Research report Randomized, placebo-controlled trial of quetiapine XR and divalproex ER monotherapies in the treatment of the anxious bipolar patient $ David V. Sheehan a,n , Kathy Harnett-Sheehan a , Rosario B. Hidalgo a , Juris Janavs a , Susan L. McElroy b , Darlene Amado a , Trisha Suppes c a Clinical and Translational Science Institute, University of South Florida, College of Medicine, 3515 East Fletcher Ave., Tampa, FL 33613, USA b Department of Psychiatry, University of Cincinnati College of Medicine, University of Cincinnatti, Lindner Center of Hope, 4075 Old Western Row Rd., Mason, OH 45040, USA c Department of Psychiatry & Behavioral Science, Stanford University School of Medicine, VA Palo Alto Health Care System, 3801 Miranda Ave. (151T), Palo Alto, CA 94304, USA article info Article history: Received 18 April 2012 Received in revised form 8 July 2012 Accepted 17 July 2012 Available online 21 August 2012 Keywords: Bipolar disorder Anxiety Panic disorder Placebo Atypical antipsychotic Generalized anxiety disorder abstract Background: Anxiety disorders complicate the treatment of bipolar disorder but are seldom the focus of bipolar treatment studies. Methods: The anxiolytic effect of quetiapine XR 50–300 mg/day compared to divalproex ER (500– 3000 mg/day) was tested in an 8-week, double-blind, placebo-controlled, randomized clinical trial in 149 patients with bipolar disorder and a co-occurring panic disorder or GAD. The primary efficacy measure was the Clinician Global Improvement-21 Anxiety Scale (CGI-21). Secondary measures included the Hamilton Anxiety Scale (HAM-A) and Sheehan Panic Disorder Scale (SPS). Results: Repeated measures last-observation-carried-forward (LOCF) analyses of variance demon- strated significant treatment-by-time interaction effects on 3 of the 4 anxiety measures. Quetiapine XR at a mean endpoint dose of 186 mg/day produced rapid sustained improvements relative to baseline, divalproex ER and placebo on anxiety. Mean baseline-to-endpoint improvement was significantly greater for quetiapine XR compared to divalproex ER and placebo on the HAM-A and SPS. Both active medications were well tolerated, but weight gain was higher on quetiapine XR. Limitations: The study was limited to 8 weeks and to patients with bipolar disorder and comorbid panic disorder or GAD. The results may not be applicable to quetiapine XR as an add-on treatment to mood stabilizers or to bipolar disorder comorbid with other anxiety disorders. Conclusions: Quetiapine XR in a dose range of 50–300 mg/day appears to reduce anxiety in bipolar patients with comorbid panic disorder or GAD treated for 8 weeks. The efficacy of other second- generation antipsychotics and mood stabilizers in patients with bipolar disorder and a co-occurring anxiety disorder should be investigated in double-blind, placebo-controlled studies. & 2012 Elsevier B.V. All rights reserved. 1. Introduction Anxiety disorders and anxious symptoms are common in bipolar disorder (Pini et al., 1997; Kessler et al., 1997; Feske, et al., 2000; Freeman et al., 2002; Boylan, et al., 2004). Indeed, in one study, Boylan et al. (2004) found that 55.8% of bipolar outpatients had at least one comorbid anxiety disorder and within this population 57% had more than two anxiety disorders. This comorbidity poses challenges for the treating physician. Compared to patients without an anxiety disorder, anxious bipolar patients have an earlier age of onset of illness (Schurhoff et al., 2000; McElroy et al., 2001; Carter et al., 2003; Henry et al., 2003; Perlis et al., 2004), higher rates of mixed states (Boylan et al., 2004), more depressive episodes, alcohol abuse and suicidal ideation (Young et al., 1993; Frank et al., 2002; Carter et al., 2003; Simon et al., 2004; Frye et al., 2003; Perlis et al., 2004) and a poorer response to lithium-based treatment (Freeman et al., 2002; McElroy et al., 2001; Frank et al., 2002). These observations, together with growing evidence that bipolar disorder with panic disorder may be a genetic subtype within the bipolar disorder spectrum (MacKinnon et al., 1997, 2003; Rotondo et al., 2002; Dilsaver et al., 2006; Nardi et al., 2007), have led to increasing interest in finding pharmacologic treatments that specifically target anxiety in bipolar disorder. Unfortunately, first line treatments for anxiety disorders (selective serotonin reuptake inhibitors [SSRIs] and selective- norepinephrine reuptake serotonin reuptake inhibitors [SNRIS]) are not recommended, at least as monotherapies, in bipolar patients with anxiety disorders because of their potential to induce rapid cycling and manic, hypomanic or depressive episodes (Ghaemi et al., 2003; Salvi et al., 2008; Young and Seim, 2009). Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/jad Journal of Affective Disorders 0165-0327/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jad.2012.07.016 $ This study was supported by a grant from Astra Zeneca. n Corresponding author. Tel.: þ1 813 956 8437; fax: þ1 813 974 4575. E-mail address: [email protected] (D.V. Sheehan). Journal of Affective Disorders 145 (2013) 83–94
Transcript
Journal of Affective Disorders 145 (2013) 83–94
Contents lists available at SciVerse ScienceDirect
Journal of Affective Disorders
0165-03
http://d
$Thisn Corr
E-m
journal homepage: www.elsevier.com/locate/jad
Research report
Randomized, placebo-controlled trial of quetiapine XR and divalproexER monotherapies in the treatment of the anxious bipolar patient$
David V. Sheehan a,n, Kathy Harnett-Sheehan a, Rosario B. Hidalgo a, Juris Janavs a, SusanL. McElroy b, Darlene Amado a, Trisha Suppes c
a Clinical and Translational Science Institute, University of South Florida, College of Medicine, 3515 East Fletcher Ave., Tampa, FL 33613, USAb Department of Psychiatry, University of Cincinnati College of Medicine, University of Cincinnatti, Lindner Center of Hope, 4075 Old Western Row Rd.,
Mason, OH 45040, USAc Department of Psychiatry & Behavioral Science, Stanford University School of Medicine, VA Palo Alto Health Care System, 3801 Miranda Ave. (151T), Palo
Alto, CA 94304, USA
a r t i c l e i n f o
Article history:
Received 18 April 2012
Received in revised form
8 July 2012
Accepted 17 July 2012Available online 21 August 2012
Background: Anxiety disorders complicate the treatment of bipolar disorder but are seldom the focus of
bipolar treatment studies.
Methods: The anxiolytic effect of quetiapine XR 50–300 mg/day compared to divalproex ER (500–
3000 mg/day) was tested in an 8-week, double-blind, placebo-controlled, randomized clinical trial in
149 patients with bipolar disorder and a co-occurring panic disorder or GAD. The primary efficacy
measure was the Clinician Global Improvement-21 Anxiety Scale (CGI-21). Secondary measures
included the Hamilton Anxiety Scale (HAM-A) and Sheehan Panic Disorder Scale (SPS).
Results: Repeated measures last-observation-carried-forward (LOCF) analyses of variance demon-
strated significant treatment-by-time interaction effects on 3 of the 4 anxiety measures. Quetiapine
XR at a mean endpoint dose of 186 mg/day produced rapid sustained improvements relative to
baseline, divalproex ER and placebo on anxiety. Mean baseline-to-endpoint improvement was
significantly greater for quetiapine XR compared to divalproex ER and placebo on the HAM-A and
SPS. Both active medications were well tolerated, but weight gain was higher on quetiapine XR.
Limitations: The study was limited to 8 weeks and to patients with bipolar disorder and comorbid panic
disorder or GAD. The results may not be applicable to quetiapine XR as an add-on treatment to mood
stabilizers or to bipolar disorder comorbid with other anxiety disorders.
Conclusions: Quetiapine XR in a dose range of 50–300 mg/day appears to reduce anxiety in bipolar
patients with comorbid panic disorder or GAD treated for 8 weeks. The efficacy of other second-
generation antipsychotics and mood stabilizers in patients with bipolar disorder and a co-occurring
anxiety disorder should be investigated in double-blind, placebo-controlled studies.
& 2012 Elsevier B.V. All rights reserved.
1. Introduction
Anxiety disorders and anxious symptoms are common in bipolardisorder (Pini et al., 1997; Kessler et al., 1997; Feske, et al., 2000;Freeman et al., 2002; Boylan, et al., 2004). Indeed, in one study,Boylan et al. (2004) found that 55.8% of bipolar outpatients had atleast one comorbid anxiety disorder and within this population 57%had more than two anxiety disorders. This comorbidity poseschallenges for the treating physician. Compared to patients withoutan anxiety disorder, anxious bipolar patients have an earlier age ofonset of illness (Schurhoff et al., 2000; McElroy et al., 2001; Carteret al., 2003; Henry et al., 2003; Perlis et al., 2004), higher rates ofmixed states (Boylan et al., 2004), more depressive episodes, alcohol
ll rights reserved.
Zeneca.
