Abstract
Purpose:
Studies comparing individuals with loss of control (LOC) eating who do and do not have objectively large binge episodes have found that degree of LOC is more important than binge size to psychological and behavioral outcomes. However, the relative importance of these characteristics has not been investigated in a population with binge eating disorder (BED), who by definition all have objectively large binge episodes.
Methods:
Persons with BED and higher weight (N=34) were enrolled in a BED treatment trial and completed the Loss of Control Over Eating Scale, the Eating Disorder Examination, and measures of eating behavior, mood, and quality of life. Body mass index (BMI) was calculated from measured height and weight. The size of the largest binge episode (measured in kilocalories) and degree of LOC were entered into multiple regression equations to determine their relationships with disordered eating symptoms, depression, quality of life, and BMI in this pilot study.
Results:
Greater LOC had a stronger independent association than binge size with higher total eating psychopathology, shape dissatisfaction, hunger, food cravings and food addiction symptoms. Larger binge size had a stronger independent association than LOC with higher weight concern and lower general and social quality of life. Both characteristics were associated with higher eating concern and neither were associated with depression or BMI.
Conclusion:
Both binge size and degree of LOC are associated with important psychosocial treatment targets in patients with BED. Future research should validate the largest binge episode measurement method and replicate the present findings in a larger sample.
Keywords: binge eating disorder, overweight, binge size, loss of control, eating disorder symptoms, food cravings
Introduction
Binge eating disorder (BED) is characterized in the Diagnostic and Statistical Manual of Mental Disorders-5 by recurrent objective binge episodes (OBEs) [1]. OBEs are defined by a subjective sense of loss of control (LOC) during an eating episode that involves consuming an unambiguously large quantity of food in a 2 hour period, i.e., the quantity of food must be definitely larger than most people would eat under similar circumstances [1]. BED is associated with a number of negative physiological and psychological outcomes, including higher odds of obesity, diabetes, hypertension, and certain chronic pain conditions, as well as mood disorders, panic disorder, and substance use disorders [2].
Researchers have struggled to define the quantity of food required for an eating episode to qualify as an OBE. One study showed that the upper limit of normal calorie consumption in a two-hour period was 1418 kcal among those with BED and higher body weight and 1049 kcal among those of a similar weight without BED [3]. Forney et al. [4] has suggested that consuming 1000 calories of food during an eating episode may be used as a cutoff for an unambiguously large amount of food. Other guidelines focus on the number of servings consumed. However, patients in clinical settings often describe episodes in which they feel LOC, but the quantity of food consumed is not clearly beyond the threshold for socially normative eating. Some individuals with BED, bulimia nervosa, or unspecified eating disorders also have subjective binge eating episodes (SBEs), which, like OBEs include a subjective sense of LOC during an eating episode, but the quantity of food is not considered unambiguously large [5].
Research findings do not lend strong support to the importance of the size of an eating episode in terms of psychological and behavioral outcomes. Initial research examined the importance of binge size by comparing individuals who primarily have SBEs to individuals who primarily have OBEs in samples with subclinical or clinical BED, bulimia nervosa, or purging disorder. In these studies, patients with SBEs did not differ from those with OBEs on level of LOC, eating disorder psychopathology, binge eating severity, depression, anxiety, stress, substance use, quality of life, or impairment associated with eating problems [6–12]. In a sample of patients identifying as men who reported at least one LOC eating episode in the past month, men with only OBEs and men with only SBEs did not differ on body mass index (BMI), weight-related medical comorbidities, dietary restraint, body image concerns, or excessive exercise [13]. In summary, these studies do not suggest that the size of patients’ LOC eating episodes is an indicator of the severity of other medical or psychological comorbidities.
Several studies have identified that women use LOC to differentiate between binge eating episodes (objective and subjective) and overeating episodes [14, 15]. In one study, women who had higher weight and BED were asked to provide free-response definitions of the components of a binge. In this sample, 82% expressed that LOC was a component of a binge, but only 43% stated that eating a large quantity was a component [15], suggesting that LOC may be the most salient feature of a binge episode. Further, researchers have shown that in a university student sample, degree of LOC, as measured by the Loss of Control Over Eating Scale, was positively associated with measures of eating disorder psychopathology, psychosocial impairment due to eating disturbance, psychological distress, and BMI, and negatively correlated with general self-control and the quality of individuals’ mental and physical health [16].
Although binge size may not be as strongly correlated with cognitive and behavioral outcomes as level of LOC, it appears to be more relevant than the degree of LOC to physical outcomes such as weight status. Unlike in Kelly et al.’s study mentioned above [13], two other studies found that individuals that had OBEs had significantly higher mean BMIs (29.0 kg/m2 and 27.4 kg/m2, respectively) than comparison groups with SBEs (25.7 kg/m2 and 24.1 kg/m2, respectively) [8, 11]. Additionally, Guss et al. [17] found that binge size in a laboratory setting was positively correlated with BMI in women with BED. These findings suggest that binge size might predict which individuals with LOC eating are more likely to develop or maintain higher body weights. Associations between binge size and body weight also may have implications for weight-related health outcomes (e.g., hypertension), physical health-related psychosocial outcomes (e.g., quality of life variables affected by health), and possibly weight and shape concerns based on these previous studies [8,11,17].
