Duloxetine in the treatment of binge eating disorder with depressive disorders: A placebo-controlled trial


  • Supported in part by a grant from Eli Lilly.

  • Paul Keck Jr, MD, is employed by the University of Cincinnati College of Medicine, University of Cincinnati Physicians, and the Lindner Center of HOPE; Dr. Keck is presently or has been in the past year a principal or co-investigator on research studies sponsored by Alkermes, AstraZeneca, Cephalon, GlaxoSmithKline, Eli Lilly and Company, Epi-Q, Jazz Pharmaceuticals, Marriott Foundation, National Institute of Mental Health (NIMH), Orexigen, Pfizer, Shire; Dr. Keck has been reimbursed for consulting to, in the past year: 2010: Sepracor, Medco, 2009: BristolMyersSquibb, GlaxoSmithKline, Pfizer, QuantiaMD, Schering Plough, Patents: Dr. Paul E. Keck Jr. is a co-inventor on United States Patent No. 6,387,956: Shapira NA, Goldsmith TD, Keck, PE Jr. (University of Cincinnati) Methods of treating obsessive-compulsive spectrum disorder comprises the step of administering an effective amount of tramadol to an individual. Filed March 25, 1999; approved May 14, 2002; Dr. Keck has received no financial gain from this patent. Susan L. McElroy, MD, is employed by the University of Cincinnati College of Medicine, University of Cincinnati Physicians, and the Lindner Center of HOPE; Dr. McElroy is a consultant to, or member of the scientific advisory boards of Alkermes, Eli Lilly and Company, Shire; Dr. McElroy is a principal or co-investigator on research studies sponsored by Agency for Healthcare Research & Quality (AHRQ), Alkermes, AstraZeneca, Bristol-Myers Squibb, Cephalon, Eli Lilly and Company, Forest Labs, GalaxoSmith Kline, Jazz Pharmaceuticals, Marriott Foundation, National Institute of Mental Health, Orexigen Therapeutics, Pfizer Shire, Takeda Pharmaceutical Company Limited; Patents: Dr. Susan L. McElroy is also inventor on United States Patent No. 6,323,236 B2, Use of Sulfamate Derivatives for Treating Impulse Control Disorders, and, along with the patent's assignee, University of Cincinnati, Cincinnati, OH, has received payments from Johnson & Johnson Pharmaceutical Research & Development, L.L.C., which has exclusive rights under the patent. James I. Hudson, MD, has been a consultant for Eli Lilly, Pfizer, and Alkermes; and has received grant support from Eli Lilly, Otsuka, and Ortho-McNeil Janssen Scientific Affairs. Erik B. Nelson, MD, has received funding for clinical trials with Astra Zeneca, Novartis and Pamlab. Drs. Erin L. Winstanley and Anna I. Guerdjikova and Mrs. Nicole Mori and Jessica McCoy have no industry/commercial support or conflicts of interest to disclose.



This study evaluated duloxetine in the treatment of binge eating disorder (BED) with comorbid current depressive disorders.


In this 12-week, double-blind, placebo-controlled trial, 40 patients with Diagnostic and Statistical Manual of Mental Disorders-IV-TR BED and a comorbid current depressive disorder received duloxetine (N = 20) or placebo (N = 20). The primary outcome measure was weekly binge eating day frequency.


In the primary analysis, duloxetine (mean 78.7 mg/day) was superior to placebo in reducing weekly frequency of binge eating days (p = .04), binge eating episodes (p = .02), weight (p = .04), and Clinical Global Impression-Severity of Illness ratings for binge eating (p = .02) and depressive disorders (p = .01). Changes in body mass index and measures of eating pathology, depression, and anxiety did not differ between the two groups.


Duloxetine may be effective for reducing binge eating, weight, and global severity of illness in BED with a comorbid current depressive disorder, but this finding needs confirmation in larger, placebo-controlled trials. © 2011 by Wiley Periodicals, Inc. (Int J Eat Disord 2012)


Binge eating disorder (BED) is characterized by recurrent, distressing episodes of binge eating without the inappropriate compensatory weight loss behaviors of bulimia nervosa.1 Though currently listed in the Diagnostic and Statistical Manual of Mental Disorders-IV-TR (DSM-IV-TR)1 appendix as a disorder provided for further study, it has been recommended that BED be formally included as an eating disorder in DSM-5.2, 3 It is common, with a lifetime prevalence of about 2–3%, often chronic, and associated with psychopathology, obesity, reduced quality of life, and disability.2, 4–6