: þ1 813 974 4575.
eehan).
abuse and suicidal ideation (Young et al., 1993; Frank et al., 2002;Carter et al., 2003; Simon et al., 2004; Frye et al., 2003; Perlis et al.,2004) and a poorer response to lithium-based treatment (Freemanet al., 2002; McElroy et al., 2001; Frank et al., 2002). Theseobservations, together with growing evidence that bipolar disorderwith panic disorder may be a genetic subtype within the bipolardisorder spectrum (MacKinnon et al., 1997, 2003; Rotondo et al.,2002; Dilsaver et al., 2006; Nardi et al., 2007), have led to increasinginterest in finding pharmacologic treatments that specifically targetanxiety in bipolar disorder.
Unfortunately, first line treatments for anxiety disorders(selective serotonin reuptake inhibitors [SSRIs] and selective-norepinephrine reuptake serotonin reuptake inhibitors [SNRIS])are not recommended, at least as monotherapies, in bipolarpatients with anxiety disorders because of their potential to inducerapid cycling and manic, hypomanic or depressive episodes(Ghaemi et al., 2003; Salvi et al., 2008; Young and Seim, 2009).
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–9484
Moreover, the presence of anxiety is associated with a poorerresponse to lithium (Boylan et al., 2004, Henry et al., 2003).
Based on its demonstrated efficacy in panic disorder (Lumet al., 1990; Keck et al., 1993), the mood stabilizer divalproex hasbeen postulated to have anxiolytic effects in bipolar disorder. Todate, however, no double-blind, placebo-controlled studies haveexamined its potential effects in reducing anxiety in bipolarpatients with a co-occurring anxiety disorder and only one smalldouble-blind placebo-controlled study has been publisheddemonstrating its superiority to placebo in reducing anxioussymptoms in a bipolar population (Davis et al., 2005).
Several second generation antipsychotics have been shown tobe effective and well tolerated in the treatment of manic andmixed episodes of bipolar I disorder (Derry and Moore, 2007;Perlis, 2005) and are increasingly being used as mood stabilizersin bipolar disorders (Kessler et al., 2005; Ghaemi et al., 2006).Their potential anxiolytic efficacy in bipolar disorder is thereforereceiving growing attention.
Early reports indicated that adding quetiapine, risperidone, orolanzapine to an SSRI produced results superior to placebo in thetreatment of anxiety disorders such as refractory obsessivecompulsive disorder (OCD), generalized anxiety disorder (GAD)and post-traumatic stress disorder (PTSD) in patients withoutbipolar disorder (McIntyre and Katzman, 2003). However, thesample sizes of these studies were small (Gao et al., 2006). Morerecently, 2 large double-blind, placebo controlled studies, havedemonstrated the anxiolytic efficacy of an extended-releaseformulation of quetiapine fumarate (quetiapine XR) in reducinganxiety in patients with a diagnosis of generalized anxiety disorder(GAD) (Khan et al., 2008; Bandelow et al., 2010) without bipolardisorder.
In addition, 3 large double-blind placebo-controlled studies haveinvestigated the effects of atypical antipsychotics on anxious symp-toms in patients with bipolar depression using the HAM-A as asecondary measure. In one study, Tohen et al. (2007) found thatolanzapine alone and together with fluoxetine were superior toplacebo in reducing HAM-A scores after 8 weeks of treatment. In apooled analysis of data from two other studies, Hirschfeld et al.(2006) found that immediate release quetiapine at doses of 300 mg/day and 600 mg/day significantly reduced total HAM-A scorescompared to placebo in 351 patients after 8 weeks of treatment.These results held up in a more recent secondary analysis of thisdata using a larger sample (n¼978) (Lydiard et al., 2009).
Unfortunately none of the above three studies specifiedwhether or not the subjects had co-occurring syndromal anxietydisorders. Such specification may matter because, while secondgeneration antipsychotics may reduce anxiety symptoms inbipolar depression, these medications have been identified asexacerbating symptoms in panic disorder and OCD, possiblybecause of their serotonergic antagonistic properties (Bakeret al., 1992, de Haan et al., 2002).
To date there is only one published double-blind, placebo-controlled trial investigating the anxiolytic efficacy of a secondgeneration antipsychotic on anxiety as the primary target oftreatment in patients with bipolar disorder and a co-occurringsyndromal anxiety disorder. That study (Sheehan et al., 2009) didnot find risperidone monotherapy (0.5–4 mg/day) to be any moreeffective than placebo in reducing anxiety or other symptoms in111 patients with bipolar disorder (I, II or NOS) and a co-occurringpanic disorder or generalized anxiety disorder (GAD). The results,however, may have been unique to risperidone and not apply toother second generation antipsychotics.
Current treatment guidelines for bipolar disorder recommendtreating anxiety disorders concurrently with bipolar disorder(Suppes et al., 2005; Perlis, 2005), but evidence based optionsfor bipolar patients with a co-occurring anxiety disorder are still
limited and treatment in this population is all too often based onanecdotal reports and open clinical experience (Perugi and Toni,2004; Singh and Zarate, 2006). Given preliminary evidence thatthe mood stabilizer divalproex and at least 2 of the secondgeneration antipsychotics (quetiapine and olanzapine) are effec-tive in reducing anxiety symptoms in patients with bipolardisorder, an important question is how effective these types oftreatments are, compared to each other and to placebo, in bipolarpatients with a co-occurring syndromal anxiety disorder.
2. Method
2.1. Study design
This randomized, double-blind, parallel-group, multicenterstudy compared the anxiolytic efficacy of the second generationantipsychotic quetiapine XR as monotherapy with the moodstabilizer divalproex ER as monotherapy and placebo in adultoutpatients with a lifetime bipolar I, II or NOS disorder, a lifetimepanic or generalized anxiety disorder and at least moderatelysevere anxiety symptoms. The study was conducted between Julyof 2007 and April of 2010 at three sites in the United States(University of South Florida, University of Cincinnati, University ofTexas Southwestern Medical Center). The institutional reviewboard for each site approved the protocol and written informedconsent was received from each participant before any study-related procedures were performed. Following a 2-day to 4-weekscreening, patients were randomized in a 1:1:1 ratio to receivequetiapine XR, divalproex ER or placebo in a flexible dose regimenof 50–300 mg/day quetiapine XR or 500–3000 mg/day divalproexER for 8 weeks.
2.2. Patients
Eligible patients included men and women aged 18–65 yearsof age who met Diagnostic and Statistical Manual, 4th edition(DSM-IV) criteria for a lifetime bipolar I, II, or NOS disorder and alifetime panic disorder (PD) or generalized anxiety disorder(GAD). Diagnoses had to be documented on the Mini InternationalNeuropsychiatric Interview (MINI; Sheehan et al., 1998). How-ever, for the purpose of the study, the GAD Criterion F clause,‘‘does not occur exclusively during a mood disorder,’’ wassuspended. Each patient’s bipolar symptoms could be no morethan moderately severe, defined as a score of r4 on the ClinicalGlobal Impressions Scale for use in Bipolar Illness (CGI-BP;Spearing et al., 1997) and his or her anxiety symptoms had tobe at least moderately severe (defined as a score of Z4 on theClinician Global Impression Severity Scale (CGI-S; Guy, 1976)).Patients were not required to be in a euthymic state (i.e. withoutmood elevation or depression) at baseline. Patients were excludedif they had an acute, serious or unstable medical illness or clinicalabnormality, were currently receiving an antimanic or moodstabilizing medication, met DSM-IV substance dependencecriteria within the past 6 months, had current psychotic symptomsor a lifetime psychotic disorder (e.g. schizoaffective disorder orschizophrenia), or were judged clinically to be at a serious risk forsuicide. All patients had to discontinue psychotropic drugs 7 daysbefore baseline or 4 weeks in the case of fluoxetine and depotantipsychotics.
2.3. Assessments
Efficacy assessments were conducted at baseline and atweeks 1 to 8. All assessments were conducted by raters blind totreatment assignment. The Clinical Global Improvement Scale for
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–94 85
Anxiety (CGI-21 Anxiety; Sheehan et al., 1993) was the primaryefficacy measure. Secondary efficacy measures were the HamiltonAnxiety Scale (HAM-A; Hamilton, 1959), Sheehan Panic DisorderScale (SPS; Leon et al., 1997; Sheehan, 1983), the Patient GlobalImprovement for Anxiety (PGI-21 Anxiety; Sheehan et al., 1993),the Young Mania Rating Scale (YMRS; Young et al., 1978),the Montgomery Asberg Depression Rating Scale (MADRS;Montgomery and Asberg, 1979), the Clinical Global ImpressionScale for Bipolar Disorder (CGI-BP; Spearing et al., 1997), theSheehan Irritability Scale (SIS; Sheehan, 1999, 2010), the RapidIdeas Scale (RISC; Shytle et al., 2002), the Suicidality TrackingScale (S-STS; Coric et al., 2009) and the Sheehan Disability Scale(SDS; Leon et al., 1997; Sheehan and Sheehan, 2008). Additionalassessments included the proportion of patients who achievedanxiolytic response (defined as Z50% improvement on theCGI-21 Anxiety Scale or alternately as Z50% reduction on theHAM-A total score at study endpoint) and remission (Z70%improvement on the CGI-21 Anxiety Scale or Z70% reductionon the HAM-A total score at study endpoint) and the proportionwho achieved response in bipolar symptoms (Z50% improve-ment in the overall severity score on the CGI-BP overall severityscore at endpoint) and remission in bipolar symptoms (Z70%improvement on the CGI-BP overall severity score at endpoint).
