SCIENTIFIC INVESTIGATIONS
Evaluation of Eszopiclone Discontinuation After Cotherapy With Fluoxetine for Insomnia With Coexisting Depression Andrew Krystal, M.D.1; Maurizio Fava, M.D.2; Robert Rubens, M.D.3; Thomas Wessel, M.D.3; Judy Caron, Ph.D.3; Phebe Wilson, M.S.3; Thomas Roth, Ph.D.4; W. Vaughn McCall, M.D., M.S.5 Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC; 2Massachusetts General Hospital, Boston, MA; Sepracor Inc., Marlborough, MA; 4Henry Ford Sleep Disorders Center, Detroit, MI; 5Wake Forest University Health Sciences, Winston-Salem, NC
1
3
Background: Insomnia and major depressive disorder (MDD) may coexist. This study evaluated hypnotic discontinuation effects following an 8-week placebo-controlled study of eszopiclone/fluoxetine cotherapy in patients with insomnia and comorbid MDD. Methods: Patients meeting DSM-IV criteria for MDD and insomnia received fluoxetine each morning for 8 weeks and were randomized to concomitant treatment with nightly eszopiclone 3mg (cotherapy) or placebo (monotherapy). Thereafter, patients received 2 weeks of continued fluoxetine plus single-blind placebo. Results: Incidence rates of central nervous system (CNS) and potentially CNS-related adverse events (AEs) during the run-out period were similar between treatment groups (8.8% with monotherapy vs 9.8% with cotherapy), and there was no evidence of benzodiazepine withdrawal AEs. Physician-assessed Clinical Global Impression improvements in depressive symptoms were maintained after eszopiclone discontinuation. Improvements in 17-item Hamilton-Depression Rating Scale (HAM-
D-17) scores with cotherapy versus monotherapy seen at Week 8 (p= .0004) were maintained at Week 10 (p< .0001) and significantly higher depression response and remission rates were observed after cotherapy at Week 10 (p< .02). Patients discontinued from eszopiclone maintained improvements in SL (sleep latency), WASO (wake after sleep onset), and TST (total sleep time) during the 2 weeks following discontinuation (p< .05). Conclusions: In this study, eszopiclone discontinuation did not result in significant CNS or benzodiazepine withdrawal AEs, rebound insomnia, or rebound depression; and improvements in sleep and depressive symptoms were maintained. Keywords: Insomnia, major depressive disorder, rebound, withdrawal, fluoxetine, eszopiclone Citation: Krystal A; Fava M; Rubens R et al. Evaluation of eszopiclone discontinuation after cotherapy with fluoxetine for insomnia with coexisting depression. J Clin Sleep Med 2007;3(1):48-55.
Disclosure Statement This was an industry supported study sponsored by Sepracor. All of the authors participated in the development of the manuscript including the analysis plan and interpretation of the results. One of the authors, Phebe Wilson, ran the statistical analysis. The investigators were the principal writers of the manuscript. Dr. Krystal has received research support from Cephalon, Cyberonics, GlaxoSmithKline, Merck & Co., Neurocrine Biosciences, Inc., Neuronetics, Organon, Pfizer, Respironics, Sanofi-Aventis, Sepracor, Astellas, Takeda Pharmaceuticals North America, and Somaxon; is a member of the speakers’ bureau for Cephalon, Cyberonics, GlaxoSmithKline, Mecta Corporation, Sanofi-Aventis, and Sepracor; and is a consultant for Cephalon, Johnson & Johnson, King Pharmaceuticals, Neurocrine Biosciences, Inc., Neurogen, Organon, Pfizer, Respironics, Sanofi-Aventis, Sepracor, Somaxon, Takeda Pharmaceuticals North America, Inc., Novartis, Astellas, Eli Lilly & Company, King, and TransOral. Dr. Fava has received research support from Abbott Laboratories, Alkermes, Lichtwer Pharma GmbH, Lorex Pharmaceuticals; as well as honoraria from Bayer AG, Biovail Pharmaceuticals, Inc., BrainCells, Inc., Compellis, Cypress Pharmaceuticals, Dov Pharmaceuticals, EPIX Pharmaceuticals, Fabre-Kramer Pharmaceuticals, Inc., Grunenthal GmbH, Janssen Pharmaceutica, Knoll Pharmaceutical Company, Lundbeck, MedAvante, Inc., Sepracor, Somerset Pharmaceuticals, Jazz Pharmaceuticals, Neuronetics, Nutrition 21, PharmaStar, and Somaxon; and has received both research support and honoraria from Aspect Medical Systems, Astra-Zeneca, Bristol-Myers Squibb Company, Cephalon, Eli Lilly & Company, Forest Pharmaceuticals Inc., GlaxoSmithKline, J & J Pharmaceuticals, Novartis, Organon Inc., PamLab, LLC, Pfizer Inc., Pharmavite, Roche, Sanofi/Synthelabo, Solvay Pharmaceuticals, Inc., and Wyeth-Ayerst Labo-
