Objective and subjective binge eating in underweight eating disorders: Associated features and treatment outcome




To define the utility of the DSM-IV-TR definition of binge eating, as it applies to anorexia nervosa (AN) and underweight eating disorder not otherwise specified (ED-NOS).


We investigated the psychopathological features associated with bulimic episodes in 105 underweight individuals with eating disorders who reported regular objective bulimic episodes with or without subjective bulimic episodes (OBE group, n = 33), regular subjective bulimic episodes only (SBE group, n = 36) and neither objective nor subjective bulimic episodes (n = 36, no-RBE group). The Eating Disorder Examination (EDE), anxiety, depression, and personality tests were administered before and upon completion of inpatient cognitive behavior therapy (CBT) treatment 6 months later.


Compared with the SBE group, OBE subjects had higher body mass index, and more frequent self-induced vomiting, while both OBE and SBE groups had more severe eating disorder psychopathology and lower self-directness than the no-RBE group. Dropout rates and outcomes in response to inpatient CBT were similar in the three groups.


Despite a few significant differences at baseline, the similar outcome in response to CBT indicates that categorizing patients with underweight eating disorder on the basis of the type or frequency of bulimic episodes is of limited clinical utility. © 2011 by Wiley Periodicals, Inc. (Int J Eat Disord 2012;)


Binge eating is a cardinal criterion for the diagnosis of bulimia nervosa (BN) and the provisional diagnosis of binge eating disorder (BED),1 but it is also reported by individuals with anorexia nervosa (AN) of the binge-eating/purging type (AN-BP).1

According to the DSM-IV-TR, an episode of binge eating is characterized both by eating an amount of food that would normally be regarded as too large, and by a sense of lack of control over eating.1 Unfortunately, there are no specific guidelines regarding the threshold of the amount of food necessary to satisfy the inclusion in the AN-BP type. This causes difficulties in classifying AN-BP, especially when clinicians see underweight patients who report regular episodes of loss of control over eating but do not have regular episodes of purging. In addition, individuals with eating disorders may fail to identify the binge episodes on the basis of food amount, giving greater emphasis to the loss of control than to the amount of food in their definition of binge eating.2, 3

The evidence supporting the validity of the DSM-IV-TR definition of binge eating as this applies to BN and variants of BN is limited.4–6 In patients with BN, “objective” and “subjective” bulimic episodes, as defined according to the Eating Disorder Examination (EDE),7 seem to be associated with a similar severity of clinical impairment and/or other clinical data (e.g., psychopathology, past-history of AN, psychiatric symptoms, interpersonal distress, low self-esteem and self-efficacy, and social adjustment).8 Other clinical studies confirmed that objective and subjective bulimic episodes are associated with similar levels of depression or anxiety,9, 10 and that the age of onset and the predictors of the two types of bulimic episodes are almost identical,11 while their frequency is different.12 Findings from community-based studies similarly suggest that individuals with bulimic-type eating disorders who report objective bulimic episodes are similar, in terms of eating disorder and comorbid psychopathology, to those who report subjective bulimic episodes.13–15

The possible effects of the type and frequency of bulimic episode on treatment response have rarely been considered. The limited data available have suggested that subjective episodes remit more slowly in patients with BED16 and both persist at the end of treatment in patients with BN.17 Subjective episodes did not respond well to self-monitoring in women with either BN or BED,18 whereas this strategy was effective in reducing objective episodes.19 As to the influence on treatment outcome, only the presence of subjective episodes predicted the response to placebo in participants with BED.20

As pointed out in a recent review,6 no meaningful information has ever been reported on the binge eating episodes in underweight ED-NOS and AN. Limited information is also available on the personality characteristics of the underweight individuals with different types of bulimic episodes and on their role on treatment response, a key aspect to define the clinical utility of a DSM diagnosis.21 The goal of this study was, therefore, to investigate the psychopathological features associated with objective and subjective bulimic episodes in underweight individuals with eating disorder, and to test their role on treatment outcome.



The sample consisted of consecutive female subjects admitted to the eating disorder inpatient unit of Villa Garda Hospital (Northern Italy) between November 2003 and November 2006. The subjects were referred to our institution from all over Italy by general practitioners or by eating disorder specialists.

Subjects were included if they met the following criteria: (a) age, 12–65 years; (b) body mass index (BMI) ≤17.5 kg/m2; (c) diagnosis of an eating disorder of clinical severity assessed by EDE; (d) failure of less intensive outpatient treatment or an eating disorder of clinical severity not manageable in an outpatient setting. According to our protocol, subjects with active substance abuse and acute psychotic disorders are not considered for hospital admission and the first author (RDG) evaluated the presence of these two comorbid conditions during an eligibility interview before admission.