Physical examinations, electrocardiograms (EKGs) and routinelaboratory tests were performed at screen. The EKG and labora-tory tests were repeated at study termination. Vital signs, includ-ing height, weight, heart rate and blood pressure, and adverseevents were recorded weekly. In addition, extrapyramidal symp-toms (EPS) were assessed at each study week using the AbnormalVoluntary Movement Scale (AIMS; Guy, 1976), the SimpsonAngus Scale (SAS; Simpson and Angus, 1970), and the BarnesAkathisia Rating Scale (BARS; Barnes, 1989).
To improve data acquisition and completion accuracy, all datawere collected at the time of each visit using a computerizeddirect entry source data system onto a Tablet PC touch screen.
2.4. Medication
Study medications were encapsulated using a double-dummydesign to protect the blind. (In double-dummy designs, eachparticipant assigned to an active medication group receives twopills (an active intervention and a placebo, called a dummy, thatappears identical to the other intervention) and each placebopatient gets 2 placebo pills. This design was necessary since the2 active treatments in this study were provided in differentcolored tablets.)
Quetiapine XR and matching placebo were provided in 50 mgcapsules. Medication was initiated at 50 mg at bedtime. The dosewas then titrated upwards based on tolerability and adverseevents but not to exceed 150 mg/day by the end of week 2.Subsequently, quetiapine XR could be increased, based on clinicalresponse and tolerability, to a maximum of 300 mg/day. Thistarget dose and titration schedule were chosen to optimizebenefit while minimizing potential side effects. Previous workhas shown that fixed daily doses of 300 and 600 mg quetiapineXR are both effective for anxiety symptoms in bipolar I and IIdepressive disorders (Hirschfeld et al., 2006), but there is someevidence that bipolar II patients experience more side effects andare more likely to drop out of studies at the higher dose of600 mg/day (Suppes et al., 2008). Although higher doses andmore rapid escalation have been recommended in cases of severeacute mania (Constant et al., 2009), our exclusion criteria (nomore than moderate severity of bipolar symptoms) made itunlikely that we would enroll significant numbers of patientswith severe mania and the study was designed to include bipolarpatients with a range of diagnoses (bipolar I, II and NOS).
Moreover, there is good evidence that quetiapine XR is effectivefor anxiety at lower doses (Bandelow et al., 2010).
Divalproex ER and matching placebo were supplied in 500 mgcapsules. Medication was initiated at 500 mg/day at bedtime. Thedose was titrated upwards based on tolerability and adverseevents but not to exceed 1500 mg/day by the end of week 2.Subsequently, divalproex ER could be increased, based on clinicalresponse and tolerability, to a maximum of 3000 mg/day. Thisdosing regimen was selected to minimize side effects, preservethe blind and achieve optimal benefit and serum valproic acidlevels as described below. Although higher initial doses and afaster titration have been recommended for acute mania (Milleret al., 2005), not all bipolar patients are candidates for high initialdoses or rapid escalation (Hirschfeld et al., 2003) and our studywas designed to enroll patients with a range of bipolar diagnosesand mood states.
Blood samples for measurement of blinded serum valproicacid levels were collected approximately 12 h after the previousdose of study medication at week 3 and at week 6. These totalserum valproic acid levels were reviewed by an unblindedmember of the study team who was not involved in any efficacyor safety ratings or treatment decisions. For the purpose of thestudy the target range for 12-hour post-dose valproic acid levelwas specified as 96–144 mg/mL. This target range for divalproexER is 20% higher than the target range of 80–120 mg/mL com-monly used for divalproex DR because the post-dose valproic acidlevel of an equipotent dose of divalproex ER is 20% higher at 12 hthan that for divalproex DR (personal communication, Dr. KellySimontacchi, Neuroscience, Abbott Laboratories). This correctionfactor is necessary to correctly interpret the 12-hour post-dosevalproic acid level of divalproex ER where the trough level wouldbe drawn at 18 h instead of at 12 h.
In the event that a subject had a valproic trough serumvalproic acid level greater than 144 mg/mL at week 3 or 6, theunblinded person alerted the appropriate investigator by emailand reported that the level was high. In the event that a subjecthad a trough serum valproic acid level less than 96 mg/mL at week3 or 6, the unblinded person alerted the appropriate investigatorand reported that the level was low. In either event, to preservethe blind, a corresponding sham email was also sent to a differentinvestigator about a placebo subject at the same time point toreport that the level was high or low. Investigators were expectedto use clinical judgment to determine if an increase or decrease indose was clinically warranted.
A low dose study-prescribed benzodiazepine (lorazepam) wasallowed as needed (up to 2 mg/day in week 1 and up to 1 mg/dayin the week 2) to manage severe anxiety but only in the first2 study weeks. Study-prescribed zolpidem (10–20 mg/day) orzaleplon (10–20 mg/day) was allowed throughout the study forthe management of insomnia and benzotropine (0.5–3 mg/day)was allowed for the management of extrapyramidal symptoms.No other psychopharmacologic agents or structured psych-otherapy were permitted during the trial. Treatment compliancewas monitored by tablet counts.
2.5. Statistical analysis
Demographic and baseline clinical characteristics of thequetiapine XR, divalproex ER and placebo groups were comparedusing analysis of variance (ANOVA) for continuous measures andchi-square for categorical variables.
The relative efficacy of quetiapine XR vs. divalproex ER andplacebo was tested using a last-observation-carried forward(LOCF) repeated-measures ANOVA in which baseline and eachof the 8 weekly assessments for the efficacy variables were thewithin subject factors (labeled ‘‘time’’) and treatment group
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–9486
(labeled ‘‘treatment’’) was the between-subjects factor with3 levels. The central focus in this analysis was on the ‘‘treat-ment-by-time’’ effect showing whether the trajectory of responsediffered over time by treatment group. Supplementary analyseswere conducted in which baseline scores for depression severity(MADRS) and mania severity (YMRS) were entered as covariatesto control for differences at baseline between the two groups. Forvariables on which there was a significant effect further analyseswere performed using analysis of covariance (ANCOVA) to explorethe significance of these differences at different time points whilstcontrolling for baseline differences. In addition, group differencesin baseline-to-endpoint changes in the efficacy measures weretested using LOCF ANOVAs followed by pairwise planned com-parisons (t-tests). These post-hoc tests were not corrected formultiplicity because of the small size of the sample. Response andremission, defined alternately as 50% and 70% improvement inanxiety (using CGI-21 Anxiety and HAM-A total scores at end-point) and as 50% and 70% reduction in bipolar symptoms (usingthe CGI-BP overall severity score at endpoint) were assessed usingchi-square with Fisher’s exact test for 2�2 comparisons.
Group differences in rates of treatment emergent adverseeffects (TAEs) were assessed using chi-square. Group differencesin treatment emergent depression, defined as a MADRS score Z18,with an increase from baseline of 44 at any two consecutiveassessments or at the final visit were assessed using chi-square.Chi-square was also used to assess group differences in treatmentemergent mania, defined as a YMRS score Z16, with an increasefrom baseline of 44 at any two consecutive assessments or at thefinal visit.
Group differences from baseline-to-endpoint in other safetymeasures, including vital signs and weight, were assessedwith ANOVA.
All efficacy analyses were based on the modified intent-to-treat population, defined as all patients who were randomized,took at least one dose of study medication, and had at least onepost-baseline assessment. Safety analyses were based on thesafety population consisting of all those who were randomizedand took at least one dose of study medication. Data for the 3 siteswere pooled because of the small number of patients at two of thesites (Texas and Cincinnati). All statistical tests were two-sidedwith the alpha set at po0.05.
Fig. 1. Flow of
3. Results
3.1. Patients and disposition
A total of 224 patients were screened. Of these 149 patientsmet eligibility criteria, signed informed consents and were ran-domly assigned to receive quetiapine XR (n¼49), divalproex ER(n¼49) or placebo (n¼51). All 149 randomized patients took atleast one dose of study medication and were included in thesafety analyses. A total of 144 patients (47 assigned to quetiapineXR, 46 assigned to divalproex ER and 51 assigned to placebo) hadat least one post-baseline assessment and were included in themodified ITT efficacy analyses. One hundred and eight patients(77% on quetiapine XR, 71% on divalproex ER and 68% on placebo)completed all 8 weeks of treatment. The mean7SD time todiscontinuation for those who withdrew early was 2.272.0weeks for quetiapine XR, 2.472.1 weeks for divalproex ER and3.671.8 weeks for placebo. The most common reasons for earlydiscontinuation were ‘‘lost to follow up’’ and ‘‘chose to leave’’(Fig. 1).
