Introduction

Binge Eating Disorder (BED) is an eating disorder characterized by binge eating episodes where there is a sense of loss of control and no compensatory behaviors following the binge episodes. Up until 2013 BED was included in the Appendix B of the Diagnostic and Statistical Manual of the American Psychiatric Association [1] among the pathologies needing further studies.

Following the DSM-IV-TR, BED was characterized by recurrent episodes of binge eating associated with at least three of the following behavioral indicators of loss of control, which are the core characteristics of the syndrome: eating until feeling uncomfortably full; eating large amounts of food when not feeling physically hungry; eating much more rapidly than usual; eating alone because of being embarrassed by how much one is eating; feeling disgusted, depressed, or guilty, after overeating. Moreover, binge eating was not associated with the regular use of compensatory behaviors (vomiting, use of laxatives, excessive physical exercise, etc.) or to occur when the patient was affected either by bulimia nervosa or anorexia nervosa.

In 2013 the American Psychiatric Association reorganized its classification system in the fifth edition of the DSM. Given its clinical relevance [2] BED was elevated from a provisional research diagnosis to a formal diagnostic category, changing its diagnostic criteria. The new, less restrictive, criteria request can be diagnosed as affected by BED the patients who have one binge episode a week for at least 3 months.

Since when the proposed DSM-5 criteria was posted for public comment in 2010, the researchers tested if the use of less restrictive criteria would lead to an increased prevalence of the syndrome [3]. Allen Frances hypothesized that the use of less restrictive criteria would lead to an increased prevalence of BED [4], potentially creating false epidemics of “misidentified pseudo-patients” (page 122). In his opinion BED diagnosis using the DSM-5 criteria had to be considered a fake eating disorder (page 183), representing nothing more than “gluttony” (page 184). This hypothesis was not confirmed in the general population [5–7]. Lifetime prevalence of BED range between 0.17 % [5] and 8.8 % [8] using DSM-IV criteria and from 0.2 % [5] to 3.6 % [6] using the proposed DSM-5 criteria. In a sample of 888 participants, the increase in the lifetime prevalence of BED using the proposed DSM-5 criteria relative to the DSM-IV TR criteria was 2.9 % in women and 3.0 % men and 7.7 % in women and 0 % in men for the point prevalence [6].

However, the debate on the prevalence of BED following the DSM-5 criteria among severe obese patients is still open [3, 9]. In a recent study, our group found that in a sample of patients seeking for bariatric surgery the prevalence of BED increased from 20.33 % using the DSM-IV-TR criteria to 48.3 % following the DSM-5 criteria [10].

The psychopathology and psychiatric comorbidity of BED is well known. The DSM-5 [11] states that patients with BED have greater functional impairment, lower quality of life, more subjective distress, and greater psychiatric comorbidity: according to Kessler [8], the 79 % of patients affected by lifetime BED also meet lifetime criteria for other DSM-IV/CID I disorders.

Although present diagnostic criteria have shown their empirical consistency [12, 13] other core psychopathologic features like body and weight overvaluation [14], impulsivity [15] may be of clinical and prognostic relevance. Body and weight overvaluation is present in 60 % of patients with BED, a level significantly higher than in non-BED obese patients [14].

Aims of this study are to discover the prevalence of new patients with BED according to the DSM-5 criteria in a sample of obese patients seeking bariatric surgery and to test if there is a difference in psychopathology between the patients seeking bariatric surgery diagnosed as being affected by Binge Eating Disorder, according to the new criteria of the DSM-5, the severely obese patients seeking bariatric surgery not affected by an eating disorder, according to the diagnostic criteria of the DSM-5 and a healthy weight Control group. We hypothesize that patients seeking bariatric surgery diagnosed as being affected by Binge Eating Disorder, according to the new criteria of the DSM-5 have higher levels of anxiety, depression and dysfunctional eating habits than severely obese patients seeking bariatric surgery not affected by an eating disorder, according to the diagnostic criteria of the DSM-5 and healthy weight participants.