Previous studies have investigated OBEs versus SBEs and LOC as correlates of psychosocial and physical outcomes in samples of primarily women with bulimia nervosa, purging disorder, or subclinical BED. Few previous studies have investigated these variables in a treatment-seeking sample with BED, and none have considered LOC simultaneously with binge size, both as continuous variables. Assessing both characteristics together in a mixed-sex BED sample would clarify how binge size and LOC relate to primary and secondary treatment targets among both males and females with BED. The findings from this study could potentially help researchers determine meaningful subtypes of BED that vary on binge size or LOC levels. Additionally, the results could help researchers design studies to determine whether tailoring the treatment approach based on baseline levels of these variables could improve treatment outcomes.
The present pilot study simultaneously assessed binge size (kilocalories in the largest recent binge) and degree of LOC overeating as independent correlates of eating disorder pathology, eating behavior, quality of life, depression, and BMI. We hypothesized that higher LOC and larger binge size would be independently associated with higher eating concern, depressive symptoms, and lower quality of life. Finally, we hypothesized that binge size, but not degree of LOC, would independently correlate with higher hunger, weight concern, shape concern, and BMI. Exploratory analyses examined associations with additional outcomes, including food addiction and trait- and state-based measures of food cravings.
Methods
Participants
The present study was a pilot, post-hoc analysis of baseline data collected in a randomized trial (NCT0327973). Participants were a total of 34 males and females (based on sex assigned at birth; the sample contained one transgendered individual who had not received surgical treatment) diagnosed with BED who were enrolled in a randomized double-blind placebo-controlled trial of liraglutide 3.0 mg/d for the treatment of BED. Inclusion criteria for the parent trial were meeting full DSM-5 criteria for BED, being 21 to 70 years of age, and having a BMI of 27 kg/m2 or above. Exclusionary criteria were: serious medical conditions (e.g., diabetes, recent cardiovascular disease or cancer, uncontrolled hypertension, history of pancreatitis, certain thyroid problems, liver or kidney dysfunction), history of psychosis, current severe major depression, active suicidal ideation, recent psychiatric hospitalization, recent substance use disorder, history of a suicide attempt in the past 5 years, being pregnant or nursing, weight loss of ≥10 lb. in the past 3 months, use of medications known to influence appetite or weight, history or plans to undergo bariatric surgery, and current psychotherapy for eating disorder symptoms or participation in a weight loss program. Participants were recruited through social media and print advertisements in a greater metropolitan area.
Procedure
The study was conducted at a university medical center. The screening visit included: completion of the informed consent; measurement of anthropometrics and vitals; semi-structured interviews including the Eating Disorder Examination to confirm BED diagnosis; and completion of questionnaires focused on disordered eating and psychosocial variables. The University’s institutional review board approved all study procedures. Participants received $30 for completing the screening visit.
Measures: Classification Variables
Demographic data.
Information concerning age, sex, race, ethnicity, and marital status was collected using a questionnaire.
Loss of control.
The Loss of Control Over Eating Scale (LOCES) [16] is a 24-item self-report measure of loss of control eating in the last 4 weeks, with item scores of 0 (never) to 5 (always). The measure consists of three subscales and a total score can be calculated as the average of all item scores. The three aspects of loss-of-control eating measured on these subscales and comprehensively measured with the total score include behavioral, cognitive/dissociative, and positive/euphoric. We did not include the subscale scores in this study. The internal reliability of the total score in our sample was Cronbach’s α = .94.
Binge size.
Binge size was defined as the number of calories in a participant’s largest binge in the previous 28 days, as assessed by the Eating Disorder Examination (see below) at screening. The mobile application MyFitnessPal [18] was used to determine the number of calories in participants’ largest reported OBE by entering the foods and amounts consumed. Calories were estimated if detailed information about the type or brand of food were not provided by the participant. For example, a participant may have stated that he or she consumed a half gallon of ice cream, but not specified the flavor or brand. To be consistent across participants, we used the same food type and brand whenever that food was unspecified. See Appendix 1 for examples of brands that were used in these calculations.
The highest-calorie binge episode for each participant was selected and calculated by the first author. The last author checked a subset (n=5, 15.6%) of the Eating Disorder Examination interviews to verify agreement of the selection of the participants’ largest reported episodes and the calorie calculation in MyFitnessPal. The intra class correlation coefficient (absolute agreement, two-way mixed model, average measure) for the five binges for the two raters was .94 (95% CI = .58 - .99), F = 18.76, p = .007. Previous research has shown a Pearson’s correlation of r=0.93 between MyFitnessPal and the Nutrition Data System for Research for total energy consumption, with no significant differences in calorie totals between the two methods [19]. Similarly, MyFitnessPal was highly correlated with the Belgian food compositional database, Nubel, for total energy intake (r = 0.96) [20].
Measures: Primary Outcome Variables
The Eating Disorder Examination, 16th Edition, Diagnostic Version
[21] is a semi-structured interview that was used to determine eating disorder diagnosis based on DSM-5 criteria, binge size, and current eating disorder psychopathology. The Eating Disorder Examination was administered by a licensed psychologist or clinical psychology postdoctoral fellow at the screening visit. The postdoctoral fellow received Eating Disorder Examination training and ongoing supervision. The Eating Disorder Examination yields a global score as well as four subscale scores, which are listed with the Cronbach’s α internal reliability scores for each in our sample: Restraint subscale (.58), Eating Concern subscale (.62), Shape Concern subscale (.64), and Weight Concern subscale (.58). The internal reliability for the total score in our sample was Cronbach’s α = .61.
The Eating Inventory
[22, 23] is a 51-item self-report questionnaire that assesses three dimensions of eating behavior: Cognitive Restraint, Disinhibition, and Hunger. We did not use the Disinhibition subscale in our analyses, because there is substantial construct overlap between the LOCES and this subscale. Internal consistencies for the current sample were: Cognitive Restraint (Cronbach’s α = .73) and Hunger (Cronbach’s α = .77).