BED frequently co-occurs with depressive disorders, including major depressive disorder and dysthymic disorder.2, 4, 6 BED patients with comorbid psychiatric disorders, including depressive disorders, may have a more severe illness than patients without comorbid psychiatric disorders, characterized by earlier age at first diet, higher lifetime body mass index (BMI), greater distress, less dietary restraint, lower self-esteem, more frequent binge eating, higher levels of negative affect, more frequent trauma and abuse histories, less successful weight loss, and overall poorer treatment outcomes.7–9

No randomized, controlled study has specifically addressed the treatment of BED patients with a co-occurring depressive disorder.10 Novel treatments that reduce binge eating, depressive symptoms, and body weight in this patient population that are also well tolerated are needed.

Duloxetine is a selective serotonin norepinephrine reuptake inhibitor (SNRI) approved by the U.S. Food and Drug Administration (FDA) for the treatment of major depressive disorder.11 Several lines of evidence suggested it might be a useful treatment for BED when comorbid with depressive disorders. First, an open-label study12 and several case reports13–16 found that duloxetine decreased binge eating in BED and the related condition bulimia nervosa. Similarly, open-label studies reported the successful treatment of BED and bulimia nervosa with the SNRIs milnacipran17, 18 and venlafaxine.19 Second, the SNRI sibutramine, a weight loss agent withdrawn from the market in 2010 for safety concerns,20 has been shown efficacious in three placebo-controlled studies in reducing binge eating behavior and excessive body weight in BED.21–23 In one of these studies, it was also effective in reducing depressive symptoms.21 Indeed, a range of antidepressants with various mechanism of action has been reported to reduce binge eating in BED and bulimia nervosa, including serotonin selective reuptake inhibitors (SSRIs) on the one hand and tricyclics with primarily noradrenergic reuptake inhibiting properties on the other.10 As noted earlier, duloxetine possesses both of these mechanisms. Third, duloxetine is generally well tolerated.11, 24 In particular, it has not been associated with the adverse cardiovascular events triggering sibutramine's withdrawal from the market.20

These observations led us to hypothesize that duloxetine would decrease binge eating, as well as excessive body weight and depressive symptoms, in BED patients with comorbid depressive disorders. To test this hypothesis we conducted a randomized, parallel-group, placebo-controlled, double-blind, flexible-dose (30–120 mg/day) study to assess the efficacy and safety of duloxetine during a 12-week course of treatment in 40 patients with BED and a comorbid current depressive disorder.



Study participants were recruited by newspaper advertisements requesting volunteers for a study of an investigational medication for binge eating and depression. Individuals were enrolled into the study if they: (1) were between 18 and 65 years of age (inclusive); (2) met DSM-IV-TR criteria for BED; (3) met DSM-IV-TR criteria for a current depressive disorder (i.e., major depressive disorder, single episode or recurrent; dysthymic disorder; or depressive disorder not otherwise specified (NOS)] for a duration of at least 1 month, including the time preceding and during the screening period; (4) displayed ≥2 binge days/week (days when the participant had one or more binge eating episodes) prospectively determined by take-home binge diaries for at least 1 week before randomization; and (5) had scores of ≥25 on the clinician-rated version of the Inventory of Depressive Symptoms Scale (IDS-C)25 at screening and baseline visits.

Individuals were excluded from participation if they: (1) displayed a significant risk for suicide; (2) were receiving psychotherapy from a mental health professional for treatment of BED or depression within 3 months before randomization; (3) had a DSM-IV-TR diagnosis of alcohol or substance abuse or dependence, bulimia nervosa, or anorexia nervosa within 6 months before randomization; (4) had a lifetime history of a psychotic disorder, a bipolar disorder, or dementia; (5) had an Axis II disorder which might interfere with study procedures; (6) had clinically unstable medical disease; (7) had a history of seizures, including febrile seizures in childhood; (8) had a known hypersensitivity to duloxetine or any of its inactive ingredients; or (9) were receiving monoamine oxidase inhibitors, tricyclics, antipsychotics, lithium, or fluoxetine within 4 weeks before randomization. Women were excluded if they were pregnant, lactating, or if fertile, not practicing a medically accepted form of contraception.