Materials and methods

During consecutive 12 months 100 severely obese patients were sent to a bariatric surgery center in the ASL of Dolo Mirano (Venice). All of the patients were already assessed by their physicians for BED using the DSM-IV-TR criteria, and none of them were diagnosed with BED because they binged only once a week. All 100 patients were asked to take part in the current study, and they all agreed to participate. The patients were further assessed using the DSM-IV-TR criteria. One of the researchers of the current study (A.d.R.), an expert in eating disorders, conducted the clinical interviews. 6 of them were diagnosed with BED (because at that time they binged at least twice a week from at least 6 months) and were excluded from the study. The remaining 94 individuals were included in the study. There was no compensation for participation.

55 of the participants were female and 39 were male. Ages ranged from 18 to 67 years, with a mean age of 44.37 years (SD 13.07). BMI ranged from 32.14 to 66.16 kg/m2 with a mean BMI of 45.42 kg/m2, (SD 7.41).

These 94 patients were further screened for eating disorders by the same researcher (A d. R.) who conducted clinical interviews using the new DSM-5 [2] criteria.

33 of the participants (35.1 %) were diagnosed as being affected by BED following the new DSM-5 criteria and were labeled NEW BED group. 21 of the participants were male and 12 were female. The average age among participants was 43.82 years, (SD 12.39) (range 18–64 years) and mean BMI was 45.75 kg/m2 (SD 8.57).

The remaining 61 participants were labeled OB group: 27 were male and 34 were female. Average age was 44.72 years, (SD 13.41) (range 18–67 years) and mean BMI was 45.23 kg/m2 (SD 6.76).

41 healthy weight control subjects not affected by eating disorders were recruited by nurses, medical students and doctors at the same center (Control group). 15 of the subjects were male and 26 were female. The age range was 20–62 years, with a mean age of 40.54 years (SD 11.61). BMI ranged from 20.03 to 25 kg/m2.

Eating habits of all the participants were assessed using The Eating Disorder Inventory (EDI) [16] and The Three-Factor Eating Questionnaire (TFEQ) [17]. Depression and anxiety levels were assessed using the Beck Depression Inventory (BDI) [18] and the State Trait Anxiety Inventory (STAI) [19].

The EDI is divided into eight subscales: drive for thinness (DT), bulimia (B), body dissatisfaction (BD), ineffectiveness (I), perfectionism (P), interpersonal distrust (ID), interoceptive awareness (IA: the ability of an individual to discriminate between sensations and feelings and between the sensations of hunger and satiety) and maturity fears (MF: the fear of facing the demands of adult life).

The TFEQ is a self-assessment questionnaire developed to measure cognitive and behavioral components of eating. The instrument contains 36 items with a yes–no response format and 14 items on a 1–4 response scale. All item responses are dichotomized and aggregated into three scales. Restraint (R) (20 items) is designed to measure dietary restraint, which is control over food intake to influence body weight and body shape. Disinhibition (D) (16 items) measures episodes of loss of control over eating, and hunger (H) (14 items) represents subjective feelings of appetite and food cravings.

The BDI [18] provides a measure of depressive symptom severity. It includes 21 items, which have four possible answer choices ranging from 0 to 3. The cutoffs are as follows: 0–9 no depression; 10–18 mild–moderate depression; 19–29 moderate–severe depression, 30–63 severe depression.

The STAI [19] is a 40-item self-report measured using a 4-point Likert scale, with 2 subscales. These subscales include the state anxiety (SA: anxiety about an event) and the trait anxiety (TA: anxiety level as a personal characteristic). Higher scores are positively correlated with higher levels of anxiety.

Statistics

Chi square was used for the categorical variables (i.e., gender), while one-way analysis of variance (ANOVA) and Bonferroni post hoc analysis was used for the continuous variables (i.e., age and BMI) to evaluate the differences between NEW BED group, OB group and controls. Moreover, Kruskal–Wallis test and Mann–Whitney post hoc analysis (with Bonferroni correction) were performed to compare the previous groups about variables with non-normal distribution. The statistical analyses were performed using SPSS for Windows 20.0 (SPSS, IBM, 2011).