Quality of Life Enjoyment and Satisfaction Questionnaire
[24] is a measure of enjoyment and satisfaction experienced by individuals in a number of life domains during the past week, including (with Cronbach’s αs): physical health/activities (.88), feelings (.87), work (.93), household duties (.90), leisure time activities (.89), and social relations (.91). It also features questions on global satisfaction in these domains in the past week (.91).
Patient Health Questionnaire-9 (PHQ-9)
[25] is a 9-item self-report measure that assesses symptoms of depression in the past two weeks, based on the DSM-IV criteria for major depression [26]. A score of 20 is indicative of possible severe depression and was a criterion for exclusion in this study. The internal reliability of this measure in our sample was Cronbach’s α = .82.
BMI.
Weight was measured on an electronic scale (Tanita Corp., BWB-800, Japan) with participants dressed in light clothing (without shoes), and height was assessed using a wall-mounted stadiometer (Veeder-Root, Elizabethtown, NC). These measures were used to calculate BMI as weight (kilograms) / [height (meters)]2.
Measures: Exploratory Outcome Variables
Food Craving Questionnaire-Trait
[27] is a 39-item self-report measure of trait or habitual food craving experiences. It assesses external cues that may trigger food craving, emotions that may trigger or be experienced during craving, physiological deficits and responses that may elicit craving, preoccupation with food, outcome expectancies of positive and negative reinforcement, intentions and plans related to consumption, anticipated lack of control when eating, and guilt. The internal reliability of this measure in our sample was Cronbach’s α = .96.
Food Craving Questionnaire-State
[27] is a 15-item self-report measure of current intense desire to consume a specific food, outcome expectancies of positive reinforcement and negative reinforcement from eating the food, anticipated lack of control when eating the food, and physiological states that may trigger craving. The internal reliability of this measure in our sample was Cronbach’s α =.89.
Yale Food Addiction Scale 2.0
[28] is a 35-item self-report measure of addictive-like eating behavior over the past 12 months. To score this measure continuously, 11 subscales representing aspects of food addiction criteria are calculated (e.g., having a persistent desire to stop eating certain foods but having trouble eliminating them). Each subscale is then dichotomized to signify whether a given food addiction criterion was met. A symptom count score is calculated as the sum of the dichotomous subscale scores, yielding a count of the number of food addiction symptoms endorsed at threshold level. The internal reliability of the Yale Food Addiction Scale 2.0 symptom count score in our sample was Cronbach’s α =.94.
Statistical Analysis
Analyses were conducted using SPSS, version 25 [29]. All variable distributions were normal. We identified one high outlier in binge size. Because this calorie value was plausible, the score was included in all analyses. We first examined bivariate correlations between binge size, LOC, and the outcome variables described above. Only outcomes that had moderate bivariate correlations (r ≥ .30) with at least LOC and/or binge size were then evaluated using multiple regression to determine the independence of these effects.
In the multiple regression models, largest binge size and LOC were entered concurrently. Semi-partial correlations (sr) are shown as a measure of effect size. The square of this value (sr2) yields how much total r2 would decrease if the variable were removed from the regression. To explore whether the relationship between binge size and psychosocial outcomes depended on degree of loss of control, we initially included the interaction term in all models. Non-significant interaction terms were removed from the final models presented below. An alpha value of .05 was used to detect statistical significance. Cohen’s effect size conventions for r and f2 were used to characterize effect sizes descriptively as small, medium, or large [30].
The sample size available for the present analyses was determined by the parent study. We conducted a sensitivity power analysis using G*Power 3.1 to identify the minimum effect sizes that could be detected in the present study. Based on the available sample size of 34 participants, we had at least 80% power to detect bivariate relationships of r = 0.339 or larger. In multiple regression, we were powered to detect independent effects of r2change = 0.198 or larger in models with all three independent variables (binge size, LOC, and the interaction term). Due to the exploratory nature of these analyses and small available sample size, we did not apply an alpha correction method to control for multiple testing.
Results
Participants’ Characteristics
The sample was 65% women and 71% non-Hispanic white, with a mean age of 41.2 ± 10.1 years and BMI of 36.2 ± 6.5 kg/m2 (Table 1). Most participants had a bachelor or graduate degree (61%). Table 2 presents participants’ mean (SD) values on the classification and primary outcome variables as well as score ranges for each measure. Their mean LOC score, as measured by the Loss of Control Over Eating Scale, was 3.25 ± .71. This score is consistent with moderate levels of LOC. This sample’s largest reported binge episode was determined to contain a mean of 2999.97 ± 1224.97 kcal (Median: 2733.5, IQR: 1649). The participants reported that in the past 28 days they had an average of 16.50 OBEs and 5.50 SBEs, which corresponded to a weekly average of 4.13 OBEs and 1.38 SBEs, corresponding with a moderate severity of BED. Participants’ mean score of 6.89 ± 4.52 (range 0 to 18) on the PHQ-9 suggests that they experienced mild symptoms of depression.
Table 1.