The Institutional Review Board at the University of Cincinnati Medical Center approved the study protocol, and the study was conducted in compliance with the Declaration of Helsinki. All participants signed approved written informed consent forms after the study procedures had been fully explained and before any study procedures were performed. Participants were enrolled from March 2007 through August 2009.

Study Design

This was a 12-week outpatient, randomized, placebo-controlled, double-blind, parallel-group, flexible-dose study conducted at the University of Cincinnati Medical Center and the Lindner Center of HOPE. The trial consisted of two phases: a 1- to 2-week screening period during which participants had to display ≥2 binge days/week to be included in the study; and a 12-week double-blind treatment period. Participants were evaluated at least twice during the screening period (the screening and baseline visits) and after 1, 2, 4, 6, 8, 10, and 12 weeks during the treatment period.

The screening evaluation included the structured clinical interview for DSM-IV-TR (SCID)26 to establish BED, depressive disorder, and other comorbid axis I diagnoses; the eating disorders examination-questionnaire (EDE-Q)27 to confirm the diagnosis of BED; a physical examination; vital signs; height and weight measurements to determine BMI; an electrocardiogram (ECG); routine blood chemical and hematological tests; and urinalysis. At this evaluation and each of the following visits, participants were given take-home diaries in which to record any binge eating episodes. At the last visit of the screening period (the baseline assessment), participants continuing to meet entry criteria were enrolled in the treatment period and randomly assigned in a 1:1 ratio to therapy with duloxetine or placebo according to computer-generated coding. Randomization was balanced by use of permuted blocks. Allocation concealment was achieved by having the research pharmacy perform the randomization, package the study medication, and maintain the integrity of the blinded information throughout the trial. At each visit following the baseline visit, participants were assessed for number of binge days and binge eating episodes experienced since the last visit; medication compliance ascertained by capsule count; adverse events; use of non-study medications; vital signs; and weight. Outcome rating scales were also performed.

All study medication was in identical 30 mg capsules supplied in numbered containers and dispensed to participants according to a predetermined randomization schedule. Study medication was begun at 30 mg/day for the first 7 days. At the beginning of the second week of treatment, study medication was increased, as tolerated, to 60 mg/day and this dose was kept constant for two weeks. In the absence of remission of binge eating or depressive symptoms and intolerable side effects, the dose could be further increased as following: 90 mg/day at the beginning of the forth treatment week and 120 mg/day at the beginning of the sixth treatment week. Dosing was either once per day or twice per day depending on tolerability.

The intent-to-treat (ITT) population for the primary longitudinal analysis included all randomized participants, and for the secondary endpoint analysis all participants who had at least one post-baseline visit. The safety population included all participants who received at least one dose of the study medication.

Outcome Measures

The primary outcome measure was the weekly frequency of binge eating days (binge day frequency), defined as the mean number of binge days (days when the participant had one or more binge eating episodes) per week in the interval between visits (total number of binge days in the interval divided by number of days in the interval, then multiplied by 7). Binge eating episodes were defined using DSM-IV-TR criteria1 and assessed via clinical interview and review of participant take-home diaries, upon which participants recorded number of binge episodes. Secondary outcome measures were weekly frequency of binge episodes, weight (kg); BMI (calculated by dividing body weight in kg by height in m2); Clinical global impression-severity (CGI-S) and improvement scale (CGI-I)28 for binge eating (CGI-S-BE and CGI-I-BE) and depressive disorder (CGI-S-DD and CGI-I-DD) symptomatology scores; Yale-brown obsessive-compulsive scale modified for binge eating scale (YBOCS-BE)29 scores; Three factor eating questionnaire (TFEQ)30 scores; Hamilton anxiety scale (HAM-A)31 total scores; and the clinician-rated IDS-C25 total scores. All scales were administered at all post-baseline visits except for the TFEQ and HAM-A, which were administered at baseline and at weeks 2, 6, 10, and 12.

Other secondary outcome measures included categorical response categories for binge eating and depressive symptoms based on change from baseline to endpoint. For binge eating episodes, these were defined as follows: remission = cessation of binges; marked = 75–99% decrease; moderate = 50–74% decrease; and none = less than 50% decrease. For depression, response was defined as a decrease in IDS score ≥50% from baseline to final study visit; remission was defined as an endpoint IDS score <12. The following safety measures were assessed: adverse events, clinical laboratory data, physical examination findings, and vital signs.