ratories; is on the speakers’ bureau of Astra-Zeneca, Boehringer-Ingelheim, Bristol-Myers-Squibb Company, Cephalon, Eli Lilly & Company, Forest Pharmaceuticals, Inc., GlaxoSmithKline, Novartis, Organon Inc., Pfizer Inc., PharmaStar, Wyeth-Ayerst Laboratories. Drs. Rubens, Wessel, Caron, and Wilson are employees of Sepracor. Dr. Roth has received research support from Aventis, Cephalon, GlaxoSmithKline, Neurocrine, Pfizer, Sanofi, Sepracor, Somaxon, Syrex, Takeda, and Xenoport; is a consultant for Accadia, Acoglix, Arena, AstraZeneca, Aventis, Cephalon, Eli Lilly, GlaxoSmithKline, Hypnion, King, Ludbeck, McNeil, Novadel, Merck, Neurocrine, Organon, Orginer, Proctor and Gamble, Pfizer, Roche, Sanofi, Sepracor, Shire, Somaxon, Syrex, Takeda, TransOral, Vanda, Vivometrics, Wyeth, Xenoport; and has participated in speaking engagements supported by Sanofi and Takeda. Dr. McCall has received research support from Sepracor, Takeda, Sanofi, GlaxoSmithKline, and Somaxon; is a consultant for King, Sepracor, and GlaxoSmithKline; and has received equipment and materials from Sepracor and Mini Mitter.
Journal of Clinical Sleep Medicine, Vol. 3, No. 1, 2007
Submitted for publication May 02, 2006 Accepted for publication August 15, 2006 Address correspondence to: Andrew Krystal, MD, Psychiatry and Behavioral Sciences, Trent Drive Room 54221, Durham, NC 27710, Tel: (919) 6818728; Fax: (919) 681-8744; Email:
[email protected]
I
nsomnia is a common health problem with considerable clinical and socioeconomic consequences. Insomnia may be unrecognized or overshadowed by comorbid disorders such as depres48
Eszopiclone Discontinuation in MDD With Insomnia
sion. Major depressive disorder (MDD) and chronic insomnia may coexist, and insomnia can be a prelude to the development of depression.1 Insomnia and depression have a complex and clinically important relationship,2 with 60%-90% of depressed patients having comorbid insomnia.3,4 Insomnia associated with depression can also persist after remission of the depressive symptoms and is the most common residual symptom in otherwise successfully treated cases of depression.5 Further, residual insomnia may predict an increased risk of relapse in MDD.2,6 Currently there is no consensus regarding use of hypnotic agents for independent treatment of insomnia in patients with MDD. Although the plurality of hypnotic use in the United States is in patients with psychiatric disorders,7 and, although many depressed outpatients on selective serotonin reuptake inhibitors (SSRIs) also receive a sleep aid,8 the first choice for treatment of insomnia in depressed patients is trazodone, not hypnotics.9 Many physicians are reluctant to treat insomnia associated with MDD with a hypnotic agent due to potential discontinuation effects including rebound insomnia and withdrawal.10 These effects might also hasten relapse or impede further antidepressant response. As previously reported,11 eszopiclone/fluoxetine co-administration led to significantly greater improvements in sleep, daytime function and depression than placebo/fluoxetine treatment across 8 weeks of cotherapy in patients with insomnia and MDD. Eszopiclone cotherapy was also associated with more rapid antidepressant effects. This report presents the 2-week single-blind placebo run-out phase of that study, which specifically assessed the effects of discontinuing eszopiclone on withdrawal-related adverse events, rebound insomnia, and depression.