Participants’ demographic characteristics.
| Variable | |
|---|---|
| Age, years, mean (SD) | 41.15 (10.07) |
| Body mass index (kg/m2) | 36.15 (6.49) |
| Sex, n (%) | |
| Women | 22 (65%) |
| Men | 11 (32%) |
| Other | 1 (3%) |
| Race, n (%) | |
| White | 24 (71%) |
| Black | 10 (29%) |
| Education, n (%) | |
| High School | 4 (12%) |
| Some College or | 7 (21%) |
| Associate’s Degree | |
| Bachelor’s Degree | 11 (32%) |
| Graduate Degree | 10 (29%) |
| Not reported | 2 (6%) |
| Marital Status | |
| Single | 12 (35%) |
| Married/In Committed Relationship | 15 (44%) |
| Divorced | 5 (15%) |
| Not reported | 2 (6%) |
Table 2.
Participants’ mean (SD) values on the independent, primary, and exploratory variables
| Variable | Mean (SD) | Possible Score Range |
|---|---|---|
| LOCES | 3.25 (.71) | 1–5 |
| Binge size (kcals in largest binge) | 2999.97 (1224.97) | N/A |
| EDE-Total | 2.70 (.84) | 0–6 |
| EDE-Restraint | 1.88 (1.31) | 0–6 |
| EDE-Eating Concern | 2.14 (1.38) | 0–6 |
| EDE-Weight Concern | 3.20 (1.09) | 0–6 |
| EDE-Shape Concern | 3.57 (1.11) | 0–6 |
| EI-Cognitive Restraint | 7.71 (3.52) | 0–21 |
| EI-Hunger | 10.44 (2.97) | 0–14 |
| PHQ-9 | 6.89 (4.52) | 0–27 |
| QLES-General | 50.56 (8.40) | 16–80 |
| QLES-Physical Health/Activities | 39.52 (8.35) | 13–65 |
| QLES-Feelings | 52.13 (7.1) | 14–70 |
| QLES-Work | 51.68 (7.63) | 13–65 |
| QLES-House | 38.53 (6.61) | 10–50 |
| QLES-Social | 41.54 (7.06) | 11–55 |
| FCQ-Trait | 160.89 (34.55) | 39–234 |
| FCQ-State | 50.55 (10.95) | 15–75 |
| YFAS 2.0-Symptom Count | 6.50 (3.19) | 0–11 |
LOCES: Loss of Control over Eating Scale. EDE: Eating Disorder Examination. EI: Eating Inventory. PHQ-9: Patient Health Questionnaire-9. QLES: Quality of Life Enjoyment and Satisfaction Questionnaire. FCQ: Food Craving Questionnaire. YFAS 2.0: Yale Food Addiction Scale 2.0
Correlates of Binge Size and LOC
Largest binge size and LOC both had a minimal or small bivariate correlation (r < .30) with BMI, Eating Disorder Examination Restraint, Eating Inventory Cognitive Restraint, and certain domains of quality of life (work, household duties, leisure time activities) (Table 3). Therefore, multiple regression analyses were not run for these dependent variables. Largest binge size was moderately correlated with higher eating and weight concern and lower quality of life (general, physical health/activities, feelings, and social subscales). Degree of LOC was moderately correlated with higher hunger, depression, eating concern, shape concern, and total eating pathology, and lower quality of life (general and physical health/activities subscales).
Table 3.
Correlation matrix of key variables
| Binge size | LOCES | EDE_Res | EDE_EC | EDE_SC | EDE_WC | EDE_Tot | EI_CR | EI_H | QLES_PH/A | QLES_Feel | QLES_Work | QLES_HD | QLES_Leis | QLES-Soc | QLES_Gen | PHQ-9 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Binge size | |||||||||||||||||
| LOCES | .03 | ||||||||||||||||
| EDE_Res | −.11 | .18 | |||||||||||||||
| EDE_EC | .35* | .42* | −.04 | ||||||||||||||
| EDE_SC | .26 | .43* | .16 | .45** | |||||||||||||
| EDE_WC | .37* | .28 | .22 | .35* | .71** | ||||||||||||
| EDE_Tot | .31 | .48** | .50** | .67** | .81** | .79** | |||||||||||
| EI_CR | −.15 | −.28 | .23 | −.15 | −.14 | .12 | .02 | ||||||||||
| EI_H | −.22 | .45** | −.22 | .04 | .18 | −.13 | −.06 | −.31 | |||||||||
| QLES_PH/A | −.32 | −.33 | −.22 | −.19 | −.16 | −.23 | −.30 | .15 | −.25 | ||||||||
| QLES_Feel | −.33 | −.28 | −.38* | −.29 | −.07 | −.09 | −.33 | .08 | −.11 | .58** | |||||||
| QLES_Work | −.05 | −.29 | −.37 | −.05 | .15 | .05 | −.10 | −.10 | −.23 | .42* | .54** | ||||||
| QLES_HD | −.07 | −.11 | −.19 | −.09 | .10 | .03 | −.07 | −.11 | −.07 | .28 | .32 | .50* | |||||
| QLES_Leis | −.26 | .09 | −.14 | .14 | .21 | −.00 | .07 | −.33 | .09 | .27 | .50** | .49** | .09 | ||||
| QLES_Soc | −.45* | −.12 | −.17 | −.10 | −.02 | −.05 | −.13 | −.04 | −.21 | .50** | .66** | .56** | .39* | .61** | |||
| QLES_Gen | −.41* | −.30 | −.21 | −.08 | −.04 | .01 | −.13 | .13 | −.07 | .53** | .72** | .53** | .52* | .50* | .76** | ||
| PHQ-9 | .22 | .31 | .06 | .26 | .09 | −.14 | .12 | −.15 | .22 | −.45* | −.51** | −.17 | −.17 | −.32 | −.52** | −.60** | |
| BMI | .25 | −.09 | −.25 | .15 | .34* | .26 | .16 | −.24 | .00 | −.35 | −.15 | .01 | .30 | −.04 | −.10 | −.18 | .06 |
Binge size: Calories in largest binge eating episode in past month. LOCES: Loss of Control over Eating Scale. EDE_Res: Eating Disorder Examination Restraint Subscale. EDE_EC: Eating Disorder Examination Eating Concern Subscale. EDE_SC: Eating Disorder Examination Shape Concern Subscale. EDE_WC: Eating Disorder Examination Weight Concern Subscale. EDE_Tot: Eating Disorder Examination Total. EI_CR: Eating Inventory Cognitive Restraint Subscale. EI_H: Eating Inventory Hunger Subscale. QLES: Quality of Life Enjoyment and Satisfaction Questionnaire. PH/A: Physical Health/Activities Subscale. Feel: Feelings Subscale. Work: Work Subscale. HD: Household Duties Subscale. Leis: Leisure Subscale. Soc: Social Subscale. Gen: General Subscale. PHQ-9: Patient Health Questionnaire-9.