Statistical Methods

The baseline characteristics of each group were compared by using chi-square or Fisher's exact tests for categorical variables and independent-sample t-tests for continuous variables.

The primary efficacy analysis was a longitudinal analysis comparing the rate of change of binge day frequency during the treatment period between groups. The same analysis was applied to binge episode frequency, weight, BMI, and scores on the CGI-Severity, YBOCS-BE, and IDS scales. The difference in rate of change was estimated by random regression methods, as described in the study by Fitzmaurice et al.32 and Gibbons et al.33 and as used in a number of pharmacotherapy studies of BED.34–41 We used a model for the mean of the outcome variable that included terms for treatment, time, and treatment-by-time interaction. Time was modeled as a continuous variable, expressed as the square root of days since randomization (baseline) unless there was evidence for a better fit with another functional form. For the analyses of binge frequency and binge day frequency, we used the logarithmic transformations log [(binges/week) + 1] and log [(binge days/week +1)], respectively, to normalize the data and stabilize the variance. The longitudinal analyses included all participants. First-order antedependence, heterogeneous autoregressive, or autoregressive covariance structures were compared for each model and the best fitting model, with the lowest AIC value, was retained.

Several secondary analyses were performed using the ITT sample. Using the last observation carried forward (LOCF), baseline to endpoint change scores were computed for each measure (on the logarithmic scale for the binging measures) and independent-samples t-tests were used to compare these changes between the treatment groups. An extension of the Wilcoxon rank-sum test (“nptrend” in Stata) was used to analyze categorical response to treatment (as defined above) for the ITT group. Time to recovery (defined as the first four consecutive binge-free weeks after baseline) was analyzed with a log-rank test of survival function for the ITT population. For laboratory measures, including weight, the mean difference between endpoint and baseline measures was computed for each treatment group and then compared using the t-test.

SAS software (version 9.1, Cary, N.C.) was used only for the primary analysis. Other analyses were conducted using Stata software (version 10.2, College Station, TX). All statistical tests and confidence intervals were two-sided, α = 0.05.


Of 64 individuals screened, 40 met entry criteria and were randomized to duloxetine (N = 20) or placebo (N = 20) (see Fig.1). Of these 40 participants, 35 (88%) were women and 33 (83%) were Caucasian; the mean (SD) baseline weekly binge day frequency of the group was 4.0 (1.7); and the mean (SD) baseline BMI of the group was 40.6 (7.4) (Table1). The most prevalent depressive disorder diagnosis was recurrent major depressive disorder (63%). The mean (SD) IDS score was 35.5 (6.7), indicating a moderate to severe level of depressive symptoms.42 There were no significant differences between the treatment groups in demographic or clinical variables at baseline, except participants receiving duloxetine were older (p = .02).

Figure 1.

Participants who entered a 12-week placebo-controlled trail of duloxetine in binge eating disorder with comorbid depressive disorders. (a) Intent to treat (ITT) defined as all randomized participants; (b) Safety population defined as all randomized participants who received atleast one dose of study medication; (c) Inadequate protocol adherence = inadequate compliance with study medication (N = 1), use of disallowed medication (hydrocodone, N = 1), and refusal of participant follow nonstudy physician's medical recommendations (N = 1).