The research was reviewed and approved by the Institutional Review Board of Villa Garda Hospital, Verona, and all participants (or their legal guardians for the 17 patients under 18) gave written informed consent to the anonymous use of personal data.

Inpatient Treatment Protocol

The treatment has been described in detail elsewhere.22, 23 The program is derived from the cognitive behavior treatment of eating disorders (CBT-E),24 but has been adapted for an inpatient setting. The treatment is manual-based,22 lasts 20 weeks and comprises 13 weeks of inpatient therapy followed by 7 weeks of day-hospital admission, with the patient living close to the hospital and spending the weekends at home.

Assessment and Measures

All data were collected on the first day of admission and on the last day of day-hospital treatment. Demographic and clinical variables were assessed during interview. Weight (to the nearest 0.1 kg) was measured by a calibrated scale and height (to the nearest 0.5 cm) by a stadiometer. Patients were measured with underwear and without shoes.

A validated Italian translation of EDE (EDE12.0D),7, 25 completed by a senior specialist in the field (RDG), was used to generate the operational definition of eating disorder diagnosis according to DSM-IV and to evaluate the eating disorder psychopathology, including bulimic episodes. The EDE inter-rater reliability has been estimated to be 0.97–0.99.26 The EDE section about “bulimic episodes and other episodes of overeating” produces an accurate measure of objective and subjective bulimic episodes, as it is an investigator-based interview where bulimic episodes are classified as objective or subjective according to specific guidelines after obtaining a detailed report of the amount of food consumed during each episode.8 For the purpose of this study, bulimic episodes were defined “regular” if they occurred at least once a week in the previous 4 weeks, and “nonregular” if they were recorded less than once a week.27, 28

As detailed below, the number of participants reporting in the EDE interview regular objective bulimic episodes but not regular subjective bulimic episodes was small. We therefore conducted a preliminary analysis in which the characteristics of participants who reported only objective bulimic episodes (n = 14) were compared with those of participants who reported both objective and subjective bulimic episodes (n = 19). Results of this analysis failed to reveal any difference in the clinical characteristics in relation to the co-presence of subjective bulimic episodes, with the exception of the eating concern scale of EDE, higher in subjects with subjective episodes (4.6 ± 1.1 vs. 3.6 ± 1.1; p = .012). We therefore took the decision to group participants who reported objective but not subjective bulimic episodes with those who reported both objective and subjective bulimic episodes (OBE group). Then, we compared the OBE group with the participants who reported regular subjective (but not objective) bulimic episodes (SBE group) and with those who reported neither objective nor subjective bulimic episodes (no-RBE group).

The Temperament and Character Inventory (TCI)29 was used to assess personality characteristics, based on a psychobiologic model of personality. This model includes four temperament dimensions (novelty seeking, harm avoidance, reward dependence, and persistence) and three character dimensions (self-directedness, cooperativeness, and self- transcendence). The TCI has good internal consistency,29, 30 inter-rater and test–retest reliability,29 and it is validated in its Italian version.31

The Beck Depression Inventory (BDI)32 and the State-Trait Anxiety Inventory (STAI Form Y-1)33 were used to assess the presence and severity of depression and trait levels of anxiety, respectively. Both inventories have excellent internal reliability, good test–retest reliability, good criterion validity32, 34 and have been validated in their Italian versions.35, 36

Statistical Analyses

Statistical analyses were carried out by means of SPSS Version 15.0 (SPSS, Chicago). Continuous variables were categorized as mean ± SD or as median [interquartile range, IQR] and categorical variables as frequency and percentage. The differences in demographic and clinical variables between the groups were tested for significance by means of ANOVA, Kruskal Wallis test, or χ2 test, as appropriate. Repeated-measures analysis of variance for continuous variables or McNemar test for categorical variables was used to analyze changes in clinical variables between groups. Spearman's correlation was used to analyze the association between the number of objective and subjective bulimic episodes and other eating disorder behaviors and psychopathology scores.