Primary Analyses
Across all models, there were no significant interactions between binge size and degree of LOC (r2change .001 to .06), indicating that the relationship between binge size and these psychosocial outcomes did not depend on level of LOC. The interaction terms were therefore removed from the final models.
Larger binge size and greater LOC were both independently associated with higher levels of Eating Disorder Examination Eating Concern (Binge size: β = .340, sr2 = .116, p = .032; LOC: β =.411, sr2 = .168, p = .011) (Table 4). Higher LOC, but not binge size, was independently correlated with higher global scores on the Eating Disorder Examination (Binge size: β = .296, sr2 = .088, p = .056; LOC: β = .468, sr2 = .219, p = .004), Eating Disorder Examination Shape Concern (Binge size: β = .245, sr2 = .060, p = .127; LOC: β = .421, sr2 = .177, p = .011), and Eating Inventory Hunger (Binge size: β = −.234, sr2 = .055, p = .141; LOC: β = .455, sr2 = .207, p = .006). Greater binge size, but not LOC, was independently associated with higher Eating Disorder Examination Weight Concern (Binge Size: β = .361, sr2 = .130, p = .031; LOC: β =.267, sr2 = .071, p = .105) and lower general (Binge Size: β = −.409, sr2 = .167, p = .016; LOC: β =−.296, sr2 = .088, p = .074) and social quality of life (Binge Size: β = −.448, sr2 = .201, p = .012; LOC: β =−.128, sr2 = .016, p = .452). The significant associations were all medium in size.
Table 4.
Binge size and loss of control as explanatory factors of eating psychopathology, eating behavior, and quality of life
| Outcome Variables | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| EDE_EC | EDE_SC | EDE_WC | EDE_Glo | EI_H | QLES_PH/A | QLES_Feel | QLES_Soc | QLES_Gen | PHQ-9 | ||
| Full Model | |||||||||||
| Unadjusted R 2 | .293 | .244 | .208 | .316 | .255 | .206 | .188 | .215 | .259 | .134 | |
| F | 6.42** | 4.99* | 4.07* | 7.14** | 5.31* | 3.76* | 3.24 | 3.85* | 5.06* | 2.40 | |
| p | .005 | .013 | .027 | .003 | .010 | .035 | .054 | .033 | .013 | .108 | |
| Predictors | Beta values | ||||||||||
| Binge size | .340* | .245 | .361* | .296 | −.234 | −.312 | −.328 | −.448* | −.409* | .211 | |
| LOCES | .411* | .421* | .267 | .468** | .455** | −.325 | −.278 | −.128 | −.296 | .292 | |
Binge size: Calories in largest binge eating episode in past month. LOCES: Loss of Control over Eating Scale. EDE_EC: Eating Disorder Examination Eating Concern Subscale. EDE_SC: Eating Disorder Examination Shape Concern Subscale. EDE_WC: Eating Disorder Examination Weight Concern Subscale. EDE_Glo: Eating Disorder Examination Global Score. EI_H: Eating Inventory Hunger Subscale. QLES: Quality of Life Enjoyment and Satisfaction Questionnaire. PH/A: Physical Health/Activities Subscale. Feel: Feelings Subscale. Soc: Social Subscale. Gen: General Subscale. PHQ-9: Patient Health Questionnaire-9.
p<0.05.
p<0.01.
p<0.001
The overall model did not reach statistical significance for predicting emotional quality of life or depression (though a medium amount of variance was explained by both models). The combination of binge size and degree of LOC predicted lower physical health and activity-related quality of life, but neither variable was a significant independent predictor in a sample of this size (Binge Size β = −.312, sr2 = .097, p = .069; LOC β =−.325, sr2 = .106, p = .059).
Exploratory Analyses
In the exploratory analyses, greater LOC, but not largest binge size, had a positive independent association with trait (Binge Size: β = −.127, sr2 = .015, p = .388; LOC: β =.719, sr2 = .486, p < .001) and state food cravings (Binge Size: β = .008, sr2 < .001, p = .953; LOC: β =.671, sr2 = .450, p < .001), and number of symptoms of food addiction (Binge Size β = −.094, sr2 = .009, p = .421; LOC β =.801, sr2 = .635, p < .001) (Table 5). The independent associations between LOC and these variables were all large in size.
Table 5.
