Association of Major Depression and Binge Eating Disorder with Weight Loss in a Clinical Setting


Department of Medicine, Division of Preventive and Behavioral Medicine, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655. E-mail:


Objective: Obesity has been linked to both major depressive disorder (MDD) and binge eating disorder (BED) in clinical and epidemiological studies. The present study compared weight loss among patients with and without MDD and BED who participated in a hospital-based weight loss program modeled after the Diabetes Prevention Program.

Research Methods and Procedures: Of 131 obese patients who enrolled in treatment, 17% were diagnosed with MDD only, 13% were diagnosed with BED only, 17% were diagnosed with both MDD and BED, and 53% lacked either diagnosis in a pretreatment clinical interview.

Results: After treatment, patients with MDD only attained 63% of the weight loss that non-depressed patients attained. Patients with BED only attained 55% of the weight loss that non-binge eaters attained. The effect of MDD on weight loss was not accounted for by the presence of BED or vice versa. Only 27% of patients with both MDD and BED achieved clinically significant weight loss compared with 67% of patients who had neither disorder. Results were not significantly altered when gender, age, and diabetes status were adjusted.

Conclusion: Both MDD and BED were prevalent among this obese clinical population, and each disorder was independently associated with worse outcomes. Research is needed to investigate how to increase the efficacy of behavioral weight loss programs for individuals with MDD and/or BED.


Clinical trials have established the efficacy of behavioral weight loss interventions (1, 2). However, efficacy in clinical settings with patients who are not research volunteers is unknown. Patients in clinical settings have much higher comorbidity rates than research volunteers (3). Major depression disorder (MDD)1 and binge eating disorder (BED) are highly comorbid with obesity (32% and 20% to 40%, respectively) (4, 5, 6) but are often excluded from clinical trials. A better understanding of the impact of these comorbidities on weight loss outcomes is needed. The purpose of this study was to determine the rate of MDD and BED in a clinic sample, and the association of MDD and BED with weight loss outcome.

Research Methods and Procedures


Data from all patients (N = 131) who enrolled in the University of Massachusetts Memorial Medical Center Weight Center program from November 2004 to September 2006 were obtained from medical records retrospectively with permission from the Human Subjects Board. Patients (72% female; 92% white) had an average BMI of 43.08.


MDD and BED diagnoses were made via clinical interview (addressing DSM-IV criteria) by licensed psychologists before enrollment. Weight was measured on an electronic scale at baseline and treatment sessions.


Patients who were bulimic, suicidal, substance dependent, cognitively impaired, or psychotic were referred elsewhere for treatment. Patients meeting criteria for other Axis I disorders were entered into the program but received referrals to the University of Massachusetts Psychiatry Department. No tracking system was in place to assess how many referrals were followed.


The Diabetes Prevention Program (DPP) lifestyle intervention (7) is a 16-visit outpatient weight loss program with demonstrated efficacy on diabetes prevention (8) and cardiovascular disease (9). Goals of DPP include 7% weight loss, 500 to 1000 daily calorie reduction, and 150 minutes of moderate activity weekly. Several modifications to the DPP were made to increase feasibility in this setting as detailed elsewhere (10). Most importantly, the intervention was group-delivered instead of individual. In the present study, outcomes were assessed after the 16-session Core phase of the DPP intervention (7).

Statistical Analysis

ANOVA adjusted for age, gender, baseline BMI, and anti-depressant use, or χ2 tests were used.


Of 131 patients, 17% had MDD only, 13% had BED only, 17% had both MDD and BED, and 53% had neither diagnosis (Table 1). Individuals with MDD were more likely to have BED (57%) than those without MDD (21%) (χ2 = 17.29, p = 0.001).

Table 1. . Demographic and outcome variables
 No MDD (N = 87)MDD (N = 44)No BED (N = 88)BED (N = 43)
  • MDD, major depression disorder; BED, binge eating disorder; SD, standard deviation.

  • *

    p < 0.05.