3.2. Demographic and clinical characteristics
As shown in Table 1, there were no statistically significantdifferences between the groups on any demographic or baselineclinical characteristics. The mean age was approximately 39years. Fifty-nine percent were women. Eighty percent met diag-nostic criteria for bipolar I disorder, 17% for bipolar II disorder and3% for bipolar disorder NOS. Mean baseline HAM-A scores for the3 treatment groups were similar and consistent with moderate tosevere anxiety: quetiapine XR (24.876.6); divalproex ER(22.277.5); and placebo (23.278.6). Mean baseline MADRSscores were also similar and consistent with moderate depres-sion: quetiapine XR (26.478.2); divalproex ER (24.577.3); andplacebo (25.877.9). Mean baseline YMRS scores were low butcomparable: quetiapine XR (11.276.1); divalproex ER (12.377.9); and placebo (11.275.7). Only two patients, one on quetia-pine XR and one on divalproex ER, could be described as being ina euthymic state (YMRSo12 and MADRSo10).
Patients.
Table 1Baseline demographic and clinical characteristics of patients (n¼149).
Quetiapine XR (N¼49) Divalproex ER (N¼49) Placebo (N¼51) Analysis
Mean (SD) Mean (SD) Mean (SD) F P
Age, years, mean 41.4 (12.1) 37.5 (12.0) 37.6 (11.6) 1.7 0.18
Age of onset of symptoms 20.5 (13.5) 18.5 (11.0) 18.4 (11.0) 0.49 0.61
Duration of illness, years 20.9 (13.8) 19.0 (11.4) 19.2 (11.9) 0.33 0.71
Note—@2 could not be calculated for some variables because of insufficient numbers.
Abbreviations: CGI-21¼Clinician Global Improvement Scale, 21 items (rated from �10 to þ10).a Work disability ratings were only available for patients who were working at all full or part time or going to school during the study (N¼130).b STS ratings were only available for patients randomized after November 2008 (N¼74).c Mood episode was identified only for patients with bipolar disorder, Type I (N¼119).d Past 30 days only. All subjects were free of medication in the 7 days before baseline;
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–94 87
3.3. Study medication and dosage
The reported mean7SD dose of quetiapine XR was 115.5739.9 mg/day at the end of the second week, 184.1797.2 mg/dayat the end of week 5, and 186.47100.3 mg/day at study termina-tion or the end of week 8. At study termination or week 8, 13patients (27%) reported taking o100 mg/day, 10 (21%) reportedtaking 100–200 mg/day, and 24 (52%) reported taking 200–300 mg/day.
Comparable mean7SD doses of divalproex ER were 12367394 mg/day at the end of the second week, 19927824 mg/day atthe end of week 5, and 19917866 mg/day at study terminationor the end of week 8. At study termination or week 8, 6 patients(13%) reported taking o1000 mg/day, 14 (32%) reported taking1000–2000 mg/day, and 24 (57%) reported taking 2000–3000 mg/day.Twelve-hour post-dose mean7s.d. serum valproic acid levels(micrograms per milliliter) were 87.4734.7 at week 3 and at100.8742.4 at week 6.
Three patients (2 on divalproex ER and 1 on placebo) hadstudy prescribed lorazepam (0.25–2 mg prn in week 1 and/or0.25–1 mg prn in week 2) for anxiety. Nine patients (2 onquetiapine XR, 2 on divalproex ER and 5 on placebo) used
zolpidem (10–20 mg prn) for insomnia in the first 2 study weeks.An additional 5 (2 on quetiapine XR, 2 on divalproex ER and 1 onplacebo) took zolpidem prn at a later week.
3.4. Efficacy
Table 2 shows the mean values for the baseline-to-endpointchanges on the primary and secondary efficacy measures. Table 2also gives the results of the tests of significance for the repeatedmeasures ANOVAs and the ANOVAS of baseline-to-endpointchange.
3.4.1. Anxiety ratings
The repeated measures ANOVA of the CGI-21 Anxiety, used tomeasure overall improvement in anxiety symptoms, demonstrateda significant time effect (p¼0.001) (all 3 groups showed improve-ment over the study) and a significant treatment-by-time inter-action effect (p¼0.02) (the 3 groups displayed a different pattern ofimprovement over the course of the study). Similar results wereobtained on the repeated measures ANOVA of the HAM-A (timeeffect, po0.001; treatment-by-time effect, po0.001) and the
Table 2Outcomes by treatment group.
Least squares mean change from baseline Repeated measures ANOVA
ANOVA Time Effect Rx�TimeEffect
Quetiapine XR (N¼47) Divalproex ER (N¼46) Placebo (N¼51) F p Post-hoc contrasts F p F p
Abbreviations: LOCF¼ last observation carried forward; CGI-21¼Clinician Global Improvement Scale, 21 items (rated from �10 to þ10); PGI-21¼Patient Global
Improvement Scale; MADRS¼Montgomery Asberg Depression Rating Scale; CGI-BP¼Clinical Global Impressions Scale for Bipolar Disorder.a CGI-21 and PGI21 Anxiety Scale ratings range from –10 (very bad, could not be worse) to þ10 (major improvement. Back to normal self).b Past week only.
Fig. 2. Mean CGI-21 Anxiety, HAM-A, SPS and PGI-21 Anxiety scores by treatment, baseline to LOCF endpoint.
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–9488
repeated measures of the SPS (time effect, po0.002; treatment-by-time effect, po0.001). In contrast the repeated measures ANOVAof the PGI-21 Anxiety produced a significant time effect (po0.001),but the treatment-by-time interaction was not significant (p¼0.5).
As shown in Fig. 2, patients on quetiapine XR had a more rapidand more sustained improvement compared to divalproex ER andplacebo on all four of these anxiety measures. Overall, meanimprovement on the CGI-21 Anxiety was 40% higher for quetia-pine XR compared to divalproex ER (an average increase of4.9 points vs. 2.9 points) and 30% higher compared to placebo(an average increase of 4.9 points vs. 3.5 points). These differencesapproached significance (po0.07) on the ANOVA of baseline-to-
endpoint change. For the HAM-A, the mean decrease (improve-ment) from baseline was 45% greater for quetiapine XR comparedto divalproex ER (an average decrease of �11.7 points vs. �6.4points) and 30% greater for quetiapine XR compared to placebo(�11.7 vs. �8.4). This difference was statistically significant(po0.003) on the baseline-to-endpoint ANOVA of change withpairwise contrasts (po0.05) showing superiority for quetiapineXR over both divalproex ER and placebo. Comparable results wereobtained with the SPS. The mean decrease (improvement) frombaseline was 40% greater for quetiapine XR compared to dival-proex ER (�24.4 vs. �16.4) and 25% greater compared to placebo(�24.4 vs. �18.3). This difference was significant on the
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–94 89
baseline-to-endpoint ANOVA of change (po0.01) with pairwisecontrasts (po0.05) indicating significant superiority for quetia-pine XR over both divalproex ER and placebo.
3.4.2. Depression, mania and irritability ratings
The repeated measures ANOVA of the MADRS, used to measuredepression, demonstrated a significant time effect (po0.02)(all 3 groups showed improvement each week over the study)and a significant treatment-by-time interaction effect (po0.03)(the 3 groups displayed a different pattern of improvement overthe course of the study). As was the case with the anxietymeasures, the quetiapine XR group had an earlier and moresustained improvement compared to the divalproex ER andplacebo. At endpoint, the mean reduction on the MADRS totalscore was 50% greater (�11.5 vs. �5.5) than that on divalproexER and 36% greater (�11.5 vs. �7.5) than that on placebo. Thisgroup difference was significant (po0.009) on the ANOVA ofbaseline-to-endpoint change with pairwise contrasts showingstatistically significant superiority (po0.05) for quetiapine XRover both divalproex ER and placebo.
The repeated measures ANOVAs of the YMRS, a measure ofmania, demonstrated a significant time effect (po0.001), but nota significant treatment-by-time interaction effect. Similar resultswere obtained on the repeated measures ANOVAs of the mania,depression and overall severity scores of the CGI-BP, but thebaseline-to-endpoint ANOVAs of the CGI-BP depression and over-all severity scores were both statistically significant (po0.04)with pairwise contrasts favoring quetiapine XR compared todivalproex ER on depression severity.
By contrast, the repeated measures ANOVA for the RIS, ameasure of rapid ideas, and the SIS, a measure of irritability, bothdemonstrated significant treatment-by-time interaction effects(po0.02 for the RIS and po0.009 for the SIS) as well assignificant time effects (po0.0002 for the RIS and po0.0005 forthe SIS). On the RIS, quetiapine XR produced a 30% greater meandecrease on the RIS compared to divalproex ER (�28.9 vs. �19.7).On the SIS, quetiapine XR produced a 25% greater mean decreasecompared to divalproex ER (�29.8 vs. �22.6) and a 33% greaterdecrease compared to placebo (�29.8. vs. �19.4) (Fig. 2d). Forthese 2 scales, however, the ANOVAs of baseline-to-endpointchange were not significant (po0.09 for the RIS and po0.08 forthe SIS) and pairwise contrasts were therefore not performed.
3.4.2.1. Effects of baseline depression and mania on anxiety
outcomes. The inclusion of depression severity (MADRS score atbaseline) as a covariate in the repeated measures of CGI-21Anxiety scores produced a significant between-subjects effectfor depression (po0.0007) and a significant depression-by-timeinteraction effect (po0.01). The ANCOVAs conducted to explorethe significance of these differences at different time pointsindicated that the severity of baseline depression had asignificant effect at every time point from week two onwardswhile quetiapine XR demonstrated a significant effect at weeks 3,4, 7 and 8. The inclusion of mania severity (YMRS score atbaseline) as a covariate in the repeated measures of the CGI-21Anxiety scores did not produce any significant between-subject orinteraction effects. These results suggest that baseline depressionseverity (but not mania severity) played a key role in response.
