Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Dec 16.
Published in final edited form as: Int J Eat Disord. 2011 May;44(4):311–316. doi: 10.1002/eat.20821

Clinical Significance and Distinctiveness of Purging Disorder and Binge Eating Disorder

Pamela K Keel 1,*, Jill M Holm-Denoma 2, Ross D Crosby 3,4
PMCID: PMC8674752  NIHMSID: NIHMS1762688  PMID: 20354996

Abstract

Objective:

In the DSM-IV, individuals with binge eating disorder (BED) and those with purging disorder (PD) receive a diagnosis of eating disorder not otherwise specified (EDNOS), suggesting no meaningful differences between clinical presentations. This article compares PD and BED on eating disorder severity and comorbid disorders.

Method:

Individuals with PD (n = 33), DSM-IV BED (n = 23 with BMI >30 kg/m2, and n = 18 with BMI between 18.5 and 26.5 kg/m2), and noneating disorder controls (n = 35) completed SCID-I interviews and questionnaires.

Results:

Eating disorder groups reported significantly greater depression, body dissatisfaction, and dietary restraint and more Axis I disorders compared with controls. Compared with both the obese and normal weight BED groups, PD reported significantly greater dietary restraint and body dissatisfaction. Compared with obese BED, PD reported lower prevalence of impulse control disorders.

Discussion:

Findings support differentiating among EDNOS based on behavioral presentation in both research and future nosological schemes such as the DSM-V.

Keywords: binge eating disorder, purging disorder, nosology

Introduction

The core behavioral feature of binge eating disorder (BED) is recurrent objectively large binge episodes in the absence of inappropriate compensatory behaviors.1 The core behavioral feature of purging disorder (PD) is recurrent purging in the absence of objectively large binge episodes.2,3 Despite no overlap in behavioral features, individuals with BED and those with PD would receive the same diagnosis of eating disorder not otherwise specified (EDNOS) in the DSM-IV, suggesting no meaningful differences between these clinical presentations. Following this lead, papers (e.g.,4–6) reporting comparisons between bulimia nervosa (BN) and EDNOS often collapse all EDNOS into one broad category. In some instances, this reflects inadequate power to test for possible heterogeneity among different forms of EDNOS (e.g.,5) which include not only partial forms of BN but also partial forms of AN. However, in other instances, it may reflect a presumption that partial BN syndromes are interchangeable. To our knowledge, only two studies7,8 have directly compared individuals with PD to individuals with BED to examine evidence of distinctiveness based on clinical presentation. One of these studies7 was marked by small sample sizes, resulting in low statistical power, and the other study8 relied on self-report measures obtained through an on-line survey. The current study utilized a combination of self-report and structured clinical interview assessments to compare PD and BED with the goal of determining whether these behaviorally distinct syndromes differ on eating disorder severity and comorbid disorders.

Several factors support the potential for BED and PD to demonstrate clinical distinctiveness. First, significant differences were found between BED and PD on levels of dietary restraint, disinhibition, depression,8 as well as anxiety.7 Second, latent structure studies have found that syndromes resembling BED fall into a separate latent group from syndromes that resemble PD.3 In one of the earliest latent class analyses of individuals reporting bulimic symptoms, Sullivan et al.9 found a latent group resembling BED with objectively large binge episodes and a latent group in which a majority of individuals purged but did not binge. Striegel-Moore et al.10 replicated these findings using a similar method of sample selection and analytic approach. Employing latent profile analyses in community-based samples, three studies11–13 found that individuals with PD fell into a large purging group while individuals with BED fell into their own latent group. Thus, the preponderance of evidence from latent structure studies supports placing PD and BED in separate groups rather than both falling within a single diagnosis of EDNOS. However, an important caveat is that results could reflect either behavioral distinctions between BED and PD or differences in associated characteristics, such as body mass index (BMI), given that BED is commonly associated with obesity but not defined by it.

The purpose of this study was to compare individuals with BED to individuals with PD on indicators of distress, eating disorder severity, and comorbid disorders using a combination of self-report and structured clinical interview assessments. Given the potential influence of differences in BMI between BED and PD, we recruited individuals who met DSM-IV criteria for BED with BMI between 18.5 and 26.5 kg/m2, matching criteria for recruitment of individuals with PD and noneating disorder controls. In addition, we recruited individuals with BED who were obese (i.e., BMI >30 kg/m2) to reflect populations studied in most previous studies of BED. This approach allowed us to determine the role of BMI in explaining differences between PD and BED. In addition, this made it possible to examine the role of obesity in contributing to the clinical significance of BED in comparison to individuals without eating disorders. The current study advanced previous comparisons of PD and BED by including larger numbers of individuals with PD and in-person assessments that included structured clinical interviews.