Table 1. Baseline demographic and clinical characteristics of participants with binge eating disorder and depressive disorders randomly assigned to 12 weeks of double-blind treatment with duloxetine or placebo
Characteristic/VariableTotal Group (N = 40)Duloxetine (N = 20)Placebo(N = 20)p
Demographic characteristics
 Female, N (%)35(88)16 (80)19 (95)0.34
 Caucasian, N (%)33(83)18(90.0)15(65)0.41
 Age (years), Mean (SD)40.1(12.0)44.4(12.1)35.7(10.4)0.02
Diagnostic characteristics
 Eating disorders examination-questionnaire
  Mean (SD)3.7(0.9)3.7(1.0)3.7(0.9)0.99
 Depressive disorder diagnosis, N (%):    
  Major depressive disorder, Recurrent25(63)10(50)15(75)0.19
  Major depressive disorder, Single episode9(23)5(25)4(20)1.0
  Dysthymic disorder5(13)4(20)1(5)0.34
 Lifetime anxiety Disorders, N (%):12(30)6(30)6(30)1.0
 Lifetime substance use disorders N, (%):5(13)4(20)1(5)0.34
Primary outcome variable
 Binge day frequency (per week), Mean (SD)4(1.7)4.3(1.7)3.8(1.7)0.36
Secondary outcome variables
 Binge episode frequency (per week), Mean (SD)4.5(2.3)4.7(1.9)4.2(2.6)0.49
 Weight (kg), Mean (SD)114.7(23.6)111.1(24.1)118.3.7(23.1)0.34
 Body mass index, Mean (SD)40.6(7.4)38.7(6.8)42.8(7.7)0.09
Clinical global impression—severity of illness scale for binge eating, Mean (SD)4.8(0.7)5.0(0.8)4.6(0.7)0.09
 Clinical global impression—Severity of Illness Scale for depressive disorders, Mean (SD)4.2(0.7)4.3(0.7)4.2(0.8)0.50
 Yale-brown obsessive—compulsive scale (modified for binge eating), Mean (SD)21.9(3.1)22.3(3.5)21.6(2.8)0.52
 Inventory of depressive symptoms Mean (SD)35.5(6.7)35.6(7.9)35.4(5.4)0.93
 Three-factor eating questionnaire mean (SD)
  Cognitive restraint subscale5.2(3.8)6(3.3)4.4 (4.3)0.24
  Disinhibition Subscale13.6(1.8)13.6(2.0)13.7(1.7)0.97
  Hunger subscale11.5(2.4)10.9(2.7)12.1(2.0)0.16
 Hamilton anxiety scale, Mean (SD)16.5(7.5)16.9(9.1)16.2(5.7)0.79

Seven (35%) of the 20 participants in the duloxetine group discontinued prematurely (adverse events, N = 3 scheduling problems, N = 1; non-adherence with the protocol, N = 3) and six (30%) of the 20 participants in the placebo group discontinued prematurely (lack of efficacy, N = 1; lost to follow-up, N = 5) (see Fig.1). The remaining 27 (68%) participants completed the 12 weeks of treatment (N = 13 receiving duloxetine and N = 14 receiving placebo). The mean (SD) daily dose of duloxetine at endpoint evaluation for the 18 participants receiving drug was 78.7 (19.6) mg; for those receiving placebo it was 80.3 (25.2) mg.

The mean frequency of binge eating days per week decreased over the study period in both treatment groups, but more so in the duloxetine group (see Fig.2). The primary efficacy analysis using random regression showed that participants receiving duloxetine had a significantly greater reduction in binge day frequency than participants receiving placebo (p = .04) (Table2). Duloxetine was also associated with a significantly greater improvement than placebo for binge episode frequency, body weight, and scores on the CGI-S-BE and CGI-S-DD scales. The associated standardized effect sizes (Cohen's d) were 0.67, 0.78, 0.66, 0.77, and 0.82 respectively; these are all either medium (0.50–0.79) or large (>0.80) range. However, there were no significant differences between the treatment groups in the change in BMI or scores on the YBOCS-BE or IDS-C scales. In the secondary analysis of baseline-to-endpoint change scores, there were no significant differences between groups in change in binge day frequency or in the secondary outcome measures (except for CGI-S-DD scale scores) (Table2).

Figure 2.

Mean (± SD) weekly binge day frequency over 12 weeks of treatment in participants with binge eating disorder and depressive disorders randomly assigned to duloxetine or placebo (error bars are 95% confidence intervals).

Table 2. Outcome measures before and after 12 weeks of treatment with duloxetine versus placebo and analysis of change in outcome
Outcome measureMean (SD)Estimated Difference Between Groups at 12 Weeks
Baselinea (N = 40)Last Observationb (N = 38)Longitudinal Analysisc (N = 40)Endpoint Analysisd (N = 38)
Placebo (N = 20)Duloxetine (N = 20)Placebo (N = 20)Duloxetine (N = 18)Estimate [95% CI]pdEstimate [95% CI]pd
  • a

    Baseline value for the entire sample.

  • b

    Last observation for the ITT sample.

  • c

    Longitudinal analysis included the entire sample.

  • d

    Last Observation Endpoint was defined using last observation carried forward for the ITT sample.

  • e

    Measured at baseline and weeks 2, 6, 10, and 12 only.