Results

Chi-square analysis found no significant difference between the groups in gender (female 55.7 % vs 63.6 % vs 63.4 %, χ 2 = 0.84, df = 2, p = 0.65) and ANOVA analysis (F = 1.35, df = 2, p = 0.262) showed no significant differences between the groups in age (M SD 43.81 ± 12.38 vs 44.67 ± 13.52 vs 40.53 ± 11.61). One-way analysis of variance (F = 182.58, df = 2, p < 0.001) showed that there were differences on BMI between the Control group and the NEW BED group (M SD 22.98 ± 1.57 vs 45.75 ± 8.57, Bonferroni post hoc p < 0.001) and between the Control group and obese group (M SD 22.98 ± 1.57 vs 45.24 ± 6.76, Bonferroni post hoc p < 0.001) while there were no significant differences in BMI between the Obese group and the NEW BED group (M SD 45.24 ± 6.76 vs 45.75 ± 8.57, Bonferroni post hoc p = 1.00).

Given that the three groups did not have a normal distribution, we used the Kruskal–Wallis test to find out if there were significant differences between the groups on EDI, BDI, STAI and TFEQ. Post hoc analysis were tested by Mann–Whitney test. To reduce Type I error, Bonferroni correction was applied; therefore, we considered only the p value under α = 0.016.

The average scores and the differences between the three groups on the EDI subscales are summarized in Table 1.

Table 1 Comparison of average scores on the Eating Disorders Inventory of NEW BED group, OB group and Control group
Full size table

Beck Depression Inventory

On the Beck Depression Inventory, the NEW BED group scored in the range of moderate to severe depression (mean 20.12, SD 9.20). The OB group showed slight signs of depression (mean 9.03, SD 7.26), while the Control group showed no signs of depression (mean score 5.95, SD 4.66). Kruskal–Wallis test showed significant differences between all the three groups (Table 2).

Table 2 Comparison of average scores on the Three-Factor Eating Questionnaire, the Beck Depression Inventory and the State–Trait Anxiety Inventory of NEW BED group, OB group and Control group
Full size table

State–Trait Anxiety Inventory

In terms of state anxiety, the NEW BED group scored 40.27 (SD 10.06), the OB group scored 32.75 (SD 6.93) and the Control group scored 30.02 (SD 7.19).

In terms of trait anxiety the NEW BED group scored 44.33 (SD 10.31), the OB group scored 33.87 (SD 9.07), and the Control group scored 33.22 (SD 8.79). Kruskal–Wallis test showed significant differences on both state and trait subscales of the STAI between the Control group and the NEW BED group and between the OB and the NEW BED group. There were no significant differences between the OB group and the Control group (see Table 2).

Three-Factor Eating Questionnaire

On the restraint subscale of the Three-Factor Eating Questionnaire (TFEQR), the NEW BED group scored an average of 5.64 (SD 3.61), the OB group’s average score was 6.77 (SD 4.10) and the Control group’s was 7.44 (SD 4.75).

On the disinhibition subscale, the NEW BED group scored 9.88 (SD 2.33), the OB group 5.93 (SD 2.89) and the Control group 4.41 (SD 2.61).

On the hunger subscale the NEW BED group scored 7.94 (SD 2.85), the OB group 3.79 (SD 3.03) and the Control group 3.24 (SD 2.48). The significant differences between the groups are summarized in Table 2.