Method

Participants

Women with DSM-IV BED (n = 41) and women with PD (n = 33) based on previously published criteria,2 and noneating disorder female controls (n = 35) were recruited through community-based advertisements for studies on eating disorders. Because BED is frequently associated with obesity, and PD has been identified in individuals of normal weight, we specifically recruited individuals with BED who fell into one of two BMI ranges: 18.5–26.5 kg/m2 (BED normal weight; BEDnw; n = 18) and >30 kg/m2 (BED obese, BEDob; n = 23). The BMI range for the normal weight group was selected to match that used for recruitment of women in the PD and control groups. Reflecting these selection criteria, there was a significant group effect on BMI (F(3, 105) = 133.16, p < .001) that was explained by significantly greater BMI in the BEDob group (Mean = 39.81 ± 7.55 kg/m2) compared with the BEDnw (22.51 ± 1.99), PD (22.11 ± 1.14), and Control (21.41 ± 1.93) groups, which did not differ significantly from one another.

Procedures and Measures

Participants completed structured clinical interviews and self-report questionnaires during a single in-person assessment. During this assessment, height and weight were objectively measured using a digital scale and stadiometer.

Structured Clinical Interview for Axis I Disorders (SCID-I)14 was used to assess lifetime and current Mood, Anxiety, Substance Use, and Impulse Control disorders. The SCID-I was modified to assess presence of eating disorders, including BED and PD, by not following standard skip rules. Inter-rater reliability of interview assessments was high with kappa ranging from 0.74 to 1.00 across diagnoses, mean kappa = 0.87.

Beck Depression Inventory (BDI)15 was used to assess levels of depression and personal distress. The BDI has excellent internal consistency reliability, test-retest reliability, and good criterion validity.15 Chronbach’s alpha in the current study was 0.93.

Body Shape Questionnaire (BSQ)16 was used to assess feelings and attitudes about weight and shape that characterize individuals with eating disorders. The BSQ has good internal reliability and concurrent validity.16 Chronbach’s alpha in the current study was 0.97.

Three Factor Eating Questionnaire (TFEQ)17 Cognitive Restraint was used to assess intentions to restrict food intake. The TFEQ Cognitive Restraint scale has good internal consistency, test-retest reliability, and criterion validity.17 Chronbach’s alpha in the current study was 0.92.

Data Analyses

Parametric analyses (ANOVA and ANCOVA) were used to compare groups on continuous measures, and non-parametric analyses (chi-square) were used to examine associations between group membership and presence versus absence of other Axis I disorders. A Bonferoni-adjusted p-value of 0.0083 was used to evaluate statistical significance of post-hoc comparisons to control for family-wise error rate.

Results

Table 1 presents comparisons among groups on continuous measures including age, BDI, BSQ, and TFEQ Cognitive Restraint scale scores. Groups differed significantly on age. The BEDob group was significantly older than the PD and Control groups, and the BEDnw group was significantly older than the Control group. Because age was significantly correlated with BDI (r(150) = .31, p < .001) and BSQ (r(147) = .30, p < .001) but not TFEQ Restraint (r(150) = −0.06, p = .46) scale scores, models for BDI and BSQ included age as a covariate.

TABLE 1.

Comparisons among subgroups on continuous measures

BEDob BEDnw PD Control
N = 23 N = 18 N = 33 N = 35
M (SD) M (SD) M (SD) M (SD) F (3, 105) p
Age 36.30 (10.46)a 30.50 (8.83)a,b 25.27 (6.70)b,c 23.89 (4.78)c 14.94 <.001
Beck depression Inventory 17.54 (10.52)a 14.44 (14.18)a 11.58 (8.36)a 1.94 (2.45)b 10.16 <.001
Body shape Questionnaire 133.86 (28.39)a,b 116.37 (30.06)a 137.14 (27.99)b 51.57 (13.14)c 73.82 <.001
TFEQ Restraint 6.93 (3.98)a 10.11 (4.38)a 16.48 (4.10)b 3.25 (2.76)c 73.35 <.001
Open in a new tab

Superscripts that differ reflect significant differences between groups with p < .0083.TFEQ, three factor eating questionnaire.

An ANCOVA with age as a covariate demonstrated that all three eating disorder groups reported significantly greater levels of depression on the BDI than the Control group, but no significant differences were found among the eating disorder groups (see Table 1). Similarly, all three eating disorder groups reported significantly greater body image disturbance on the BSQ compared with the control group. In addition, women with PD reported significantly greater body image disturbance compared with women with BED who were within a healthy weight range. Finally, all three eating disorder groups reported significantly greater dietary restraint on the TFEQ compared to controls, and women with PD reported significantly greater dietary restraint compared to women with BED, regardless of their weight group. No significant differences were found between the BEDob and BEDnw groups with regard to depression, body image disturbance, or dietary restraint.