Additional post-hoc tests indicated that quetiapine XR treatedpatients with low-to-moderate depression severity (defined asMADRS o30, Muller et al., 2003) had consistently greaterimprovement on the CGI-21 Anxiety Scale compared to thosewith high depression severity (MADRS 430) from week oneonwards. These differences were significant (po0.02) from week6 to study endpoint. Similar results were obtained for compar-isons of divalproex ER treated patients with high and low
depression severity at baseline, but the magnitude of effect waslower and significant (po0.05) only at week 6. At study endpointthe mean CGI-Anxiety improvement on quetiapine XR wassignificantly higher for patients with a baseline MADRS of o30vs. those with a baseline a MADRS of 430 (6.6 vs. 2.8; po0.001).
3.4.3. Disability
The repeated measures ANOVA of the total SDS score, used tomeasure overall impairment in functioning, produced a signifi-cant treatment-by-time interaction effect (po0.001). Overall, all3 groups demonstrated a reduction in overall disability starting atweek 1. This reduction was more consistent and sustained for thequetiapine XR group compared to the divalproex ER group. Atstudy endpoint, the mean decrease in the total disability scorewas �6.5 for quetiapine XR, �3 for divalproex ER and �5.3 forplacebo. This difference was statistically significant on the ANOVAof baseline-to-endpoint change (po0.03) with pairwise contrastsindicating superiority for quetiapine XR over divalproex ER (po0.05)but not over placebo.
The repeated measures analyses of all 3 SDS subscales,measuring impairment in work, social and family life, alsoproduced significant treatment-by-time interactions (po0.009for work disability, po0.001 for social disability and po0.001for disability in family life). On all 3 subscales, mean improve-ment was 2–3 times greater for quetiapine XR vs. divalproex ERbut not substantially greater than placebo. This difference wasstatistically significant only on the ANOVAs of baseline-to-end-point change for the SDS subscales, social disability and familydisability. (In both cases, pairwise contrasts showed superiority(po0.05) for quetiapine XR over both divalproex ER but notplacebo.) Two additional SDS items (number of days missed workin the past week and number of days underproductive at work inthe last week) both demonstrated significant treatment-by-timeeffects (po0.02 and po0.05, but the ANOVAs of baseline-to-endpoint change were not significant.
3.4.4. Suicidality
S-STS data, measuring suicidality in the past month, wasavailable for the last 74 enrolled patients (19 on quetiapine XR,28 on divalproex ER and 27 on placebo). Of these, 23 (6 onquetiapine XR, 11 on divalproex ER and 6 on placebo) had ahistory of a previous suicide attempt. At baseline, the mean S-STStotal score was low (1.572.3, range¼0–12) and there were nostatistically significant differences across the 3 groups. On therepeated measures ANOVA there was significant time effect(p¼0.03), but the treatment-by-time interaction effect was notsignificant. Although the magnitude of the mean S-STS scoredecrease was higher in the quetiapine XR group (�0.9) comparedto the other 2 groups (�0.07 for divalproex ER and �0.29 forplacebo) this difference was not significant on the ANOVA ofbaseline-to-endpoint change. Overall, 36% (N¼27) improved and49% (N¼36) of the patients experienced no change on thismeasure, but 15% (N¼11) (1% on quetiapine XR, 5% on placeboand 8% on divalproex ER) experienced a worsening.
3.4.5. Response and remission
3.4.5.1. Anxiety. Response rates (Z50% improvement in CGI-21Anxiety score) were higher for those treated with quetiapineXR compared to divalproex ER and placebo (62% vs. 35% and 47%).This difference was significant (@2
¼6.9, df¼2, po0.03) withquetiapine XR showing superiority compared to divalproex ER(po0.01) but not to placebo. Response rates were also higher forquetiapine XR compared to divalproex ER and placebo whenresponse was defined as Z50% reduction in the HAM-A total
Fig. 3. Responders (Z50% improvement at LOCF endpoint).
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–9490
score: 60% vs. 28% and 43%. Again, this difference was significant(@2¼9.4, po0.008), but pairwise comparisons only showed statistical
superiority (po0.01) for quetiapine compared to divalproex ER(Fig. 3).
Remission rates (Z70% improvement) were 45% for quetia-pine XR, 22% for divalproex ER and 29% for placebo on the CGI-21Anxiety score. This difference was significant (@2
¼5.8, df¼2,po0.0.05) with pairwise comparisons showing superiority forquetiapine XR compared to divalproex ER (po0.02). Similarremission rates were using the criterion Z70% reduction insymptoms on the endpoint HAM-A: 40% for quetiapine XR, 24%for divalproex ER and 20% for placebo. This difference, however,did not quite reach statistical significance (@2
¼5.7, df¼2,po0.06). Somewhat different results were obtained when anxietyremission was defined as a score of 7 or less on the HAM-A. In thiscase, anxiety remission was achieved by 49% of patients onquetiapine XR vs. 24% on divalproex ER and 25% on placebo. Thisdifference was significant (@2
¼8.2, df¼2, po0.02) with pairwisecontrasts indicating a higher rate of remission for quetiapine XRcompared to divalproex ER and placebo (po0.02).
3.4.5.2. Bipolar symptoms. Overall, response and remission rateswere lower for bipolar symptoms compared to response rates foranxiety symptoms. Response (450% improvement on the CGI-BPoverall severity rating at study endpoint) was 40% for quetiapine XR,17% for divalproex ER and 36% for placebo. This difference wassignificant (@2
¼6.8, df¼2, o0.03) with pairwise contrasts indicatingsuperiority for quetiapine XR compared to divalproex ER (po0.02)
Remission, defined as a 70% or greater improvement on theCGI-BP overall severity rating at study endpoint, was low for allthree groups but higher for quetiapine XR (19%) compared todivalproex ER (11%) and placebo (12%). This difference was notstatistically significant (@2
¼1.5, df¼2, o0.46). When remissionwas defined alternatively as YMRSo12 and MADRSo10, betterresults were obtained (40% for quetiapine XR, 28% for divalproexER and 39% for placebo). This difference, however, was notstatistically significant (@2
¼51.81.9, df¼2, po0.39).
3.4.6. Subgroup analyses
3.4.6.1. Bipolar I, II and NOS disorders. In the bipolar I subgroup ofpatients (N¼115), the mean change in the HAM-A total scorefrom baseline to endpoint, after adjusting for baseline differenceson the HAM-A, was �12.2 for quetiapine XR, �6.4 for divalproexER, and �8.2 for placebo (F¼4.5, df¼2, po0.01) (po0.05 for theplanned comparison of quetiapine XR vs. divalproex ER and forthe planned comparison of quetiapine XR vs. placebo). In the
smaller subgroup of patients with bipolar II or bipolar NOSdisorder (N¼29), the mean change on the HAM-A total scorefrom baseline to endpoint was smaller for quetiapine XR andlarger for divalproex ER compared to the changes in bipolar Idisorder �9.4 for quetiapine XR, �10.3 for divalproex ER and�8.9 for placebo. These group differences were not significant.
3.4.6.2. Mood state. In the bipolar I subgroup of patients presentingwith depressed mood, as identified on the MINI, at study entry(N¼46), the mean change in the HAM-A score from baseline toendpoint, after adjusting for baseline differences on the HAM-A,was �14.4, �8.7, and �9.4 for quetiapine XR, divalproex ER andplacebo respectively. In the subgroup presenting with manic orhypomanic mood (N¼39), comparable decreases were �11.3,�5.7, and �8.2 for quetiapine XR, divalproex ER and placeborespectively. In the subgroup presenting with mixed state (N¼34),the decreases were �10.6, �4.2, and �7.3. Although these groupdifferences were not statistically significant, the pattern of resultssuggests that quetiapine XR has greater anxiolytic effect thandivalproex ER across a range of mood states. Additionally, itappears that both active compounds (quetiapine XR and divalproexER) produce more anxiolytic effect for depressed mood states thanfor mixed states or hypomanic symptoms.
To further investigate the role of manic symptoms on outcome,we examined the improvement on the CGI-21 Anxiety Scale forpatients with baseline YMRS scores o12, o8 and o5. Withinthe subgroup with baseline YMRS scores o12 (n¼90), meanimprovement was significantly higher for quetiapine XR com-pared to divalproex ER and compared to placebo (po0.05).Within the subgroup YMRS o8 (n¼46), improvement wassignificantly higher for quetiapine XR compared to placebo(po0.05). No group differences were found for the much smallerYMRS o5 subgroup (n¼21).
3.4.6.3. Panic disorder. In the subgroup with current panicdisorder (n¼113), the mean decrease on the SPS, after adjustingfor baseline differences, was �24.4 for quetiapine XR, �14.2 fordivalproex ER and �18.9 for placebo (F¼3.0.4, df¼2, po0.05,po0.05 for quetiapine XR vs. divalproex ER). Similarly in thissubgroup, the mean HAM-A decrease, after adjusting for baselinedifferences, was �11.5 for quetiapine XR, �5.9 for divalproex ERand �8.6 for placebo (F¼4.4, df¼2, po0.01, po0.05 forquetiapine XR vs. divalproex ER).
