Symptoms of Eating Disorders With Obsessive-Compulsive
in Patients Disorder
Teresa A. Pigott, M.D., Margaret Altemus, M.D., Cheryl S. Rubenstein, M.A., James L. Hill, Ph.D., Katalin Bihari, M.D., Francine L’Heureux, M.D., Suzanne Bernstein, B.S., and Dennis L. Murphy, M.D.
Objective: sive-compuisive tured, puisive
This
self-rating disorder
Inventory
scale, clinic
has
behavioral
study disorder
been
was designed and eating
potential Method:
overlap of the symptoms The authors administered
the Eating Disorder Inventory, to 59 outpatients and to 60 sex-matched normal volunteers.
previously
dimensions
to explore disorders.
validated
of the
as a reliable
psychopathology
measure
typical
of patients
of obsesa struc-
at an obsessive-comThe Eating Disorder
of the
specific
cognitive
with
eating
disorders.
and The
scores ofthe patients with obsessive-compulsive disorder and ofthe healthy comparison jects were compared with those of 32 female inpatients with anorexia nervosa (N=1 bulimia nervosa (N=22) who had also been given the inventory. Results: The patients obsessive-compulsive disorder scored significantly higher jects on all eight subscales ofthe Eating Disorder Inventory: dissatisfaction, ineffectiveness, perfectionism, interpersonal and maturity fears. Relative to the healthy subjects, male
than the healthy comparison subdrive for thinness, bulimia, body distrust, interoceptive awareness, patients with obsessive-compulsive
disorder
obsessive-compulsive
scores of the
had
more
symptoms
than
female
patients
with
ofthe female patients with obsessive-compulsive 32 female patients with eating disorders and
Conclusions:
These
results
suggest
that
patients
significantly more disturbed eating attitudes and and that they share some of the psychopathological common to patients with eating disorders. (Am J Psychiatry 1991; 148:1552-1557)
T
here has been much recent interest in the clinical and biological overlap of obsessive-compulsive disorder and the eating disorders. Clinically, the focal and extreme preoccupation with food and body image charactenistic of patients with anorexia nervosa and bulimia nervosa resembles to some extent the repetitive and ritualistic behavior exhibited by patients with obsessive-compulsive disorder (1-4). In addition to shared clinical features, there are biological similarities, since serotonin dysnegulation has been implicated in obsessive-compulsive disorder, anorexia nenvosa, and bulimia nervosa (5, 6). These factors have led to speculation concerning a potential link between obsessivecompulsive disorder and eating disorders (2-4).
Presented at the 143rd annual meeting of the American Psychiatric Association, New York, May 12-17, 1990. Received Sept. 26, 1990; revision received April 25, 1991; accepted May 31, 1991. From the Laboratory of Clinical Science, NIMH. Address reprint requests to
Dr. Pigott, Bldg. 10, Rm. thesda, MD 20892.
1552
sub0) or with
3D/41,
NIMH,
9000
Rockville
Pike,
Be-
with
disorder.
The
disorder were midway between those those of the 35 female normal subjects. obsessive-compulsive
disorder
behavior than healthy comparison eating attitudes and behavior
display subjects that are
Several studies have investigated a potential association between obsessive-compulsive disorder and eating disordens, but most have been neither controlled non systematic in using validated rating scales to measure symptom overlap (2-4, 7). Ofthe several controlled studies, the majority have investigated retrospectively, through chart review, the prevalence of obsessive-compulsive disorder symptoms in patients with eating disorders (1 , 2). One exception (8), however, used the National Institute of Mental Health (NIMH) Diagnostic Interview Schedule and found that 33% of the active bulimic subjects and 32% of the inactive bulimic subjects had met the DSM-III cmitenia for obsessive-compulsive disorder at some time in their lives. To our knowledge, there has been only one study (7) that specifically measured the current eating attitudes and behavior of patients with obsessive-compulsive disorder; that study had a small sample size, and the investigators concluded that there was no significant association between obsessive-compulsive disorder and the eating disorders.
Am
J Psychiatry
148:1
1, November
1991
PIGO1T,
TABLE 1 . Data on Patients With Obsessive-Compulsive Disorder Inventory Patients
ALTEMUS,
RUBENSTEIN,
ET AL.
Disorder or Eating Disorders and on He afthy Comparison Subjects Wh o Were Given the Eating With
Obsessive-
Healthy
Compulsive Disorder (N=59)a
Patients
Comparison Subjects
Anorexia
(N=60)’
With
Patients
Nervosa
Bulimia
(N=10)c
With
Nervosa
(N=22)c
Variable
Mean
SD
Mean
SD
Mean
SD
Mean
SD
Age (years) Total group Females Males
37.4 38.1 36.6
10.2 12.1 7.3
30.8 28.6 33.8
9.7 8.7 10.5
23.7
10.0
22.7
2.6
Maudsley Obsessive-Compulsive Inventory score Totalgroup
4.2 4.2 4.3 24.3
3.4 3.3 3.5 5.0
8.8
4.8
10.4
7.2
16.6
5.5
Females
15.9
5.7
Males Body mass
17.9 24.0
5.1 4.4
index
aTwenty..seven ‘Twenty-five CAll female
male and 32 female male patients.
and
35 female
patients. subjects.