ings, and a set of items rated on a Likert scale of 0-10: sleep quality, daytime alertness, ability to concentrate, physical well-being, and ability to function. Daytime alertness captured feelings of alertness during the day (ratings from very sleepy to wide awake and alert), while ability to concentrate captured the patient’s ability to concentrate or think clearly during the day (ratings from poor to excellent). Sense of physical well-being and ability to function were graded on a scale of poor to excellent. Patients with a minimum of three complete IVRS assessments received open-label fluoxetine 20 mg each morning and were randomly assigned to adjunctive treatment with either eszopiclone cotherapy (ESZ+FLX group) or fluoxetine monotherapy (PBO+FLX group), administered immediately before bed, for 8 weeks. At the end of 8 weeks, all patients continued open-label fluoxetine monotherapy treatment along with single-blind placebo (administered immediately before bed) for two weeks. For the last 3 weeks of the study (the last week of doubleblind treatment and the 2-week run-out period), patients called the IVRS daily. The morning-administered IVRS questionnaires were used to capture sleep and daytime function. To capture patient-reported depressive symptoms (including Bech and Maier subscales of HAM-D-17, data not presented), the Kellner questionnaire (data not presented), and a 26-item Daily Telephone Assessment (DTA) symptom inventory (data not presented) IVRS calls were completed once daily between Weeks 8 and 10 and on either the morning prior to or the morning of Week 10. Additionally, physician-administered HAM-D-17 was captured at Baseline and Weeks 4, 8 and 10; patients completed the Insomnia Severity Index (ISI) while at the clinic at Weeks 0 (baseline), 2, 4, 8, and 10; and physician-assessed Clinical Global Impression (CGI) were collected at baseline (except CGI-Improvement), every clinical visit during the double-blind period, at the end of the double-blind period, and Week 10. Safety measurements including vital signs, laboratory results, physical examinations, and electrocardiograms (ECGs) were assessed during the doubleblind period as well as at the end of the double-blind and run-out periods. Adverse events (AEs) reported during the single-blind run-out period were also collected.
METHODS The methods of this 10-week study have been fully described in the original publication.11 A synopsis of that information is presented below with details specific to the single-blind placebo/ fluoxetine run-out phase. Patients and Screening
Study Endpoints
All patients were required to be 21-64 years old (inclusive) and meet DSM-IV criteria for insomnia with comorbid MDD. The current depressive episode was required to have lasted 2 weeks to 6 months (inclusive), and the insomnia symptoms must not have predated the symptoms of MDD by more than 10 weeks. Additionally, patients were required to have a patient-reported score of ≥ 14 on the 17-item Hamilton Depression Rating Scale (HAM-D17)12 after subtracting the three sleep-related item scores and must also have reported total sleep time (TST) ≤ 6.5 hours, sleep latency (SL) ≥ 30 minutes, and wake time after sleep onset (WASO) ≥ 45 minutes per night at least three times a week for at least one month prior to screening. Finally, patients were required to either not be taking antidepressant medications at screening or to be taking a subtherapeutic antidepressant dose.
Withdrawal effects were evaluated by examining: 1) the prevalence of new or worsening central nervous system (CNS) or CNS-related adverse events, 2) rebound insomnia, and 3) worsening of MDD symptoms. Rebound insomnia variables assessed with IVRS during the run-out period included: 1) SL, 2) WASO, 3) TST, and 4) measures of daytime functioning (daytime alertness, concentration, physical well-being, and ability to function measured on 10-point scale, with higher scores indicating better next day functioning). Sleep and daytime function were additionally assessed subjectively with the ISI.13 The ISI is a validated scale consisting of seven 0-4 Likert self-ratings of difficulties with sleep onset, sleep maintenance, and early morning awakening, as well as daytime functioning, degree of impairment, concern, and satisfaction with current sleep pattern. The ISI total score ranges from 0 to 28. An ISI value ≤ 7 was used to indicate “no clinically meaningful insomnia,” as this range was shown in treatment studies to identify subjects who no longer met diagnostic criteria.14 A threshold of 8-14 has been suggested as a cutoff for “subclinical insomnia”
Study Procedures Patients meeting screening criteria at Visit 1 used Interactive Voice Response System (IVRS) daily for 3 to 7 days in order to assess baseline sleep and daytime function. These assessments consisted of estimates of the SL, WASO, TST, number of awakenJournal of Clinical Sleep Medicine, Vol. 3, No. 1, 2007
49
A Krystal, M Fava, R Rubens et al
of all screened patients were randomized and received treatment (275 to monotherapy and 270 to cotherapy), and, of the 387 subjects entering the single-blind run-out period, 373 completed the study. This report describes analyses performed in the single-blind run-out population. Thus, results reported here will differ from those reported in the original paper.11 No significant differences between treatment groups were observed for any demographic, sleep history, or efficacy parameter at baseline or at the start of the run-out period. Demographics and baseline characteristics of the run-out population were generally similar to those who discontinued prior to the single-blind run-out: more patients were Caucasian (65.4%), and female (67.7%), and the mean age was 41.9 years (versus 38.7 years in the double-blind population). Mean baseline (pre–double-blind treatment) insomnia characteristics of the original double-blind population and the run-out population were similar; WASO and SL were slightly higher in the singleblind run-out population compared with the double-blind population (92.0 vs 90.0 min and 202.6 vs 184.9 min, respectively), and baseline TST was slightly lower in the run-out population (220.8 vs 230.0 min). Mean baseline total ISI scores were equivalent (19.7 in each population).