Finally, we tested the independent association between the frequency of both types of bulimic episodes at baseline (independent variables) and the changes in EDE global score or BMI (dependent variables) following treatment in two different linear regression models. Both models were adjusted for initial BMI, age, frequency of self-induced vomiting, laxative and diuretic misuse, and intense exercising to control shape or weight, considered potential predictors of outcome. Change in BMI was added as additional dependent variable in the EDE model. BMI and EDE scores were used as outcome variables in the regression analysis as weight gain is the primary goal for AN patients,37 and improved EDE-measured eating disorder psychopathology is the main target of inpatient cognitive behavioral intervention.



The participants were 105 underweight eating disorder patients (mean age, 26.0 ± 9.0 years; mean BMI, 14.6 ± 1.6 kg/m2); 66 (62.9%) were classified as AN, and 39 as ED-NOS. According to the definitions described in the Methods section, 33 participants were classified in the OBE group, 36 in the SBE group, and 36 in the no-RBE group. In the OBE group, 19 (57.6%) had a diagnosis of AN, and 14 were ED-NOS, while in the SBE group, 27 (75%) were AN and 9 ED-NOS, and in the no-RBE group, 20 (55.6%) were AN and 16 were ED-NOS (χ2 = 3.49, p = .175)

Clinical Characteristics of Participants at Baseline by Subgroup

Individuals with regular bulimic episodes, and particularly those with both types of episodes, were characterized by significantly higher BMI, a higher number of episodes of self-induced vomiting and higher EDE eating concern subscale and global scores, as well as lower scores on the self-directedness scale of TCI. There were no baseline differences between groups on any of the remaining study variables (Table 1).

Table 1. Descriptive statistics for age, body weight and measures of eating disorder and comorbid psychopathology in patients with regular objective with/without subjective bulimic episodes (OBE group), with regular subjective without subjective bulimic episodes (SBE group) and without regular bulimic episodes (no-RBE group)
 OBE Group (N = 33)SBE Group (N = 36)No-RBE Group (N = 36)Testep Value
  • a

    p < .05 vs. OBE group.

  • Data are presented as mean (SD), as

  • b

    median [interquartile range] or as

  • c

    number of cases (%).

  • d

    In the last 28 days before examination.

  • e

    ANOVA, Kruskal Wallis, χ2 or Fisher's exact test as necessary.

  • f

    p < .05 vs. SBE group

Present BMI (kg/m2)15.4 (1.6)14.0 (1.3)a14.4 (1.7)a8.140.001
Maximum BMI (kg/m2)22.4 (3.7)20.7 (4.0)19.5 (2.5)a6.090.003
Minimum BMI (kg/m2)13.9 (1.7)13.1 (1.5)13.4 (1.7)2.310.105
Pre-morbid BMI (kg/m2)21.3 (3.9)20.2 (3.8)19.0 (2.0)a3.960.022
Menarche (years)13.0 (2.4)12.4 (1.7)12.7 (1.7)0.670.513
Age (years)27.9 (6.9)25.8 (10.5)24.3 (9.0)1.520.225
Age at onset (years)16.6 (3.8)16.0 (6.7)17.9 (5.7)0.950.391
ED duration (months)b129 [138]66 [168]30 [425]a8.270.016
Suicidal attemptsc7 (21.2%)8 (22.9%)4 (11.1%)1.920.383
Eating Disorder Examinationd
  Objective bulimic episodesc33 (100%)00--
  Subjective bulimic episodesc19 (57.6%)36 (100%)0-<0.001
  Self-induced vomiting episodec27 (81.8%)8 (22.2%)3 (8.3%)a,f44.89<0.001
  Laxative misuse episodesc10 (30.3%)5 (13.9%)4 (11.1%)4.680.085
  Diuretics misuse episodesc3 (9.1%)3 (8.3%)1 (2.8%)1.350.510
  Intense exercising episodesc12 (36.4%)21 (58.3%)20 (55.6%)3.890.143
  Restraint3.5 (1.7)3.9 (1.5)3.5 (1.7)0.630.532
  Eating concern4.1 (1.2)3.4 (1.4)2.5 (1.3)a,f12.71<0.001
  Weight concern3.6 (1.5)3.8 (1.6)2.9 (1.7)3.070.051
  Shape concern3.7 (1.0)3.9 (1.2)3.1 (1.3)f4.530.013
 Global score3.7 (1.0)3.7 (1.2)3.0 (1.3)a,f4.600.012
State-Trait Anxiety Inventory (Form Y)59.3 (11.1)60.5 (12.5)55.7 (16.8)1.140.325
Beck Depression Inventory30.8 (15.7)31.8 (12.5)27.2 (13.1)1.070.346
Temperament and Character Inventory
 Novelty Seeking19.3 (6.7)16.1 (4.6)17.1 (5.4)2.850.063
 Harm avoidance21.4 (6.4)23.0 (5.7)22.3 (7.1)0.500.611
 Reward dependence14.6 (4.3)14.9 (3.7)15.3 (3.0)0.3220.725
 Persistence4.9 (1.6)5.3 (1.8)5.7 (1.8)1.950.147
 Self-directedness19.0 (8.5)18.4 (6.7)24.4 (8.0)a,f6.330.003
 Cooperativeness30.8 (6.1)28.4 (7.0)31.9 (5.4)2.930.058
 Self-transcendence13.6 (7.2)12.9 (5.6)13.9 (5.3)0.240.787