In the subgroup without current panic disorder (N¼33), themean SPS decrease, after adjusting for baseline, was �23.3 forquetiapine XR, �18.1 for divalproex ER and �16.7 for placebo.This difference was not statistically significant nor were statisti-cally significant differences found in this subgroup on analyses ofbaseline-to-endpoint decreases on the HAM-A where meandecreases, after adjusting for baseline, were �11.7 for quetiapineXR, �8.9 for divalproex ER and �7.7 for placebo.
3.4.7. Dose–response relationships
To evaluate potential dose–response relationships, we exam-ined the mean HAM-A decrease, adjusted for the baseline HAM-A,for three endpoint dose levels of quetiapine XR and separatelyfor three endpoint dose levels of divalproex ER. In the quetiapineXR group, the mean HAM-A decrease was �10.1, �14.6, and�12.1 for endpoint dose levels of o100 mg/day (N¼13), 100–199 mg/day (N¼10), and 200–300 mg/day (N¼24) respectively.These differences were not statistically significant (F¼0.8, po0.47).In the divalproex ER group, the mean HAM-A decrease was �5.7,�3.4 and �7.8 for endpoint dose levels of o1000 mg/day (N¼6),1000–1999 mg/day (N¼14), and 2000–3000 mg/day (N¼26)respectively. These differences were not statistically significant
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–94 91
(F¼1.7, po0.18) nor was there any evidence that when divalproexER treated patients were grouped at endpoint according to 12-hourpost-dose valproic acid level concentrations (low, in range, high), theresponse on the HAM-A increased or decreased as a function ofvalproic acid level.
3.4.8. Requirement for benzodiazepine
Only 1 of the 10 patients using a benzodiazepine in the monthbefore baseline withdrew early. This patient was on divalproex-ER and withdrew at week 2. Further, only 3 patients (2 ondivalproex ER and 1 on placebo) required study prescribedlorazepam in the first 2 weeks.
3.5. Safety and tolerability
3.5.1. Adverse events
One patient in the quetiapine XR group, 3 in the divalproex ERgroup and one in the placebo group discontinued because of anadverse event. The quetiapine XR patient left the study at week7 because of excessive sedation. One divalproex-treated patientdiscontinued at week 2 because of worsening depression, asecond left the study at week 2 because of worsening socialwithdrawal and a third had to be withdrawn from the study atweek 7 after becoming violent and striking his girlfriend’sdaughter. One placebo-treated patient withdrew at week 4 aftercomplaining of worsening depression.
The most common treatment emergent adverse events (TAEs),reported by 5% or more patients in each group, are shown inTable 3. The quetiapine XR group had a significantly higherfrequency of treatment-emergent dry mouth (31%) compared todivalproex ER (6%) and placebo (14%) (@2
¼12.4, po0.006). Treat-ment emergent headache was more common in the divalproex ERand placebo groups (24% for each vs. 8% in the quetiapine XRgroup). This difference, however, only approached statistical sig-nificance (po0.07). There were no other statistically significant orclinically relevant differences. Almost all of the treatment emer-gent adverse events were judged as mild or moderate in severity.
3.5.2. Extrapyramidal symptoms
Mean baseline-to-endpoint changes in extrapyramidal symp-toms (EPS), as measured on the AIMS, SAS and BARS, wereminimal and did not differ significantly by treatment group.The change scores over the 8 week study were þ0.02, �0.14,�0.19 on the AIMS, þ0.10, �0.78 �0.25 on the BARS and �0.02,þ0.008, �0.009 on the SAS for quetiapine XR, divalproex ER andplacebo respectively.
Table 3Treatment emergent adverse events with frequency 45%.
Quetiapine XR (N¼49) Divalproex E
N % N
Drowsiness/Sleepiness 24 49 18
Dry mouth 15 31 3
Nausea or nausea/vomiting 9 18 12
Tinglinga 8 16 1
Increased appetite 7 14 6
Sedation 6 12 3
Headache 4 8 12
Lightheadednessa 4 8 2
Tirednessa 3 6 1
Diarrheaa 1 2 2
Dizzinessa 1 2 3
a Chi-square could not be calculated because the numbers in some groups were to
3.5.3. Vital signs, weight, laboratory measures and EKG
Mean baseline-to-endpoint changes in systolic or diastolicblood pressure and in heart rate, taken sitting and standing, werenot clinically or statistically significant.
The mean7SD and median weight increase was 4.076.8 lb(median: 3 lb) in the quetiapine XR group, 2.775.9 lb (median:3 lb) in the divalproex ER group and �0.277.9 lb (median: 0 lb)in the placebo group. Because the weight increase data was notnormally distributed, we used the nonparametric Kruskal–Wallistest to evaluate the significance of these differences. This testdemonstrated a significant overall effect (@2
¼11, df¼2,po0.003) with pairwise post-hoc tests indicating greater weightgain for quetiapine XR compared to placebo (po0.001) and fordivalproex ER compared to placebo (po0.03). Overall, however,weight increases of Z7% over baseline were low (12% in thequetiapine XR group; 8% in the divalproex ER; and 4% in theplacebo group). These differences were not statistically signifi-cant. There was no evidence that weight gain was associated withmore or less improvement in the primary outcome measure (CGIAnxiety score). Pearson’s correlation coefficient, used to assess thesignificance of the association between weight change and theCGI-Anxiety at endpoint, was low and non-significant (r¼0.27,p¼0.06 for quetiapine XR and r¼0.09, p¼0.54 for divalproex ER).
There were no clinically relevant treatment emergent changesin any of the groups in mean changes on laboratory values or EKG.
3.5.4. Treatment-emergent depression and mania
The rate of treatment-emergent depression (defined as aMADRS score Z18, with an increase from baseline of 44 atany two consecutive assessments or at the final visit) was low:6.3% with quetiapine XR vs. 15.2% with divalproex ER and 17.7%with placebo. This difference was not statistically significant andin a third of the cases on active medication (1 of 3 on quetiapineXR and 2 of 7 on divalproex ER), treatment resistant depressionresolved before the final visit. The rate of treatment-emergentmania, defined as a YMRS score Z16 at any two consecutiveassessments or at the last observation, was also low and com-parable across groups: 8.5% with quetiapine XR vs. 6.5% withdivalproex ER and 9.8% with placebo. For most patients in theactive treatment groups (3 of 4 on quetiapine XR, 2 of 3 ondivalproex ER), treatment emergent mania resolved before thefinal visit.
4. Discussion
To our knowledge this is the first randomized, double-blind,placebo-controlled study comparing a conventional mood stabilizer
R (N¼49) Placebo (N¼51)
% N % @2 p
37 17 33 2,8, df¼2 0.24
6 7 14 12.4, df¼2 0.006
24 14 27 1.2, df¼2 0.55
2 1 2 – –
12 7 8 0.09 0.95
6 3 6 – –
24 12 24 5.4, df¼2 0.07
4 1 2 – –
2 3 6 – –
4 4 8 – –
6 1 2 – –
o small.
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–9492
with a second generation antipsychotic in the treatment of asyndromal anxiety disorder associated with a bipolar disorder inadult outpatients. The results, while preliminary, suggest that que-tiapine XR is an effective, well-tolerated anxiolytic for patients with acomorbid panic disorder or GAD. Compared with divalproex ER andplacebo, quetiapine XR, taken once daily at bedtime, at a meanendpoint dose of 1867100 mg/day produced an earlier and moresustained improvement from baseline on the CGI-21 Anxiety scale,HAM-A and the SPS. Quetiapine XR also produced significantly moreimprovement than divalproex ER and placebo in depression symp-toms, as measured by the MADRS, and it was superior to divalproexER in reducing total, social and family life disability.
The results are consistent with previous findings showing thatquetiapine reduces anxious symptoms in bipolar depression(Hirschfeld et al., 2006; Thase et al., 2006, Tohen et al., 2007,Lydiard et al., 2009). In those studies, the results were morerobust among bipolar I patients than bipolar II or NOS patients. Inthe current study of bipolar patients with a comorbid anxietydisorder, anxiety reduction, as measured on the HAM-A, wassignificantly greater for quetiapine XR compared to both dival-proex ER and placebo in the bipolar I subgroup (N¼115) but notin the smaller bipolar II subgroup (N¼29).
The mean baseline-to-endpoint HAM-A decrease for quetia-pine XR (�11.7) was numerically greater than the mean decreasewe observed for risperidone (�7.6) in a comparable 8-weekdouble-blind placebo-controlled study of patients with bipolardisorder and comorbid panic disorder or GAD using the same sites(Sheehan et al., 2009). However, it was similar to the �10.1 and�10.5 mean HAM-A reductions reported by Lydiard et al. (2009)in an analysis of the effects of quetiapine IR at fixed doses of300 mg/day and 600 mg/day on anxious symptoms in patientswith bipolar depression.
Overall, response and remission rates for overall bipolarsymptom severity were lower than response and remission foranxiety symptoms. These results suggest that changes in bipolarsymptoms may lag behind change in anxiety. More prolonged useof quetiapine XR in a larger sample of bipolar patients withcomorbid anxiety is worth exploring.