Discussion

The recently revised DSM-5 criteria were used to evaluate the prevalence of BED among a sample of obese patients. The individuals who were diagnosed with BED (NEW BED group) and the remaining sample of obese patients who were not diagnosed with BED (OB group) were compared to a healthy weight group (Control group) on depression, anxiety and eating habits. We sought to examine whether new patients with BED diagnosed under the criteria of the DSM-5 would show higher levels of pathology than obese individuals without a BED diagnosis. The prevalence of NEW BED in our sample was high: 34.6 %. This prevalence may support the hypothesis that less restrictive criteria may induce an “epidemic” of BED [4] among severe obese patients seeking for bariatric surgery. On the other hand, this study shows many significant differences in eating habits and psychopathology between NEW BED patients and other obese patients, justifying the changes in diagnostic criteria for BED in the DSM-5. There were significant differences between the NEW BED group and the Control group on all the subscales of the EDI excluding the ID subscale and significant differences between the NEW BED group and the OB group on the subscales: DT, B, BD, and IA of the EDI. There were no significant differences between the NEW BED group and the OB group on the subscales: I, P, ID and MF. The results suggest that BED patients experience difficulty dealing with their body image and their inner sensations. The difficulty in distinguishing bodily sensation from emotion may result in binge eating. Our most valuable result is the presence of a significant difference on the Bulimia subscale of the EDI between the NEW BED group and the OB group. This result confirms that the DSM-5 criteria for BED are efficient in identifying binge eating behaviors among obese patients.

We found significant differences between the NEW BED group and the other groups on BDI, confirming the well-known relationship between depressive symptoms and BED [20]. There were also significant differences on the BDI between the OB group and the Control group. It is noteworthy that the average score of the OB group on the BDI was 9.03 (SD 7.26). This average score was lower than the levels of depressive symptoms reported among obese patients not affected by BED, using the diagnostic criteria from the DSM-IV-TR criteria [11–13], confirming the efficacy of BED DSM-5 criteria in identifying patients with higher level of depressive symptoms among severe obese patients.

On both subscales of the STAI there were significant differences between the NEW BED group and both the OB group and the Control group. We did not find any significant difference between the OB group and the Control group for anxiety; both groups scored within the normal range. These results were unexpected, given that many studies report high levels of anxiety among obese individuals seeking bariatric surgery [20–23]. This datum suggests that anxiety is more related to BED than to obesity alone.

On The TFEQ there was a significant difference between the NEW BED group and the OB group on both the disinhibition and hunger subscales, but no difference was found on the restraint subscale. From a clinical perspective, a combination of high disinhibition and low restraint has been associated with a higher susceptibility to weight gain and sedentary behavior [24]. Our results reinforce the need to assess disinhibition, which may help identify subjects at a high risk for BED among obese patients. Disinhibition is a typical characteristic among BED patients [24, 25], but it has been also reported among obese patients [26]. There was a significant difference between the OB group and the Control group on this subscale, while there was no significant difference between these two groups on hunger subscale of the TFEQ suggesting that (if BED patients are carefully excluded), the difficulty in managing hunger is not related to obesity in itself. This datum, if confirmed by further studies, may encourage clinicians to modify how they treat obesity.

Conclusion

The results of the current study offer a contribution in clarifying the controversial relationship between obesity and BED, and may provide support for the efficiency of the new diagnostic criteria for BED in the DSM-5. As reported in previous studies, obese patients have several common features with patients affected by BED. Higher scores on several subscales of the Eating Disorders Inventory confirm the important role played by psychological factors [27] in the onset and maintenance of obesity and the necessity of a psychological assessment for these patients. However, several psychiatric disorders such as depression and anxiety, which were once attributed to obesity, appear to be more associated with comorbidity. The new diagnostic criteria published in the DSM-5 are useful in identifying patients with high levels of psychopathology, showing that patients who rarely binge have peculiar psychopathological characteristics. This factor supports the validity of the new criteria of the DSM-5 in differentiating patients with BED from other obese patients. The new diagnostic criteria appear to be useful in identifying BED patients, which may contribute to better treatment for BED patients who are at high risk for obesity [28, 29].

Limits

In our research were only included severely obese patients undergoing bariatric surgery. Findings may not be generalized beyond this specific population. Moreover, given the small sample of patients included in our study, more research in this area needs to be conducted to support our results.