of bipolar disorders. Psychiatr Serv 2007;58:85–91.
20. Ballenger JC, Post RM. Carbamazepine in manic-depressive illness: A new treatment. Am J Psychiatry 1980;137:782–90.
21. Post RM, Uhde TW, Roy-Byrne PP, et al. Correlates of antimanic response to carbamazepine. Psychiatry Res 1987;21:71–83.
22. Emilien G, Maloteaux JM, Seghers A, et al. Lithium compared to valproic acid and carbamazepine in the treatment of mania: A statistical meta-analysis. Eur Neuropsychopharmacol 1996;6:245–52.
23. Weisler RH, Kalali AH, Ketter TA. A multicenter, randomized, double-blind, placebo-controlled trial of extended-release carbamazepine capsules as monotherapy for bipolar disorder patients with manic or mixed episodes. J Clin Psychiatry 2004;65:478–84.
24. Weisler RH, Keck PE, Jr., Swann AC, et al. Extended-release carbamazepine capsules as monotherapy for acute mania in bipolar disorder: A multicenter, randomized, double-blind, placebo-controlled trial. J Clin Psychiatry 2005;66:323–30.
25. Ketter TA, Kalali AH, Weisler RH. A 6-month, multicenter, open-label evaluation of beaded, extended-release carbamazepine capsule monotherapy in bipolar disorder patients with manic or mixed episodes. J Clin Psychiatry 2004;65:668–73.
26. Weisler RH. Carbamazepine extended-release capsules for the treatment of bipolar I disorder. Therapy 2005;2:533–44.
27. Ginsberg LD. Safety and efficacy of extended-release carbamazepine capsules in patients with bipolar disorder: QD vs. BID dosing. Presented at US Psychiatric and Mental Health Congress; November 18-21, 2004; San Diego, CA.
28. Young RC, Biggs JT, Ziegler VE, et al. A rating scale for mania: reliability, validity and sensitivity. Br J Psychiatry 1978;133:429–35.
29. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23:56–62.
30. Montgomery SA, Asberg M. A new depression scale designed to be sensitive to change. Br J Psychiatry 1979;134:382–9.
31. Spearing MK, Post RM, Leverich GS, et al. Modification of the Clinical Global Impressions (CGI) Scale for use in bipolar illness (BP): The CGI-BP. Psychiatry Res 1997;73:159–71.
32. Post RM. Extended-release carbamazepine: targeting acute bipolar mania. Curr Psychiatry 2005;4:72–9.
33. McElroy SL, Frye MA, Suppes T, et al. Correlates of overweight and obesity in 644 patients with bipolar disorder. J Clin Psychiatry 2002;63:207–13.
34. Malhotra S, McElroy SL. Medical management of obesity associated with mental disorders. J Clin Psychiatry 2002;63(Suppl 4):24–32.
35. Basu R, Brar JS, Chengappa KN, et al. The prevalence of the metabolic syndrome in patients with schizoaffective disorder–bipolar subtype. Bipolar Disord 2004;6:314–8.
36. Fagiolini A, Frank E, Scott JA, et al. Metabolic syndrome in bipolar disorder: Findings from the Bipolar Disorder Center for Pennsylvanians. Bipolar Disord 2005;7:424–30.
37. Weisler RH, Hirschfeld R, Cutler AJ, et al. Extended-release carbamazepine capsules as monotherapy in bipolar disorder : pooled results from two randomised, double-blind, placebo-controlled trials. CNS Drugs 2006;20:219–31.
38. Scott J, Pope M. Self-reported adherence to treatment with mood stabilizers, plasma levels, and psychiatric hospitalization. Am J Psychiatry 2002;159:1927–9.
39. Cramer JA, Mattson RH, Prevey ML, et al. How often is medication taken as prescribed? A novel assessment technique. JAMA 1989;261:3273–7.
40. Greenberg RN. Overview of patient compliance with medication dosing: A literature review. Clin Ther 1984;6:592–9.