Binge day frequency per week3.5 (1.5)4.0 (1.8)1.3 (1.2)1.0 (1.7)−0.77 [−1.00, −0.17]0.040.670.78 [−0.29, 1.86].150.47
Binge episode frequency per week4.0 (2.4)4.5 (2.0)1.3 (1.2)1.1 (2.0)−0.62 [−0.89, −0.03]0.020.780.79 [−0.65, 2.22].270.36
Weight (kg)118.3 (23.1)111.1 (24.1)118.0 (23.2)108.3 (23.8)−2.91 [−5.74, −0.09]0.040.662.86 [−0.21, 5.94].070.59
Body mass index, kg/m242.8 (7.6)38.7 (6.8)42.9 (7.3)37.7 (7.5)−0.83 [−1.84, 0.18]0.110.530.85 [−0.11, 1.81].080.58
Clinical global impression-severity of illness scale for binge eating4.6 (0.7)5.0 (0.8)2.7 (1.3)2.3 (1.5)−1.23 [−4.13, 0.03]0.020.770.76 [−].100.53
Clinical global impression -improvement scale for binge eating2.2 (1.3)1.7 (1.0)   0.28 [−0.50, 1.07].470.24
Clinical global impression-severity of illness scale for depressive disorders4.2 (0.7)4.3 (0.7)2.9 (1.0)2.3 (1.3)−0.79 [−2.49, −0.04]0.010.820.81 [0.07, 1.56].030.68
Clinical global impression—improvement scale for depressive disorders2.4 (1.4)1.7 (1.1)   0.48 [−0.43, 1.39].300.34
Yale-brown obsessive-compulsive scale (modified for binge eating)          
 Total21.6 (2.8)22.3 (3.5)10.3 (6.3)9.4 (7.0)−9.05 [−53.99, 1.77]0.170.441.54 [−2.78, 5.85].470.24
 Obsessions10.7 (1.4)11.0 (2.6)5.5 (3.1)5.4 (3.6)−1.98 [−13.64, 0.78]0.230.390.24 [−2.02, 2.50].830.07
 Compulsions11.0 (2.0)11.3 (2.1)4.8 (3.4)3.9 (4.0)−2.76 [−17.76, 0.79]0.200.421.30 [−1.20, 3.80].300.34
Inventory of depressive symptoms35.4 (5.4)35.6 (7.9)21.6 (12.7)19.1 (11.5)−2.70 [−56.91, 18.12]0.580.183.65 [−3.65, 10.95].320.33
Three-factor eating questionnairee          
 Cognitive restraint4.4 (4.3)6.0 (3.2)7.1 (4.8)5.6 (3.3)   1.42 [−1.29, 4.13].290.39
 Disinhibition13.6 (1.7)13.6 (2.0)11.5 (3.3)11.3 (4.3)   0.06 [−2.46, 2.59].960.02
 Hunger12.1 (2.0)11.0 (2.7)9.5 (3.9)8.7 (3.5)   −0.91 [−3.78, 1.96].520.24
Hamilton Anxiety Scalee16.2 (5.7)16.9 (9.1)7.2 (6.5)9.6 (9.0)   −1.22 [−5.93, 3.49].600.17

Regarding global impression of improvement, there were no significant differences between groups on any measure. Fourteen (78%) participants in the duloxetine group and 11 (55%) participants in the placebo group had final scores of “much” or “very much” improved on both the CGI-I-BE and CGI-I-DD scales.

Regarding categorical response analyses, there was a numerically but not statistically significantly higher level of remission of BED for duloxetine-treated participants (56%) compared with placebo-treated participants (30%) (Table3). Numerically similar proportions of participants attained response and remission of depressive symptoms in favor of duloxetine (Table3). Four duloxetine recipients achieved remission in both binge eating and depression compared with two placebo recipients. Duloxetine was not associated with shortened time to recovery of binge eating in the ITT group [hazard ratio (95% confidence interval) for recovery = 1.7 (0.67, 4.07), p = .27].

Table 3. Categorical response to treatment among participants with binge eating disorder and depressive disorders randomly assigned to 12 weeks of double-blind treatment with duloxetine or placebo
 Placebo (N = 20) N (%)Duloxetine (N = 18) N (%)p valuea
  • a

    Chi-square, fishers exact or nptrend were used to determine a statistical difference between groups.

  • b

    Percent reduction in binge eating episodes from baseline to endpoint.

  • c

    50% or greater reduction in Inventory of Depressive Symptoms Score from baseline to endpoint.

  • d

    Endpoint Inventory of Depressive Symptoms Score <12.