Associations between OBE and SBE Frequency and Eating Disorder Features

The number of OBEs was positively associated with the number of episodes of self-induced vomiting and of laxative misuse and with the EDE eating concern subscale, and negatively with the episodes of intense exercising. The number of subjective episodes was positively associated with the number of episodes of self-induced vomiting, the eating, weight, and shape concern subscales of EDE and with the EDE global score (Table 2).

Table 2. Spearman's correlation coefficients (rho values) between the frequency of objective and subjective bulimic episodes, and frequency of extreme weight-control behaviors and scores on subscales of the Eating Disorder Examination
Eating Disorder ExaminationNo. of Objective Bulimic EpisodesNo. of Subjective Bulimic Episodes
  • a

    p < .01

  • b

    p < .05.

No. of self-induced vomiting episodes0.669a0.299a
No. of laxative misuse episodes0.227b0.136
No. of diuretic misuse episodes0.0640.173
No. of intense exercising episodes−0.253a0.094
Restraint score−0.0360.039
Eating concern score0.405a0.403a
Weight concern score0.0260.234a
Shape concern score0.0120.293a
Global score0.0890.287a

Treatment Outcome of OBE, SBE, and No-RBE Groups

73 patients (69.5%) completed the treatment program (continuers), whereas 32 (30.5%) left the program before concluding the planned 20 weeks (dropouts). They had a similar BMI at baseline (14.7 ± 1.7 kg/m2 in completers and 14.2 ± 1.6 in dropouts; t = 1.66, p = .099). The dropout rate was similar in relation to bulimic episodes (OBE: 11/33, 33.3%; SBE: 12/36, 33.3%; no-RBE: 9/36, 25.0%; χ2 = 0.77, p = .679) and also the time to dropout did not differ (median, 22 days [IQR, 42] vs. 49 [14] vs. 35 [64], respectively; χ2 = 3.10, p = .212). The median number of objective and subjective bulimic episodes in the last 28 days before admission was not different between drop-outs and completers (Objective episodes: 2 [IQR, 84] and 0.5 [30] z = −0.17, p = .861; Subjective episodes: 10 [53] and 18.5 [52] z = −0.96, p = .338, respectively).

Pre- and post-treatment scores on measures of eating disorder and comorbid psychopathology for each group are shown in Table 3. BMI as well as all psychological and behavioral measures improved the persistence and self-transcendence subscales of TCI decreased, whereas the scores of harm avoidance and self-directedness increased.

Table 3. Scores pre- and post-treatment on measures of eating disorder and comorbid psychopathology in patients with objective bulimic episodes (OBE group), with subjective without subjective bulimic episodes, in subjects with regular subjective bulimic episodes (SBE group) and in those without regular bulimic episodes (no-RBE group) who completed the program
 OBE Group (N = 22)SBE Group (N = 24)No OBE Group (N = 27)ANOVA for Repeated Measures F (df = 1,2,68)
PretreatmentPosttreatmentPretreatmentPosttreatmentPretreatmentPosttreatmentTimeGroupTime × Group
  • Data are presented as mean (SD) or as °number and (percentage).

  • *

    p < .001;

  • §

    p < .01;

  • £

    P < .05.