Table 2 presents comparisons among groups on Axis I disorders. Women with eating disorders reported significantly higher prevalence of lifetime and current mood and anxiety disorders as well as lifetime impulse control disorders compared with controls. Women in the BEDob and PD groups reported significantly greater lifetime history of substance use disorders and current impulse control disorders compared with controls whereas prevalence in the BEDnw group was intermediate and did not differ significantly from the other eating disorder groups or controls. Of interest, the BEDob group reported significantly greater lifetime history of impulse control disorders compared with the PD group. Finally, no significant differences were found between the BEDob and BEDnw groups on comorbid Axis I disorders.

TABLE 2.

Comparisons among subgroups on comorbid axis I disorders

BEDob BEDnw PD Control
N = 23 N = 18 N = 33 N = 35
% (n) % (n) % (n) % (n) χ2 (3) p
Lifetime
 Mood 96 (22)a 67 (12)a 79 (26)a 23 (8)b 38.08 <.001
 Anxiety 70 (16)a 61 (11)a 45 (15)a 14 (5)b 20.87 <.001
 Substance use 70 (16)a 50 (9)a,b 61 (20)a 17 (6)b 19.75 <.001
 Impulse control 65 (15)a 33 (6)a,b 27 (9)b 0 (0)c 29.98 <.001
Current
 Mood 26 (6)a 28 (5)a 33 (8)a 0 (0)b 11.11 <.001
 Anxiety 61 (14)a 50 (9)a 34 (11)a 6 (2)b 22.13 <.001
 Substance use 0 (0) 0 (0) 12 (4) 0 (0) 9.44 <.025
 Impulse control 39 (9)a 17 (3)a,b 21 (7)a 0 (0)b 15.25 <.0025
Open in a new tab

Superscripts that differ reflect significant differences between groups with p < .0083.

Discussion

Results supported the clinical significance of all three eating disorders based on the presence of higher levels of depression and greater levels of Axis I disorders compared with control participants. Results further supported several distinctions between PD and BED. Specifically, women with PD reported higher body image disturbance compared with normal weight BED despite no objective difference in BMI and higher dietary restraint compared with both BED groups. Differences in dietary restraint replicate findings from previous comparisons of PD and BED.8 Findings are also consistent with the different behavior patterns observed between PD and BED in which PD is characterized by the use of purging to influence weight or shape and the undue influence of weight and shape on self-evaluation—features that are not required for a diagnosis of BED. Whereas previous studies reported greater depression8 and anxiety7 in individuals with BED compared with those with PD, we did not find significant differences among our groups. The failure to replicate these findings may reflect the division of our BED participants into separate weight groups, resulting in smaller group sizes for comparisons and lower statistical power. Despite this, obese BED was associated with higher rates of lifetime impulse control disorders compared with PD. Of interest, although there were significant differences in BMI between the obese and normal weight BED groups, no significant differences were found between groups on any psychological measure.

Simple inspection of the examples of EDNOS provided in the DSM-IV suggest considerable heterogeneity in this group, and the current results support that this behavioral heterogeneity is associated with clinically meaningful differences that may underlie the presence of binge-eating versus purging behaviors. Specifically, although dietary restraint has been long associated with risk for binge eating,18 this feature was most elevated in the PD group. Potentially, strict rules regarding what, when, and how much they should eat leave individuals with PD particularly vulnerable to violation of dietary rules that then trigger purging in the absence of objectively large binge episodes. Body shape concerns also were elevated in PD compared with the normal weight BED group, suggesting a close link between body image disturbance and purging when BMI is within a healthy range. In contrast, the very high levels of body image disturbance observed in obese BED may reflect the presence of objective overweight in this group. Importantly, the elevation of body image disturbance in both the obese and nonobese BED groups relative to controls suggests that body image disturbance may be a valuable addition to the diagnostic criteria for BED.19

Differences between PD and obese BED on lifetime impulse control disorders resemble differences found between PD and BN on impulsivity20 and disinhibition.2,21 Moreover, difficulties controlling impulses in general and difficulties controlling food intake in particular may explain the propensity to binge in disorders characterized by objectively large binge episodes. Of note, these differences were not found for comparisons of PD and normal weight BED suggesting the possibility that body weight serves as an indicator of severity of binge eating. Importantly, like differences between BED and control participants on body image disturbance, differences in impulse control disorders were observed between PD and controls. Thus, while body image concerns and problems with mood, anxiety, and regulation of impulses may increase the risk for having an eating disorder in general, elevations on specific vulnerability domains may contribute to which disorder of eating an individual develops.