% female68797174
Age (yrs) [mean (SD)]48.41 (11.08)49.88 (10.55)47.64 (10.85)51.48 (10.62)
Baseline BMI (kg/m2) [mean (SD)]43.02 (8.16)42.91 (7.17)43.38 (8.02)42.18 (7.40)
Baseline weight (kg) [mean (SD)]121.88 (28.93)117.37 (21.85)122.25 (29.11)116.50 (20.96)
% BED2157*0100
% MDD01002258*
On antidepressant2855*2854*
Weight change (%)[mean (SD)]−5.28 (3.29)−3.29 (3.63)*−5.35 (3.39)−3.10 (3.36)*
% losing ≥7% weight3816*3816*
Attendance [mean (SD)]13.54 (2.13)12.75 (2.55)13.53 (2.36)12.74 (2.10)
% dropout1025*1419

To allow for comparisons with DPP (7), the last value carried forward was used when end-of-treatment weight was missing. ANOVA on percentage weight change revealed a main effect of BED, F(7,122) = 7.04, p = 0.009, and MDD, F(7,122) = 4.10, p = 0.036, but no interaction between MDD and BED, F(7,122) = 0.58, p = 0.39 (Table 1). Baseline anti-depressant use was not significant in either model.

Analyses revealed that MDD patients were less likely to achieve the DPP weight loss goal (≥7%) than non-depressed patients (16% vs. 38%, χ2 = 6.68, p = 0.01), and BED patients were less likely to achieve the weight loss goal than non-BED patients (16% vs. 37%, χ2 = 6.13, p = 0.01). MDD patients were more likely to drop out than non-depressed patients (25% vs. 10%, χ2 = 4.85, p = 0.03, but drop-out rates were not different between BED and non-BED patients (19% vs. 14%, χ2 = 0.55, p = 0.45).


Both MDD and BED were prevalent in this clinical sample. Each was independently associated with worse weight loss outcomes, but having both disorders was not significantly worse than having one. MDD but not BED was associated with a higher dropout rate.

Our finding that MDD and BED are associated with less weight loss is consistent with past studies (11, 12, 13). The present study extends the literature by using a “real world” sample, examining patients with clinical levels of depression and binge eating, and by examining the compound effect of MDD and BED on weight loss.

Fewer patients (31%) in the present study met the 7% weight loss goal than in the original DPP (49%) (14)(Table 2). However, DPP participants had far lower rates of depression (10%) and BED (9%). When patients with MDD and BED are excluded from this sample, the percentage attaining the 7% weight loss goal is much closer to that reported in the DPP (42% vs. 49%), suggesting that DPP outcomes generalize to patients who do not have these disorders.

Table 2. . Percentage of patients reaching ≥7% weight loss goal
 DPP weight loss goal (≥7%) (%)
  1. DPP, Diabetes Prevention Program; UMMC, University of Massachusetts Memorial Medical Center; BED, binge eating disorder; MDD, major depression disorder.

 Total sample31
 BED only16
 MDD only16
 MDD + BED14
 Neither MDD or BED42
Original DPP49

This study has some limitations. While control groups were not used, the choice to forfeit a control group is consistent with the goal of examining a “real-world” sample. Second, MDD and BED diagnoses were derived by single raters and not assessed at end of treatment. Biological markers, diet, and physical activity measures would broaden the scope of findings but were unavailable. Long-term outcomes were not explored; thus, few conclusions can be drawn about weight regain. However, weight regain is a secondary issue in patients who have difficulty losing weight in the initial treatment stages. Finally, results may not generalize to other clinic samples, including those of lower socioeconomic status, where comorbidity rates may be even higher. The unique feature of this investigation is that outcomes are reported on all patients who enrolled in a treatment program in the first 2 years of its existence. Patients were paying consumers, not research volunteers and, thus, more closely represent the real-world clinic than trial samples.

Further investigation is needed to understand psychological disorders that are associated with obesity. Future studies using real world samples, standardized measures, and long-term outcomes would further elucidate the impact of psychological disorders on treatment outcome.


This study was supported by University of Massachusetts Medical School Grant K23HL073381 (to S.P.).


  • 1

    Nonstandard abbreviations: MDD, major depression disorder; BED, binge eating disorder; DPP, Diabetes Prevention Program.