Pergamon 0005-7967(95)00011-9 Behav. Res. Ther. Vol. 33, No. 6, pp. 685-689, 1995 Copyright © 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0005-7967/95 $9.50 + 0.00 CASE HISTORIES AND SHORTER COMMUNICATIONS Exposure with response prevention treatment of anorexia nervosa-bulimic subtype and bulimia nervosa SIDNEY H. KENNEDY, l RANDY KATZ, 2 CHRISTINA S. NEITZERT, I ELIZABETH RALEVSKI 1 and SANDRA MENDLOWlTZ 3 t The Clarke Institute of Psychiatry, Mood & Anxiety Division, 250 College Street, Toronto, Ontario, Canada M5T IR8, -'The Toronto Hospital, Toronto General Division, 200 Elizabeth Street, Toronto, Ontario, Canada and ~The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada MSG IX8. (Received 31 October 1994) Sunnnary--Binge exposure with response-prevention of bingeing (ERP-B) was evaluated in 20 female Ss within an inpatient eating disorders unit over 9 sessions as an adjunct to standard milieu therapy. Subjects met DSM-III-R criteria for either bulimia nervosa (BN) (n = 13) or the bulimic subtype of anorexia nervosa (AN-B) (n = 7). The average age of the Ss in each group was 26.5 (+ 8.8) and 24.1 (_ 6.0)yr, respectively. Results indicate significant within-session and pre-post treatment effects in self-report measures 'urge to binge', 'lack of control', 'feelings of guilt' and 'tension'. Further analysis revealed that the AN-B subgroup had significantly greater reduction in 'depression' and 'urge to vomit' compared to the BN group. This study provides preliminary evidence that ERP-B deserves further investigation with long-term follow-up in both BN and AN-B patients and may be particularly advantageous in the AN-B subpopulation. INTRODUCTION Since the original description of bulimia nervosa (BN) (Russell, 1979) and of bulimic (AN-B) and non-bulimic (AN-NB) subtypes of anorexia nervosa (Garfinkel, Moldofsky & Garner, 1980; Casper, Eckert, Goldberg & Davis, 1980), several treatment approaches for reducing binge eating and vomiting have been evaluated. These include pharmacological interventions, cognitive and interpersonal psychotherapies, and exposure with response prevention (ERP). In virtually all cases treatments have been evaluated in BN Ss while patients with the AN-B diagnosis have been excluded from these trials. ERP involves planned, sustained, and repetitive exposures in the context of response prevention (de Silva & Rachman, 1981). Two types of ERP have been systematically evaluated in the treatment of BN. Rosen and Leitenberg (1982) proposed an anxiety-reduction model of 'bulimia' in which vomiting served as the negative reinforcer of binge eating by removing the fear of weight gain. Thus, patients would be required to consume binge foods up to the point at which they would typically induce vomiting. They would then be supervised for up to 21 hr until the urge to vomit had dissipated (ERP-V) (Leitenberg, Rosen, Gross, Nudelman & Vara, 1988). Although this treatment was effective in a small number of cases, high attrition rates along with the impracticality of organizing binge conditions for each patient have limited the application of this form of ERP, which has been criticized for its emphasis on the centrality of vomiting as a maintaining factor in binge eating (Carter & Bulik, 1994). An alternative approach involves prevention of the actual binge (ERP-B). In this case, small amounts of binge foods are presented to the patient who may "touch it, smell it, lick it or eat a small amount" (Jansen, Broekmate & Heymans, 1992). Schmidt and Marks (1989) compared ERP-V and ERP-B among hospitalized bulimic patients and reported similar reductions in binge-vomit frequencies and other between-session measures. However, within-session comparisons showed greater reductions in urge to binge, anxiety, and liking of food in the ERP-B group. In this study we set out to examine the feasibility of performing ERP-B treatment in an inpatient unit, and to compare its effect on AN-B and BN patients. Standard treatment in this unit includes supervised meals and calorie adjustments for AN patients to promote a weekly weight gain of 1 kg (Kennedy & Shapiro, 1993). We hypothesized that the AN-B group, who have demonstrated a greater capacity for restrained eating than the BN group (Polivy & Herman, 1993), would respond better to the combination of nutritional and exposure treatment than the BN group who are more often characterized by disinhibition and affective instability (Steinberg, Tobin & Johnson, 1990). METHOD Subjects Subjects were 20 drug-free female patients who were consecutively admitted to the inpatient unit of the Programme for Eating Disorders at the Toronto Hospital. They met criteria for inclusion in either the BN subgroup (n = 13) or the AN-B subgroup (n = 7) as assessed using the Structured Clinical Interview for the DSM-llI-R-Patient Edition (SCID-P) (Spitzer, Williams, Gibbon & First, 1990). They remained drug-free for the duration of the study. No significant differences were found between the groups for age, duration of illness or binge frequency. As expected, the BN subgroup had a significantly greater Body Mass Index (BMI) than the AN-B subgroup (t = 5.52, df= 12,6, P < 0.0001) (see Table 1). 685