Response of binge eating episodesb   
 None (<50%)6 (30)3 (17) 
 Moderate (50–74%)4 (20)1 (6) 
 Marked (75–99%)4 (20)4 (22) 
 Remission (100%)6 (30)10 (56)0.09
Response of depressionc8 (40)11 (61)0.19
Remission of depressiond4 (20)5 (28)0.71

Participants receiving duloxetine experienced a mean (SD) weight loss of 3.2 (6.4) kg from baseline to endpoint, whereas those receiving placebo experienced a mean (SD) weight loss of 0.30 (2.2) kg (p = .07).

The most common adverse events reported by duloxetine-treated participants were nausea (N = 9), dry mouth (N = 7), constipation (N = 5), and hyperhydrosis (N = 5). There were no statistically significant differences between treatment groups in the incidence of any individual event. More participants discontinued duloxetine (N = 3; 15%) for adverse events than placebo (N = 0) (Fisher exact p = .19). Those events were severe gastrointestinal problems and a sinus infection (N = 1), insomnia (N = 1), and confusion (N = 1). The gastrointestinal problems and sinus infection was classified as a serious adverse event as the participant required a 48-hour hospitalization; the participant recovered fully and the event was thought not to be due to duloxetine.

There were no clinically significant changes in physical examination findings, vital signs, or clinical laboratory values. There was no evidence of withdrawal symptoms in the seven participants who discontinued duloxetine prematurely or in the 13 participants who discontinued duloxetine per protocol.


In the primary longitudinal analysis of this 12-week randomized, placebo-controlled study of BED with a comorbid current depressive disorder, duloxetine (mean final dose 78.7 mg/day) was significantly superior to placebo in reducing binge day and binge episode frequency, body weight, and overall severity of illness for binge eating and depressive disorders, with the associated effect sizes being medium to large. The mean weight loss in the group receiving duloxetine was 3.4 kg, compared with 0.3 kg in the group receiving placebo. Taken together these findings suggest duloxetine might be efficacious in the treatment of BED when comorbid with a current depressive disorder, which may be less responsive to treatment than BED without a comorbid psychiatric disorder.7, 8

The potential mechanism of action of duloxetine in BED is unknown. As has been hypothesized for sibutramine, duloxetine might reduce binge eating by modifying internal signals controlling hunger and satiety through its combined effects on serotonergic and noradrenergic transmission.21–23 Reduced binge eating, in turn, might lead to weight loss. In addition, several preliminary reports have suggested that duloxetine may have beneficial effects in substance use disorders.43–45 Eating disorders, including BED, have been hypothesized to be related to addictive disorders, as they co-occur with one another and highly palatable food and drugs of abuse compete for the same brain reward circuitry.46, 47 Moreover, the dopamine-mediated reward circuits implicated in addictive disorders may also be compromised in major depressive disorder.48 Thus duloxetine might exert its beneficial effects in BED with a comorbid depressive disorder by affecting this reward system.

This study has several important limitations. First, the small sample size may have compromised the ability of the study to detect important treatment effects. It may have been under-powered to detect clinically important differences in the end-point analysis or in secondary outcome measures, such as BMI and IDS-C scale scores. Second, the attrition rate was high, with 33% of participants withdrawing before study completion, rendering the results heavily dependent on assumptions regarding missing data. However, the primary analysis was based on data from all randomly assigned participants at all time points. Moreover, the high attrition rate found in this study is consistent with attrition rates in other placebo-controlled pharmacotherapy studies in BED.34–41

A third limitation is that because the study group was primarily female and Caucasian, and the duration of treatment was short (12 weeks), the results may not generalize to males, to other racial groups, or to longer treatment periods. Fourth, because individuals with substance use disorders, severe personality disorders, and unstable medical disorders were excluded, the results may not generalize to BED and depressive disorders when also co-occurring with these conditions.

In summary, in a 12-week trial in individuals with BED and comorbid depressive disorders, duloxetine was superior to placebo in reducing binge frequency, body weight, and global severity for binge eating and depressive disorders. It was associated with fairly good tolerability, but a high discontinuation rate. Controlled trials of duloxetine and other SNRIs in larger groups of participants with BED and comorbid depressive disorders appear warranted.

Earn CE credit for this article!

Visit: http://www.ce-credit.com for additional information. There may be a delay in the posting of the article, so continue to check back and look for the section on Eating Disorders. Additional information about the program is available at www.aedweb.org