Both of the active medications were well tolerated. Only1 patient on quetiapine XR (2%) and 3 on divalproex ER (6%)withdrew from the study because of a treatment related adverseevent. The very low dropout rate for adverse events on quetia-pine-XR may have been a function of the high proportion ofpatients with BDI. This population has been found to be lesssensitive to side effects than those with BDII (Suppes et al., 2008).Compared to patients on divalproex ER and placebo, patients inthe quetiapine XR group experienced higher rates of treatmentemergent dry mouth (po0.006). Although treatment emergentsedation, drowsiness and sleepiness occurred more often in thequetiapine XR group and headaches occurred less often, groupdifferences were not statistically significant. Neither of the activetreatments was associated with clinically significant changes inEKG abnormalities or EPS symptoms. Quetiapine XR and dival-proex ER, however, were both associated with a significantlygreater median weight gain compared to placebo (3 lb. vs. 0 lb).
5. Limitations
The study had several limitations. First, the relatively smallsample size patients per treatment group may have limited theability to detect treatment differences in the ANOVAs andrepeated measures ANOVAs. Second, this 8-week study does notprovide information regarding the relative efficacy or tolerabilityof quetiapine XR and divalproex ER when used for longer-termmaintenance. Third, in light of the relatively low mean valproic
acid level at week 6, it is possible that some of the patientsreporting high doses of divalproex ER were not taking the fulldoses they reported. Fourth, data had to be pooled since most ofthe patients were enrolled at one site (University of SouthFlorida). Finally, the study was restricted to patients with bipolardisorder and a co-occurring panic disorder or generalized anxietydisorder (GAD). Results may not be generalizable to patients withother anxiety disorders such as post-traumatic stress disorder(PTSD), obsessive-compulsive disorder (OCD) or Social AnxietyDisorder (SAD).
5.1. Clinical implications
Current treatment guidelines for bipolar disorders recommendtreating co-occurring anxiety disorders concurrently with bipolardisorder (Suppes et al., 2005, Perlis, 2005). However, evidencebased treatment options for bipolar patients with co-occurringanxiety disorders are still few. The findings of the present studysuggest that quetiapine XR monotherapy reduces anxiety inpatients with bipolar disorder and a co-occurring panic disorderor GAD and is well tolerated over 8 weeks of treatent. Furtherstudies in patients with bipolar disorder and comorbid anxietydisorders are needed.
Role of funding sourceFunding for this study was provided by Astra Zeneca Pharmaceuticals. Astra
Zeneca had no further role in the study design, in the data collection, monitoring,
statistical analysis and interpretation of data, in the writing of the report or in the
decision to submit the paper for publication.
Conflict of interestDV Sheehan, MD, MBA: During the period of January 2004–April 2012, Dr. DV
Sheehan has received grant support from Avera Pharmaceuticals, Astra Zeneca,
Cenerx Biopharma, Cephalon, Eli Lilly, GlaxoSmithKline, Indevus Pharmaceuticals,
Janssen Pharmaceutica, Jazz Pharmaceuticals, LaboPharm Inc., MediciNova, MGH-
CTNI, Otsuka, Pfizer U.S. Pharmaceuticals, Repligen, Sanofi-Synthelabo, Stanley
Medical Research Institute, Sunovion/DSP, Takeda and UCB Pharma Inc. He has
served on advisory boards for Roche, Sagene Pharm, Otsuka, Forest Laboratories,
Foundation for Improving Quality, NovaDel, Labopharm, Lundbeck, Neuronetics
and the International Society for CNS Drug Development. He has been a consultant
for Avera Pharmaceuticals, Astra Zeneca, Cephalon, Eli Lilly, Eisai Pharmaceuticals,
Forest Laboratories, GlaxoSmithKline, INC Research, Janssen Pharmaceutica, Jazz
Pfizer US Pharmaceuticals, Schlesinger Medical, Shire Inc., Solvay Pharmaceuticals
and Wyeth Pharmaceuticals (shown combined as this was a business collabora-
tion on a new drug under study), and UCB Pharmaceuticals. From January 2009–
April 2012, Dr. Suppes has not served on any Consulting or Advisory Boards.
D Amado: During the period of January 2004–April 2012, Ms. Amado has no
conflict of interest to report.
AcknowledgmentsWe wish to thank Lizbhet Delgado, Pharm D. and Sumeet Roy, Ph.D. for their
invaluable assistance in the initial planning and implementation of the study.
References
Baker, R.W., Chengappa., K.N., Baird, J.W., Steingard, S., Christ, M.A., Schooler, N.R.,1992. Emergence of obsessive-compulsive symptoms during treatment withclozapine. Journal of Clinical Psychiatry 53 (12), 439–442.
Bandelow, B., Chouinard, G., Bobes, J., Ahokas, A., Eggens, I., Liu, S., Erikkson, H.,2010. Once-daily extended release quetiapine fumarate (quetiapine XR)monotherapy in generalized anxiety disorder (GAD). Data from a randomized,double-blind, placebo and active-controlled study. International Journal ofNeuropsychopharmacology 13 (3), 305–320.
Barnes, T.R., 1989. A rating scale for drug induced akathisia. British Journal ofPsychiatry 154, 672–676.
Boylan, K.R, Bieling, P.J., Marriott, M., Begin, H., Young, L.T., MacQueen, G.M., 2004.Impact of comorbid anxiety disorders on outcome in a cohort of patients withbipolar disorder. Journal of Clinical Psychiatry 65, 1106–1113.
Carter, T.D., Mundo, E., Parikh, S.V., Kennedy, J.L., 2003. Early age at onset as a riskfactor for poor outcome of bipolar disorder. Journal of Psychiatric Research 37,297–303.
Constant, E., Seghers, A., Ranalli, E., Ryckmans, V., Snauwaert, P., Parent, M.,Vandenhoven, G., 2009. Safety tolerability and efficacy of a rapid doseescalation of quetiapine in bipolar I mania: The FATIMA study. Acta Neurop-sychiatrica 21 (3), 125–131.
Coric, V., Stock, E.G., Pultz, J., Marcus, R., Sheehan, D.V., 2009. Sheehan SuicidalityTracking Scale (Sheehan—STS): preliminary results from a multicenter clinicaltrial in generalized anxiety disorder. Psychiatry (Edgemont) 6 (1), 26–31.
Davis, L., Bartolucci, A., Petty, F., 2005. Divalprex in the treatment of bipolardepression: a placebo controlled study. Journal of Affective Disorders 85 (3),259–266.
de Haan, L., Beuk, N., Hoogenboom, B., Dingemans, P., Linszen, D., 2002. Obsessive-compulsive symptoms during treatment with olanzapine and risperidone:prospective study of 11 patients with recent-onset schizophrenia or relateddisorders. Journal of Clinical Psychiatry 63 (2), 104–107.
Derry, S., Moore, R.A., 2007. Atypical antipsychotics in bipolar disorder: systematicreview of randomized trials, 2007. BMC Psychiatry 7 (40), 1–17.
Dilsaver, S.C., Akiskal, H.S., Akiskal, K.K., Benazzi, F., 2006. Dose–response relation-ship between number of comorbid anxiety disorders in adolescent bipolar/unipolar disorders, and psychosis, suicidality, substance abuse and familiality.Journal of Affective Disorders 96, 249–258.
Feske, U., Frank, E., Mallinger, A.G., Houck, P.R., Fagiolini, A., Shear, M.K.,Grochocinski, V.J., Kupfer, D.J., 2000. Anxiety as a correlate of response tothe acute treatment of bipolar I disorder. American Journal of Psychiatry 157(6), 956–962.
Frank, E., Cyranowski, J.M., Rucci, P., Shear, M.K., Fagiolini, A., Thase, M.E, Cassano,G.B., Grochocinski, V.J., Kostelnik, B., Kupfer, D.J., 2002. Archives of GeneralPsychiatry 59, 905–911.
Freeman, M.P., Freeman, S.A., McElroy, S.L., 2002. The comorbidity of bipolar andanxiety disorders: prevalence, psychobiology and treatment issues. Journal ofAffective Disorders 68, 1–23.
Frye, M.A., Altshuler, L.L., McElroy, S.L., et al., 2003. Gender differences inprevalence, risk, and clinical correlates of alcoholism comorbidity in bipolardisorder. American Journal of Psychiatry 160, 883–889.
Gao, K., Muzina, D., Gajwani, P., Calabrese, J.R., 2006. Efficacy of typical and secondgeneration antipsychotics for primary and comorbid anxiety symptoms ordisorders—a review. Journal of Clinical Psychiatry 67 (9), 1327–1340.
Ghaemi, S.N., Hsu, D.J., Soldani, F., Goodwin, F.K., 2003. Antidepressanets in bipolardisorder: the case for caution. Bipolar Disorders 5 (6), 421–433.
Ghaemi, S.N., Hsu, D.J., Thase, M.E., Wisniewski, S.R., Nierenberg, A.A., Miyahaa, S.,Sachs, G., 2006. Pharmacological treatment patterns at study entry for the first500 STEP-BD participants. Psychiatric Services 57, 660–665.