Binge size and loss of control as exploratory factors of cravings and food addiction
| Outcome Variables | ||||
|---|---|---|---|---|
| FCQ-T | FCQ-S | YFAS 2.0 Symptom Count | ||
| Full Model | ||||
| Unadjusted R 2 | .489 | .451 | .636 | |
| F | 12.43*** | 12.34*** | 24.47*** | |
| p | <.001 | <.001 | <.001 | |
| Predictors | Beta values | |||
| Binge size | −.127 | .008 | −.094 | |
| LOCES | .719*** | .671*** | .801*** | |
Binge size: Calories in largest binge eating episode in the past month. LOCES: Loss of Control over Eating Scale. FCQ-T: Food Craving Questionnaire-Trait. FCQ-S: Food Craving Questionnaire-State. YFAS 2.0: Yale Food Addiction Scale 2.0.
p<0.05.
p<0.01.
p<0.001
Discussion
This pilot study provides a novel contribution to the literature by examining whether largest binge size or degree of LOC over eating were more strongly associated with eating disorder pathology, quality of life, depression, BMI, and eating behavior in a mixed-sex treatment-seeking sample of individuals with BED. Higher LOC had stronger independent associations with overall eating psychopathology (Eating Disorder Examination global score), shape dissatisfaction, hunger, food cravings and food addiction symptoms. Largest binge size had relatively stronger independent associations with weight concern and lower general and social quality of life. Eating concern was the only variable for which both predictors reached statistical significance. We note that due to the study’s small sample size (34 participants), we were only able to reliably detect effects that were medium in size.
Total eating psychopathology, eating concern, shape concern, and hunger were more strongly associated with degree of LOC than with binge size. The medium-sized associations between LOC and aspects of disordered eating are consistent with previous research that has identified LOC as a key feature of binge episodes [14, 15] and found associations between LOC and eating disorder pathology [16]. Further, this sample’s score on the Eating Inventory’s Hunger subscale score (10.44 ± 2.97) was consistent with scores reported in other samples with BED [3, 17, 31]. It may be that individuals who feel hungry more frequently experience more LOC when they do eat. Binge size had a stronger association with weight concern than did LOC and was a significant predictor of eating concern (medium effect). These findings may suggest a stronger relationship between binge size and eating psychopathology than has been detected in previous studies comparing individuals with OBEs to those with SBEs [6–13].
Neither binge size nor degree of LOC had medium bivariate correlations with dietary restraint as measured by the Eating Disorder Examination or the Eating Inventory, and these variables were not evaluated in multiple regression. This sample’s score on the Eating Inventory’s Cognitive Restraint subscale score (7.71 ± 3.52) was consistent with scores reported in other samples with BED [3, 17, 31]. However, the negative association between Eating Inventory Cognitive Restraint and LOC approached a medium size and would be worthy of investigation in future studies.
Binge size and degree of LOC had independent associations with depression and several aspects of quality of life that approached or exceeded a medium effect size. The only relationships that reached statistical significance in our sample were between binge size and lower general and social quality of life. Emotional and physical health/activity-related quality of life were predicted by these two binge characteristics combined, but the independent effects did not reach statistical significance. Bivariate correlations with quality of life related to work, household duties, and leisure were small or minimal in size and were not tested in multiple regression. These findings generally suggest that more severe binge pathology is associated with lower quality of life across several domains. This pattern is again consistent with previous research on LOC (16), but not with studies reporting that patients with OBEs vs SBEs do not differ in quality of life [8, 11, 12].
Although we had hypothesized that binge size would be associated with BMI, the bivariate correlation between binge size and BMI was small (r = .25) and did not reach our threshold for testing in multiple regression. Previous research has found that binge size is associated with higher BMI in those with BED [17] and that individuals with OBEs have higher BMIs than those with SBEs [8, 11]. Although the direction of our result was consistent with these findings, a larger sample would be needed to determine its reliability. The correlation between LOC and BMI was minimal in size, suggesting that there is not likely to be a relationship between these variables.
Exploratory outcomes including trait and state food cravings and food addiction symptoms also were strongly related to LOC but were not associated with binge size. Degree of LOC may serve as a relatively strong predictor of hedonic eating and eating involving compulsivity (i.e., a sense of urgency to act or a need to fulfill a desire), which are often components of food craving episodes and addictive-like eating behavior.
A strength of this study was that calories in binge episodes were used to provide an indicator of relative OBE size in a population with BED. This differentiates this study from previous research on the importance of binge size, which has typically compared groups with OBEs to groups with SBEs. The participants’ mean largest binge episode was estimated to contain 2999.97 (Median = 2733.5) kcal. This size was considerable larger than the 1898 to 2388 kcal average binge size observed in a study in which participants with BED were instructed to binge during a laboratory meal [17]. This difference may be attributable to differences in research methods, including our selection of the participants’ largest recent binge (vs. an average binge), observation effects in the laboratory study, or accuracy limitations inherent to self-report. However, the frequency of OBEs and SBEs in the current study was similar to the 15.5 to 20.4 OBEs and 2.6 to 4.8 SBEs found in two other treatment-seeking samples with BED who were also assessed with the EDE [32–34]. Our sample’s mean global Eating Disorder Examination score was 2.70 ± .84, which was also consistent with the score of 2.5 found in a primary care sample with subthreshold and threshold BED [35]. Of note, largest binge size was not related to degree of LOC in our study, nor did the relationship between binge size and any outcome variable depend on level of LOC. These findings suggest that binge size and degree of LOC are independent characteristics of an OBE.