Of interest, across multiple measures, the normal weight BED group looked somewhat healthier than the obese BED group; however, none of these reached statistical significance. This could reflect limited statistical power given that the study was adequately powered to detect only large effect sizes for post-hoc comparisons that controlled for family-wise error rate. However, every difference found between obese BED and controls was also observed between normal weight BED and controls, indicating that the clinical significance of BED is attributable to its behavioral features rather than the influence of these behaviors on BMI. Such results could have important implications for understanding the clinical significance of BED independently from the clinical significance of obesity.

This study had several strengths. We found evidence of clinically significant differences between syndromes that might otherwise be considered partial syndromes of BN. Both normal weight BED and obese BED groups were recruited in order to determine whether potential group differences reflected alterations in eating patterns or body weight associated with BED. Participants were recruited from the community, reducing the potential influence of Berkson’s Bias22 when making comparisons on distress and presence of other Axis I disorders. Finally, diagnoses were based on structured clinical interviews associated with high inter-rater reliability in this study and comparisons utilized data from both self-report and interview-based assessments.

Despite its strengths, the current study had certain weaknesses that merit consideration. First, the sample size for the BED subgroups was too small to detect differences associated with small-to-moderate effect sizes. Inspection of mean values suggested the potential for meaningful differences between obese and normal-weight individuals with BED. Importantly, however, significant differences in present distress, dietary restraint, body image disturbance and Axis I disorders were evident between normal-weight individuals with BED and noneating disorder controls supporting the conclusion that the clinical significance of BED is associated with disturbances in eating behavior rather than weight dysregulation commonly associated with BED. Second, analyses were restricted to women, raising questions about the generalizability of findings to men. This is a particular concern for findings with BED given that men are better represented among individuals with this syndrome.23 However, PD is significantly more likely to occur in women than in men,24 and recruiting adequate numbers of men with PD to control for gender differences in comparisons would have required resources beyond those available for the current investigation. Finally, comparisons are based on a cross-sectional design. Thus, findings contribute to the development of hypotheses regarding the influence of dietary restraint, body image disturbance, and impulse control on the development of binge eating versus purging behaviors; however, they do not allow adequate testing of these hypotheses.

Findings support differentiating among EDNOS based on behavioral presentation in both research and future nosological schemes such as the DSM-V. This approach is consistent with that used in recent epidemiological studies25,26 but differs from that used in longitudinal studies4,5 that lump different forms of EDNOS together when describing clinical course and outcome of eating disorders. The DSM-V could facilitate the collection of data to describe course and outcome for PD much in the same way that the DSM-IV facilitated collection of data to describe the course and outcome of BED when it included a set of provisional diagnostic criteria for the syndrome. Such efforts would support the collection of information on clinical utility and validity of including current forms of EDNOS as new disorders or expanding criteria of existing syndromes to encompass various EDNOS in future editions of the DSM.

Footnotes

Portions of this paper were presented at the 2009 Eating Disorders Research Society Meeting in Brooklyn, New York.