Guy, W., 1976. ECDEU Assessment Manual for Psychopharmacology. Revised forMental Health. DHEW Pub (ADM). National Institute for Mental Health,Rockville, MD.
Hamilton, M., 1959. The assessment of anxiety states by rating. British Journal ofMedical Psychology 32, 50–55.
Henry, C., Van de Bulke, D., Bellivier, F., Etain, B., Rouillon, F., Leboyer, M., 2003. Anxietydisorders in 318 bipolar patients: prevalence and impact on illness severity andresponse to mood stabilizer. Journal of Clinical Psychiatry 64 (3), 331–335.
Hirschfeld, R.M., Baker, J.D., Wozniak, P., Tracy, K., Sommerville, K.W., 2003. Thesafety and early efficacy of oral-loaded divalproex versus standard-titrationdivalproex, lithium, olanzapine, and placebo in the treatment of acute maniaassociated with bipolar disorder. Journal of Clinical Psychiatry 64 (7), 841–846.
Hirschfeld, R.M.A., Weisler, R.H., Raines, S.R., Macfadden, W., for the Bolder StudyGroup, 2006. Quetiapine in the treatment of anxiety in patients with bipolar Ior II depression: a secondary analysis from a randomized, double-blind,placebo-controlled study. Journal of Clinical Psychiatry 67 (3), 355–362.
Kessler, R.C., Chiu, W.T., Demler, O., Walters, E.E., 2005. Prevalence, severity, andcomorbidity of twelve-month DSM-IV disorders in the National ComorbiditySurvey Replication (NCS-R). Archives of General Psychiatry 62 (6), 617–627.
Kessler, R.C., Rubinow, D.R., Holmes, C., Abelson, J.M., Zhao, S., 1997. Theepidemiology of DSM-III-R bipolar I disorder in a general population survey.Psychological Medicine 27 (5), 1079–1089.
Khan, A, Joyce, M, Eggens, I., et al., 2008. Extended release quetiapine fumarate(quetiapine XR): a once daily monotherapy in the treatment of patients withgeneralized anxiety disorder (GAD). Paper presented at: Anxiety DisordersAssociation of America Congress. March 6–9, 2008, Savannah, GA.
Leon, A.C., Olfson, M., Portera, L., Farber, L., Sheehan, D.V., 1997. Assessingpsychiatric impairment in primary care with the Sheehan Disability Scale.International Journal of Psychiatry in Medicine 27 (2), 93–105.
Lum, M., Fontaine, R., Elie, R., Ontiveros, A., 1990. Divalproex sodium’s antipaniceffect in panic disorder: a placebo-controlled study. Biological Psychiatry 27,164A–165A.
Lydiard, R.B., Culpepper, L., Schioler, H., Gustafsson, U., Paulsson, B., 2009.Quetiapine monotherapy as treatment for anxiety symptoms in patients withbipolar depression: a pooled analysis of results from 2 double-blind, rando-mized placebo-controlled studies. Journal of Clinical Psychiatry 11 (5), 215–225.
MacKinnon, D.F., McMahon, F.J., Simpson, S.G., McInnis, M.G., De Paulo, J.R., 1997.Panic disorder with familial bipolar disorder. Biological Psychiatry 42, 90–95.
MacKinnon, D.F., Zandi, P.P., Gershon, E.S., Nurnberger, J.I., DePaulo, J.R., 2003.Association of rapid mood switching with panic disorder and familial panic riskin familial bipolar disorder. American Journal of Psychiatry 160, 1696–1698.
McElroy, S.L., Altshuler, L.L., Suppes, T., Keck, P.E., Frye, M.A., Denicoff, K.D., Nolen,W.A., Kupka, R.W., Leverich, G.S., Rochussen, J.R., Rush, J., Post, R.M., 2001. Axis Ipsychiatric comorbidity and its relationship to historic illness variables in 288patients with bipolar disorder. American Journal of Psychiatry 158 (3), 420–426.
McIntyre, R., Katzman, M., 2003. The role of atypical antipsychotics in bipolardepression and anxiety disorders. Bipolar Disorders 5 (suppl 2), 20–35.
D.V. Sheehan et al. / Journal of Affective Disorders 145 (2013) 83–9494
Miller, B.P., Perry, W., Moutier, C.Y., Robinson, S.K., Feifel, D.., 2005. Rapid oralloading of extended release divalproex in patients with acute mania. GeneralHospital Psychiatry 27 (3), 218–221.
Montgomery, S., Asberg, M., 1979. A new depression scale designed to be sensitiveto change. British Journal of Psychiatry 134, 382–389.
Muller, M.J., Himmerich, H., Kienzle, B., Szegedi, A., 2003. Differentiating moderateand severe depression using the Montgomery-Asberg depression rating scale(MADRS). Journal of Affective Disorders 77, 255–260.
Nardi, A.E., Nascimento, I., Freire, R.C., Veras, A.B., Valfrido, L, Valenca, A.M., Lopes,F.L., Soares-Filho, G., Levitan, M.N., Carvalho, M.R., da Costa, R.T., King, A.L.,Mezzasalma, M.A., Grivet, L.O., Rassi, A., Versiani, M., 2007. Demographic andclinical features of panic disorder comorbid with bipolar I disorder: a 3-yearretrospective study. Journal of Affective Disorders 106 (1–2), 185–189.
Perlis, R.H., 2005. The role of pharmacologic treatment guidelines for bipolardisorder. Journal of Clinical Psychiatry 66 (suppl 3), 37–47.
Perlis, R.H., Miyahara, S., Marangell, L.B., Wisniewski, S.R., Ostacher, M., DelBello,M.P., Bowden, C.L., Sachs, G.S., Nierenberg, A.A., 2004. Long-term implicationsof early onset in bipolar disorder: data from the first 1000 participants in thesystematic treatment enhancement program for bipolar disorder (STEP-BD).Biological Psychiatry 55, 875–881.
Pini, S., Cassano, G.B., Simonini, E., Savino, M., Russo, A., Montgomery, S.A., 1997.Prevalence of anxiety disorders comorbidity in bipolar depression, unipolardepression and dysthymia. Journal of Affective Disorders 42 (2–3), 145–153.
Rotondo, A, Mazzanti, C, Dell’Osso, L, Rucci, P, Sullivan, P, Bouanani, S, Gonnelli, C,Goldman, M.D., Cassano G., D, 2002. Catechol O-Methyltransferase, Serotonintransporter, and tryptophan hydroxylase gene polymorphisms in bipolardisorder patients with and without co-morbid panic disorder. AmericanJournal of Psychiatry 159, 23–29.
Salvi, V., Faglioni, A, Swartz, H.A., Giuseeppe, M, Frank, E., 2008. The use ofantidepressants in bipolar disorder. Journal of Clinical Psychiatry 69 (8),1307–1318.
Schurhoff, F., Bellivier, F., Jouvent, R., Mouren-Simeoni, M.C., Bouvard, M., Allilaire,J.F., Leboyer, M., 2000. Early and later onset bipolar disorders: two differentforms of manic depressive illness? Journal of Affective Disorders 58, 215–221.
Sheehan, D.V., 1983. The Anxiety Disease. Charles Scribner and Sons, New York.Sheehan, D.V., Raj, A.B., Sheehan, K.H., Soto, S., Knapp, E., 1993. The relative efficacy
of high-dose buspirone and alprazolam in the treatment of panic disorder:a double-blind placebo-controlled study. Acta Psychiatrica Scandinavica88 (1), 1–11.
Sheehan, D.V., Lecrubier, Y., Sheehan, K.H., Amorim, P., Janavs, J., Weiller, E., Hergueta,T., Baker, R., Dunbar, G., 1998. Mini-International Neuropsychiatric Interview(MINI): the development and validation of a structured diagnostic interview forDSM—IV and ICD—10. Journal of Clinical Psychiatry 59 (suppl 20), 22–33.
implementation of medication algorithms: update to the algorithms for
treatment of bipolar I disorder. Journal of Clinical Psychiatry 66 (7), 870–886.Thase, M., Macfadden, W., Weisler, R.H., Chang, W., Paulsson, B, Khan, A.,
Calabrese, J.R., for the Bolder II Study Group, 2006. Efficacy of quetiapine
monotherapy in bipolar I and II depression: a double-blind, placebo-controlled
study. Journal of Clinical Psychopharmacology 26 (6), 600–609.Tohen, M., Calabrese, J., Vieta, E, Bowden, C., Gonzalez-Pinto, A., Lin, D., Xu, W.,
Corya, S., 2007. Effect of comorbid anxiety on treatment response in bipolar
depression. Journal of Affective Disorders 104 (1–3), 137–146.Young, A., Seim, D., 2009. Review: long term use of antidepressants for bipolar
disorder reduces depressive episodes but increases risk of mania. Evidence-
Based Mental Health 12 (2), 49.Young, L.T., Cooke, R.G., Robb, J.C., Levitt, A.J., Joffe, R.T., 1993. Anxious and
nonanxious bipolar disorder. Journal of Affective Disorder 29 (1), 49–52.Young, R., Biggs, J., Ziegler, V., 1978. A rating scale for mania: reliability, validity,
and sensitivity. British Journal of Psychiatry 133, 429–435.