The greatest limitation of this pilot study was the small sample size, which we have tried to take into consideration by discussing both the statistical significance and the size of our effects. Due to this limitation, we also were not able to control for multiple testing and the statistical significance of results should be interpreted with caution. Because this study was cross-sectional, we are also not able to determine whether degree of LOC and binge size have causal relationships with the psychological and physiological outcomes discussed above. Further, calculating energy intake for the binge episodes was difficult when no brand names for foods were provided; we standardized our methods as much as possible, but this remains a limitation. Finally, this sample was treatment-seeking and the majority of participants were well educated (61% had a college or graduate degree) and identified as non-Hispanic White women, so generalizability to other groups may be limited.
Our study used a novel method of assessing binge size as calories consumed in the largest binge episode in the past month. This method allowed us to compare the relative effects of binge size in a sample with BED (who, by definition, all experience OBEs) in a free-living environment. Although our results suggested that the assessment of binge size using this method has utility, further research is needed to determine the reliability of measuring the largest binge in this manner, as well as the relationship between participants’ largest and average binge sizes. Because data on binge foods were collected via retrospective self-report, it would also be useful to validate this measure by comparing it to food record and laboratory-based assessments of binge size. Future studies might investigate whether methods like ecological momentary assessment or digital food images can be used to more accurately quantify total kilocalories consumed in binge episodes occurring outside of the laboratory.
Our findings from this pilot work suggested that within a population with BED, largest binge size may have more relevance to psychosocial outcomes than has been detected by previous studies comparing patients with OBEs and SBEs, particularly with regards to weight-related concerns that may impact quality of life. A clear next step is to further validate the largest binge episode measurement strategy and replicate the present findings in a larger sample. These findings also suggested that even in a sample of patients with BED (who by definition all experience loss of control eating), degree of LOC plays a key role in patients’ level of eating-related distress. Eating disorder interventions should continue to focus on management of this feature. We hope that this study can inform future research determining the relative importance of binge size and degree of LOC to psychopathological, psychosocial, and health-related outcomes in patients with BED.
Appendix
Examples of Rules Applied for Calorie Estimation of Binge Episodes
For fast-food items, participants usually provided enough information to match the foods directly in the database (e.g., McDonald’s McGriddle, Wendy’s medium fries). For pizza, potato chips, ice cream, cookies, and cereal, brands were only sometimes provided, so we picked standard popular brands from our region (unless the patient stated a particular brand and type/flavor). Some examples of these are in the table below. In some cases, our selections might have underestimated calories in an item for a participant (e.g., we assumed Breyers natural vanilla but the participant actually had Breyers Snickers ice cream which has more calories). In other cases, our selections might have overestimated calories in an item (e.g., we assumed Honey Nut Cheerios but the participant actually had Rice Chex which has fewer calories). This is a limitation to our study, but the choices would suggest that our over- and under-estimates happened equally.
| Food | Brand | Variety |
|---|---|---|
| Pizza | Domino’s © | 14 inch cheese pizza |
| Potato chips | Lays © | Classic potato chips |
| Ice cream | Breyers © | Natural vanilla ice cream |
| Crackers | Keebler © | Club crackers |
| Cereal | General Mills © | Honey Nut Cheerios |
| Cookies | Keebler © | Chips Ahoy chocolate chip |
References
- 1.American Psychiatric Association (2013) Diagnostic and Statistical Manual of Mental Disorders: DSM-5, 5th ed. Arlington VA [Google Scholar]
- 2.Kessler RC, Berglund PA, Chiu WT, et al. (2013) The prevalence and correlates of binge eating disorder in the WHO World Mental Health Surveys. Biol Psychiatry 73:904–914. 10.1016/j.biopsych.2012.11.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Chao AM, Wadden TA, Walsh OA, et al. (2019) Perceptions of a large amount of food based on binge-eating disorder diagnosis. Int J Eat Disord 1–8. 10.1002/eat.23076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Forney KJ, Holland LA, Joiner TE, Keel PK (2015) Determining empirical thresholds for “definitely large” amounts of food for defining binge-eating episodes. Eat Disord 23:15–30. 10.1080/10640266.2014.931763 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fairburn CG, Cooper Z (1993) The eating disorder examination (12th ed.). In: Fairburn CG, Wilson GT (eds) Binge eating: Nature, assessment, and treatment. The Guilford Press, New York, pp 317–360 [Google Scholar]
- 6.Brownstone L, Bardone-Cone A, Fitzsimmons-Craft E, et al. (2013) Subjective and objective binge eating in relation to eating disorder symptomatology, negative affect, and personality dimensions. Int J Eat Disord 46:1–17. 10.1002/eat.22066 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Keel PK, Mayer SA, Harnden-Fischer JH (2001) Importance of size in defining binge eating episodes in bulimia nervosa. Int J Eat Disord 29:294–301. 10.1002/eat.1021 [DOI] [PubMed] [Google Scholar]