  • a

    McNemar test as necessary

Body mass index (kg/m2)15.9 (1.2)19.6 (0.8)14.1 (1.2)18.9 (1.5)14.5 (1.9)19.7 (1.0)546.4*8.30*5.30§
Eating Disorder Examination
 Objective bulimic episodesa22 (100%)2 (9.1%)*0000
 Subjective bulimic episodesa14 (63.6%)9 (40.9%)£24 (100%)3 (12.5%)*05 (18.5%)
 Self-induced vomitinga16 (72.7%)1 (4.5%)*4 (16.7%)02 (7.4%)0
 Laxative misusea6 (27.3%)0§4 (16.7%)04 (14.8%)0
 Diuretics misusea2 (9.1%)02 (8.3%)01 (3.7%)0
 Intense exercisea10 (45.5%)1 (4.5%)§15 (62.5%)4 (16.7%)*14 (51.9%)6 (22.2%)§
 Restraint3.5 (1.8)0.8 (0.7)3.7 (1.4)0.5 (0.6)3.3 (1.7)0.6 (0.9)211.2*0.250.53
 Eating concern4.2 (1.4)1.3 (1.0)3.3 (1.2)1.1 (0.9)2.4 (1.2)1.2 (1.2)137.8*5.75§7.81*
 Weight concern3.6 (1.8)2.4 (1.2)3.6 (1.7)1.6 (1.1)2.9 (1.8)1.8 (1.6)44.2*1.671.16
 Shape concern3.7 (1.1)2.5 (1.3)3.9 (1.3)2.0 (1.2)3.2 (1.4)2.2 (1.5)63.8*0.662.61
 Global score3.7 (1.2)1.7 (0.9)3.6 (1.2)1.3 (0.7)3.0 (1.3)1.4 (1.2)155.4*2.052.24
Beck Depression Inventory32.7 (14.1)15.8 (10.3)30.5 (13.5)13.4 (10.4)28.3 (14.3)18.5 (16.0)98.5*0.192.97
State-Trait Anxiety Inventory (Form-Y)60.2 (9.7)52.1 (14.9)60.5 (12.7)48.3 (11.2)54.2 (16.4)49.8 (16.2)21.7*0.651.82
Temperament and Character Inventory
 Novelty Seeking19.2 (7.5)20.4 (6.6)15.4 (4.4)15.6 (4.6)16.9 (5.4)16.1 (5.7)0.13.99£0.80
 Harm avoidance21.6 (6.3)18.3 (6.2)23.0 (5.6)19.3 (8.3)22.6 (7.7)20.5 (8.6)22.0*0.310.60
 Reward dependence14.7 (4.6)15.5 (4.2)15.4 (3.5)15.5 (3.6)15.1 (3.1)15.0 (3.0)
 Persistence5.4 (1.6)5.0 (2.0)5.3 (2.0)4.3 (2.2)6.0 (1.9)5.3 (2.1)13.5*1.260.70
 Self-directedness18.0 (8.4)20.9 (7.5)18.4 (6.0)23.2 (8.5)25.4 (8.0)27.1 (9.4)8.5£6.74§0.75
 Cooperativeness30.3 (6.7)30.9 (6.2)28.9 (5.5)29.7 (8.5)32.7 (5.5)32.3 (4.8)
 Self-transcendence14.6 (6.3)12.2 (5.1)14.3 (5.2)12.0 (6.4)12.7 (4.6)11.1 (5.7)12.2§0.620.16

All groups achieved a mean normal BMI with treatment, but the no-RBE group regained significantly more weight. In contrast, the eating concern subscale of EDE was more significantly improved in the OBE group. Finally, the number of bulimic episodes, of episodes of self-induced vomiting, laxative misuse and intense exercising was reduced in all groups, whereas they were frequently reported at baseline.

Factors Associated with Treatment Outcome

In a linear regression model, changes in eating disorder psychopathology (EDE global score) were negatively associated with the number of SBEs at baseline (β = −0.24, t = −2.02, p = .048) and with the number of intense exercising episodes (β = −0.37, t = −3.35, p = .001), accounting for 31% of the variance of dependent variable. Age, baseline BMI, change in BMI, the number of objective bulimic episodes, self-induced vomiting episodes, and laxative misuse episodes were not independently significant.

The changes in BMI were negatively associated with baseline BMI (β = −0.49, t = −4.72, p < .001), again accounting for 31% of the variance.


In an inpatient CBT unit for eating disorder, we identified three groups of underweight patients who reported regular objective bulimic episodes with or without subjective bulimic episodes (OBE group), regular subjective bulimic episodes only (SBE group) and neither objective nor subjective bulimic episodes (no-RBE group). These groups were assessed on a wide range of outcomes before and after treatment, including, age, age of menarche, BMI, premorbid BMI, eating disorder duration, suicidal attempts, eating disorder psychopathology, general psychopathology, personality characteristics, and treatment outcome. The main findings were twofold. First, in terms of eating disorder and comorbid psychopathology at baseline, similarities between the OBE and SBE groups outweighed the differences, whereas differences between these groups and the no-RBE group were more pronounced. Second, all three groups had similar outcomes in response to inpatient CBT.