References

  • 1.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders,4th ed. Text Revision (DSM-IV-TR). Washington, D.C.: American Psychiatric Association, 2000. [Google Scholar]
  • 2.Keel PK, Haedt A, Edler C. Purging disorder: An ominous variant of bulimia nervosa? Int J Eat Disord 2005;38:191–199. [DOI] [PubMed] [Google Scholar]
  • 3.Keel PK, Striegel-Moore RH. The validity and clinical utility of Purging Disorder. Int J Eat Disord 2009;42:706–719. [DOI] [PubMed] [Google Scholar]
  • 4.Grilo CM, Sanislow CA, Shea MT, Skodol AE, Stout RL, Yen S, et al. The natural course of bulimia nervosa and eating disorder not otherwise specified is not influenced by personality disorders. Int J Eat Disord 2003;34:319–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Keel PK, Gravener JA, Joiner TE Jr, Haedt AA. Twenty-year follow-up of bulimia nervosa and related eating disorders not otherwise specified. Int J Eat Disord 2009. August 29; [Epub ahead of print (DOI: 10.1002/eat.20743)]. [DOI] [PubMed] [Google Scholar]
  • 6.Kendler KS, MacLean C, Neale M, Kessler R, Heath A, Eaves L. The genetic epidemiology of bulimia nervosa. Am J Psychiatry 1991;148:1627–1637. [DOI] [PubMed] [Google Scholar]
  • 7.Fink EL, Smith AR, Gordon KH, Holm-Denoma JM, Joiner TE Jr. Psychological correlates of purging disorder as compared with other eating disorders: An exploratory investigation. Int J Eat Disord 2009;42:31–39. [DOI] [PubMed] [Google Scholar]
  • 8.Roberto CA, Grilo CM, Masheb RM, White MA. Binge eating, purging, or both: Eating disorder psychopathology findings from an internet community survey. Int J Eat Disord 2009. October 27; [Epub ahead of print (DOI: 10.1002/eat.20770)]. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Sullivan PF, Bulik CM, Kendler KS. The epidemiology and classification of bulimia nervosa. Psychol Med 1998;28:599–610. [DOI] [PubMed] [Google Scholar]
  • 10.Striegel-Moore RH, Franko DL, Thompson D, Barton B, Schreiber GB, Daniels SR. An empirical study of the typology of bulimia nervosa and its spectrum variants. Psychol Med 2005;35:1563–1572. [DOI] [PubMed] [Google Scholar]
  • 11.Wade TD, Crosby R, Martin NG. Use of latent profile analysis to identify eating disorder phenotypes in an adult Australian twin cohort. Arch Gen Psychiatry 2006;63:1377–1384. [DOI] [PubMed] [Google Scholar]
  • 12.Pinheiro AP, Bulik CM, Sullivan PF, Machado PP. An empirical study of the typology of bulimic symptoms in young Portuguese women. Int J Eat Disord 2008;41:251–258. [DOI] [PubMed] [Google Scholar]
  • 13.Keel PK, Fichter M, Quadflieg N, Bulik CM, Baxter MG, Thornton L, et al. Application of a latent class analysis to empirically define eating disorder phenotypes. Arch Gen Psychiatry 2004;61:192–200. [DOI] [PubMed] [Google Scholar]
  • 14.First M, Gibbons M, Spitzer RL, Williams JBW, Benjamin LS. Structured Clinical Interview for DSM-IV Axis II Personality Disorders (SCID-II). Washington, D.C.: American Psychiatric Press, 1997. [Google Scholar]
  • 15.Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4: 561–571. [DOI] [PubMed] [Google Scholar]
  • 16.Cooper P, Taylor MJ, Cooper Z, Fairburn CG. The development and validation of the body shape questionnaire. Int J Eat Disord 1987;6:485–494. [Google Scholar]
  • 17.Stunkard AJ, Messick S. The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. J Psychosom Res 1985;29:71–83. [DOI] [PubMed] [Google Scholar]
  • 18.Polivy J, Herman CP. Dieting and binging. A causal analysis. Am Psychol 1985;40:193–201. [DOI] [PubMed] [Google Scholar]
  • 19.Grilo CM, Masheb RM. Correlates of body image dissatisfaction in treatment-seeking men and women with binge eating disorder. Int J Eat Disord 2005;38:162–166. [DOI] [PubMed] [Google Scholar]
  • 20.Keel PK, Mayer SA, Harnden-Fischer JH. Importance of size in defining binge eating episodes in bulimia nervosa. Int J Eat Disord 2001;29:294–301. [DOI] [PubMed] [Google Scholar]
  • 21.Keel PK, Wolfe BE, Liddle RA, DeYoung KP, Jimerson DC. Clinical features and physiological response to a test meal in purging disorder and bulimia nervosa. Arch Gen Psychiatry 2007;64: 1058–1066. [DOI] [PubMed] [Google Scholar]
  • 22.Berkson J Limitations of application of fourfold table analysis to hospital data. Biometrics 1946;2:47–53. [PubMed] [Google Scholar]
  • 23.Hudson JI, Hiripi E, Pope HG Jr, Kessler RC. The prevalence and correlates of eating disorders in the national comorbidity survey replication. Biol Psychiatry 2007;61:348–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Haedt AA, Keel PK. Comparing definitions of purging disorder on point prevalence and associations with external validators. Int J Eat Disord 2009. June 17; [Epub ahead of print (DOI: 10.1002/eat.20712)]. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Favaro A, Ferrara S, Santonastaso P. The spectrum of eating disorders in young women: A prevalence study in a general population sample. Psychosom Med 2003;65:701–708. [DOI] [PubMed] [Google Scholar]
  • 26.Wade TD, Bergin JL, Tiggemann M, Bulik CM, Fairburn CG. Prevalence and long-term course of lifetime eating disorders in an adult Australian twin cohort. Aust N Z J Psychiatry 2006;40: 121–128. [DOI] [PubMed] [Google Scholar]

RESOURCES