- 8.Mond JM, Latner JD, Hay PH, et al. (2010) Objective and subjective bulimic episodes in the classification of bulimic-type eating disorders: Another nail in the coffin of a problematic distinction. Behav Res Ther 48:661–669. 10.1016/j.brat.2010.03.020 [DOI] [PubMed] [Google Scholar]
- 9.Niego SH, Pratt EM, Agras S (1997) Subjective or objective binge: Is the distinction valid? Int J Eat Disord 22:291–298. [DOI] [PubMed] [Google Scholar]
- 10.Palavras MA, Morgan CM, Borges FMB, et al. (2013) An investigation of objective and subjective types of binge eating episodes in a clinical sample of people with co-morbid obesity. J Eat Disord 1:1–7. 10.1186/2050-2974-1-26 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Palavras MA, Hay PJ, Lujic S, Claudino AM (2015) Comparing symptomatic and functional outcomes over 5 years in two nonclinical cohorts characterized by binge eating with and without objectively large episodes. Int J Eat Disord 48:1158–1165. 10.1002/eat.22466 [DOI] [PubMed] [Google Scholar]
- 12.Watson HJ, Fursland A, Bulik CM, Nathan P (2013) Subjective binge eating with compensatory behaviors: A variant presentation of bulimia nervosa. Int J Eat Disord 46:119–126. 10.1002/eat.22052 [DOI] [PubMed] [Google Scholar]
- 13.Kelly NR, Cotter E, Guidinger C (2018) Men who engage in both subjective and objective binge eating have the highest psychological and medical comorbidities. Eat Behav 30:115–119. 10.1016/j.eatbeh.2018.07.003 [DOI] [PubMed] [Google Scholar]
- 14.Beglin SJ, Fairburn CG (1992) What is meant by the term “binge”? Am J Psychiatry 123–124 [DOI] [PubMed] [Google Scholar]
- 15.Telch C, Pratt E, Niego S (1998) Obese women with binge eating disorder define the term binge. Int J Eat Disord 24:313–317 [DOI] [PubMed] [Google Scholar]
- 16.Latner JD, Mond JM, Kelly MC, et al. (2014) The loss of control over eating scale: Development and psychometric evaluation. Int J Eat Disord 47:647–659. 10.1002/eat.22296 [DOI] [PubMed] [Google Scholar]
- 17.Guss JL, Kissileff HR, Devlin MJ, et al. (2002) Binge size increases with body mass index in women with binge-eating disorder. Obes Res 10:1021–1029. 10.1038/oby.2002.139 [DOI] [PubMed] [Google Scholar]
- 18.Under Armour Inc. (2019) Myfitnesspal. www.myfitnesspal.com. Accessed January 27, 2021
- 19.Griffiths C, Harnack L, Pereira MA (2018) Assessment of the accuracy of nutrient calculations of five popular nutrition tracking applications. Public Health Nutr 21:1495–1502. 10.1017/S1368980018000393 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Evenepoel C, Clevers E, Deroover L, et al. (2020) Accuracy of nutrient calculations using the consumer-focused online app MyFitnessPal: Validation study. J Med Internet Res 22:1–9. 10.2196/18237 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fairburn CG (2008) Cognitive Behavior Therapy and Eating Disorders. Guilford Press, New York [Google Scholar]
- 22.Stunkard AJ, Messick S (1985) The Three-Factor Eating Questionnaire to measure dietary restraint, disinhibition, and hunger. J Psychosom Res 29:71–83. 10.1016/0022-3999(85)90010-8 [DOI] [PubMed] [Google Scholar]
- 23.Stunkard AJ, Messick S (1988) Eating Inventory Manual. The Psychological Corporation, San Antonio, Texas [Google Scholar]
- 24.Endicott J, Nee J, Harrison W, Blumenthal R (1993) Quality of life enjoyment and satisfaction questionnaire: A new measure. Psychopharmacol Bull 29:321–326 [PubMed] [Google Scholar]
- 25.Kroenke K, Spitzer RL, Williams JBW (2001) The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med 16:606–613 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.American Psychiatric Association (1994) Diagnostic and Statistical Manual for Mental Disorders: DSM-IV, 4th ed. Washington, DC [Google Scholar]
- 27.Cepeda-Benito A, Gleaves DH, Williams TL, Erath SA (2000) The development and validation of the State and Trait Food-Cravings Questionnaires. Behav Ther 31:151–173 [DOI] [PubMed] [Google Scholar]
- 28.Gearhardt AN, Corbin WR, Brownell KD (2016) Development of the Yale food addiction scale version 2.0. Psychol Addict Behav 30:113–121. 10.1037/adb0000136 [DOI] [PubMed] [Google Scholar]
- 29.IBM Corp (2017) IBM SPSS Statistics for Windows Version 25
- 30.Cohen J (2013) Statistical power analysis for the behavioral sciences. Academic Press [Google Scholar]
- 31.Grilo CM, Masheb RM, Wilson GT (2001) Subtyping binge eating disorder. J Consult Clin Psychol 69:1066–1072. 10.1037/0022-006X.69.6.1066 [DOI] [PubMed] [Google Scholar]
- 32.Grilo CM, Masheb RM, Wilson GT (2001) Different methods for assessing the features of eating disorders in patients with binge eating disorder: A replication. Obes Res 9:418–422. 10.1038/oby.2001.55 [DOI] [PubMed] [Google Scholar]
- 33.Grilo CM, Masheb RM, Wilson GT (2001) A comparison of different methods for assessing the features of eating disorders in patients with binge eating disorder. J Consult Clin Psychol 69:317–322. 10.1038/oby.2001.55 [DOI] [PubMed] [Google Scholar]
- 34.Goldfein JA, Devlin MJ, Kamenetz C (2005) Eating Disorder Examination-Questionnaire with and without instruction to assess binge eating in patients with binge eating disorder. Int J Eat Disord 107–111. 10.1002/eat.20075 [DOI] [PubMed] [Google Scholar]
- 35.Barnes RD, Masheb RM, White MA, Grilo CM (2011) Comparison of methods for identifying and assessing obese patients with binge eating disorder in primary care settings. Int J Eat Disord 44:157–163. 10.1038/jid.2014.371 [DOI] [PMC free article] [PubMed] [Google Scholar]