At baseline, our underweight OBE, SBE, and no-RBE groups had similar sociodemographic and clinical characteristics and severity of general psychopathology, also including similar suicidal attempts, thus confirming previous studies in non-underweight clinical8–11 and community samples with objective and subjective bulimic episodes.5, 13, 14 The OBE group had significantly higher BMI than the SBE and no-RBE groups, a finding reported previously in relation to individuals with bulimic-type eating disorders,5, 15 and a higher number of episodes of self-induced vomiting, a difference not observed, to our knowledge, among individuals with bulimic-type eating disorders.5 These findings might be seen to be consistent with the view that purging behaviors are inefficient as means of reducing caloric intake.38

The OBE and SBE groups were characterized by more severe eating disorder psychopathology, in particular significantly higher eating concern and lower self-directness at TCI, a finding also observed in a previous study on AN subtypes,39 than the no-RBE group. According to Cloninger et al.,40 self-directedness is a developmental process encompassing several distinct aspects, including acceptance of responsibility for one's choices, identification of individually valued goals and purposes, resourcefulness, and self-acceptance. Lower self-directness, has been observed associated with poor outcome in patients with eating disorder,41 and with higher drop-out rates in patients with AN treated with brief outpatient individual psychodynamic therapy.42

Importantly, however, in this study the three groups had a similar dropout rate and time to dropout, and responded similarly well to the inpatient CBT, with a significant improvement of BMI, eating disorder behaviors, eating disorder and general psychopathology, and of several personality features (namely, increased self-directness—a prognostic indicator of treatment efficacy—43 and harm avoidance, and decreased persistence and self-transcendence). By the end of treatment, all groups had achieved a mean normal BMI, but the no-RBE group was more prone to regaining weight. This latter finding suggests that there may be limited clinical utility in categorizing underweight eating disorder patients on the basis of the type or frequency of bulimic episodes.

In regression analysis, a higher frequency of subjective bulimic episodes and of episodes of intense exercising prior to admission was associated with a lower reduction in eating disorder psychopathology following treatment, after controlling for initial BMI, age, frequency of self-induced vomiting, laxative and diuretic misuse, and exercising. In contrast, the frequency of objective bulimic episodes prior to admission was not independently associated with any change in eating disorder symptoms following treatment. According to these data, specific strategies should be developed to address subjective bulimic episodes and intense exercising, to improve the outcome of our CBT-based inpatient treatment.

The strengths of the study are the inclusion of several ED-NOS cases, a group scarcely evaluated in previous research on bulimic episodes, and the assessment of bulimic episodes using the accurate EDE interview before and after treatment,8 but several limitations should be also noted. First, because of the limited sample size we could not split our OBE group according to the presence/absence of subjective bulimic episodes as used in other studies.5 The preliminary analysis within the OBE group, showed a higher eating concern in patients with subjective bulimic episodes which may be partly explained by the loss of control over eating that characterizes both OBE and SBE patients. However, no other significant differences emerged between the two groups at baseline. Second, sample size precluded a more stringent separation of “regular binge eating” (“once a week or more”) from “nonregular binge eating” (“less than once a week”) when assigning participants to groups. This may have had the effect of minimizing differences between groups. Although the once a week threshold has been recently shown to capture individuals with clinically significant levels of binge eating psychopathology,27, 28 our definition might have minimized differences between groups. Third, data were derived from a single inpatient unit, mainly treating adult underweight patients; external validation is needed, and the results might not apply to underweight outpatients or adolescent subjects or to individuals with AN- or AN-type EDNOS receiving inpatient treatment in other regions. Fourth, life-time history data on objective and subjective bulimic episodes were not fully available, and the possible effects of crossover between the two types on outcome cannot be evaluated. Fifth, the lack of follow-up precludes any inference on the long-term outcome of patients with OBE and SBE.

In conclusion, our data show that underweight eating disorder patients have similar treatment response to an intensive inpatient program of CBT regardless of whether or not they report objective and/or subjective bulimic episodes. Although methodological limitations preclude any firm conclusions, the findings suggest that there may be limited clinical utility in categorizing underweight eating disorder individuals on the basis of the type and frequency of bulimic episodes and that further investigation of the validity of distinguishing OBE and SBE among underweight eating disorder patients is warranted. Finally, greater attention may need to be given to the role of SBE and excessive exercise in improving treatment outcome for individuals with AN and variants of AN.

Earn CE credit for this article!

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