Table 1—Frequency of CNS and Potentially CNS-Related AEs During the Two-week Single-Blind Placebo Run-out Period Event Headache Pain Accidental Injury Back Pain Nausea Dizziness Anxiety Insomnia Memory Impairment
PBO+FLX N=194 (%) 3.1 1.5 0.5 0 1.0 1.5 0.5 0.5 0.5
ESZ+FLX N=192 (%) 2.1 2.1 2.1 1.0 1.0 0.5 0.5 0.5 0
PBO indicates placebo; FLX, fluoxetine; ESZ, eszopiclone
because it identifies subjects who rate their symptoms as “below clinical threshold”. Several additional items were added to the ISI (referred to as the extended ISI) that are not included in the overall score. These items assess sleep quality, restedness upon arising, daytime fatigue, attention/concentration, and relationship and mood disturbances. Depressive symptoms were assessed during the run-out period (Week 8 through 10) with the HAM-D-17 (clinician-version) and the Bech and Maier subscales of the HAM-D-17 (both patient- and clinician-versions).15,16 The HAM-D-17 was also used to evaluate antidepressant response (defined as the percent of patients with at least 50% reduction in HAM-D-17 from baseline) and remission rates (defined as HAM-D-17 ≤ 7).17 Additionally, the investigators completed the CGI at Weeks 8 and 10. The CGI captured the investigator’s subjective assessment of severity (CGI-S) and improvement (CGI-I) in the subject’s depressive symptoms.
Safety Analysis of the safety data (clinical laboratory assessments, vital signs and ECGs) during the 2-week single-blind placebo run-out period revealed no systematic group differences. The incidence of AEs during the run-out period was 15.9% in the placebo/fluoxetine monotherapy group, and 20.3% in the eszopiclone cotherapy group; no AEs occurred at a rate ≥ 5% and there were no occurrences of abnormal dreams, hyperesthesia, neurosis, photosensitivity, or somnolence in either group. Incidence rates of CNS and potentially CNS-related AEs during this period were similar between treatment groups (Table 1). No AEs associated with benzodiazepine withdrawal were reported after eszopiclone discontinuation. Three serious AEs were reported during the run-out period; two in patients previously treated with placebo and fluoxetine (monotherapy) and one in a patient previously treated with eszopiclone and fluoxetine (cotherapy). One event of gastric hemorrhage in the monotherapy group was considered to be possibly related to study drug by the investigator, and the patient was discontinued. The other serious AE in the monotherapy group was an event of cholelithiasis with diarrhea, vomiting, and dizziness that was also considered possibly related to study drug; again, the patient was discontinued. The one serious AE in the cotherapy group was the development of suicidal ideation that occurred on Day 9 of singleblind run-out. This patient was hospitalized on the last day (Day 14) of the single-blind run-out, at which time the patient admitted to suicidal ideation 5 days previously. The event was considered to be possibly related to study drug by the investigator, and the event resolved without sequelae 8 days later. This was the only report of worsening depression or suicidal ideation in the cotherapy group during the study (there was one report of suicidal ideation in the monotherapy group during the double-blind period).
Statistical Analysis Analyses of the run-out period assessed data only from the subset of patients who entered the run-out phase of the study. All statistical tests were two-sided and were conducted at the 5% significance level. All continuous variables were compared across treatment groups using an analysis of covariance model (ANCOVA) to assess change from baseline (prior to double-blind treatment), with treatment and site as fixed effects and baseline as the covariate. Sleep latency and WASO endpoints were log-transformed prior to analysis. Categorical variables were compared using the Cochran-Mantel-Haenszel test for general association controlling for site. All efficacy parameters were summarized using descriptive statistics. Post hoc analyses included: assessments of rebound and withdrawal HAM-D-17 excluding insomnia items; 50% antidepressant response and remission (HAM-D-17) during the run-out period; HAM-D-17 scores at end of run-out in patients with more severe depression (HAM-D-17 > 22 at baseline); and the shift in the ISI from baseline to the end of the double-blind period and end of run-out period. RESULTS
Depression Results
Patient Disposition
Relative to the fluoxetine monotherapy group, HAM-D-17 scores with and without insomnia items remained significantly improved from baseline in the eszopiclone cotherapy group at
A total of 985 patients were screened for study inclusion; 545 Journal of Clinical Sleep Medicine, Vol. 3, No. 1, 2007
50
Eszopiclone Discontinuation in MDD With Insomnia
Figure 1—HAM-D-17 Change from Baseline at Weeks 8 and 10. Note: p-values reflect results from change from baseline analyses using ANCOVA.
Weeks 8 and 10 (p