With this in mind, we administered the Eating Disorden Inventory (9) to 59 patients meeting the DSM-III-R criteria for obsessive-compulsive disorder. This is a 64item self-report measure that consists of eight subscales measuring drive for thinness, bulimia, body dissatisfaction, ineffectiveness, perfectionism, interpersonal distrust, interoceptive awareness, and maturity fears. Since the Eating Disorder Inventory focuses on the specific cognitive and behavioral dimensions that appear to be important in the development and penpetuation of the symptoms of patients with eating disorders, we wanted to assess patients with obsessive-compulsive disorder for the presence of the same characteristics. In previous studies, it has been shown that in large samples of patients with eating disorders and age- and sexmatched control subjects, the inventory reliably differentiates subgroups of patients with eating disorders and also distinguishes those with serious psychopathology from “normal” dieters (9). While the Eating Disorder Inventory has been administened to large groups of young women (mean±SD age=20.3± 1 .6 years, N=271 ) and the results then cornpared to those for women with anorexia nenvosa (aged 22.5±5.4 years, N=1SS) and bulimia nenvosa (aged 22.6±4.0 years, N=92) (9), it has not been validated in older subjects (more than 30 years of age). Since our group of obsessive-compulsive disorder patients was older than the eating disorder and control groups previously reported, we recruited a comparison group specifically for this study. We also administered the Eating Disorder Inventory to a newly studied group of female patients with anorexia nenvosa and bulimia nervosa in order to compare the obsessive-compulsive disorder group to a group of patients with eating disorders. Furthen, we chose to include men in both the group with obsessive-compulsive disorder and the comparison group because preliminary analysis indicated that the eating attitudes and behaviors of males with obsessivecompulsive disorder were at least as disturbed as those of females (10).
Am
J Psychiatry
1 48:1
1, November
1991
Our objectives for this study were 1 ) to assess systematically the eating attitudes and behavior of patients with obsessive-compulsive disorder as measured by the Eating Disorder Inventory and compare their scones to those of sex-matched healthy comparison subjects and 2) to compare the scores of our sample of patients with obsessive-compulsive disorder with those generally reported for patients with anorexia and bulimia nenvosa.
METHOD Fifty-nine outpatients (27 male and 32 female) at the NIMH obsessive-compulsive disorder clinic who met the DSM-III-R criteria for obsessive-compulsive disonden completed the Eating Disorder Inventory. At the time the inventory was administered, 23 (39%) of the patients had been medication free for at least I month, 21 (36%) were on a stable regimen of clomipramine, and 14 (24%) were on a stable regimen of fluoxetine. A body mass index was calculated from each individual patient’s height and weight ( 1 1 ). Data on these patients are summarized in table 1. The predominant obsessive-compulsive disorder symptoms manifested by the patients included checking and/or fear of harm (36%, N=21), cleaning and/or contamination fears (34%, N=20), repeating and/or preoccupation with numbers (19%, N=11), and scrupulousness (14%, N=8). Five (19%) of the men and six (19%) of the women reported obsessions or cornpulsions that were directly related to their body shape or appearance. Among the female patients with obsessive-compu!sive disorder, 44% (N=14) met the DSM-III-R cnitena for an additional axis I diagnosis. The most common comorbid diagnosis was depression (34%, N=11), and an additional 13% (N=4) met the DSMIII-R criteria for a past episode of depression. In addition, among these patients, one met the criteria for generalized anxiety disorder, social phobia, and panic
1553
EATING
DISORDERS
IN OBSESSIVE
PATIENTS
TABLE 2. Eating Disorder Inventory Subscale Scores of Patients With Obsessive-Compulsive
Disorder and Healthy Comparison Subjects Male
Total Group of Patients With
Total
Obsessive. Eating
Compulsive Disorder (N=S9)
. Disorder
Inventory
Subscale
Mean
Driveforthinness Bulimia Body dissatisfaction
Ineffectiveness Perfectionism
Interpersonal distrust Interoceptive awareness Maturityfears aSignificantly bSignificantly CSignificantly dSignificantly
different different different different
from from from from
of
Patients
SD
Mean
With
Obsessive-
Comparison Subjects (N=60)
Healthy
Compulsive Disorder (N=27) SD
Mean
Comparison Subjects (N=2S) SD
Mean
SD
3.9 7.2
1.0 0.2 3.2
1.6 0.5 5.4
0.9
1.2
6.4
4.7
4.4
3.5
5#{149}3C
4.1
1.6
2.2
0.2 1.6
0.6 1.9
33b
3.8a 2.oa I 1.4 8.4a 73a 4.2a
5.1 3.9 8.5 6.6 44 3.7
1.9 0.5 6.2 0.9 5.2 1.3
3.1 1.5 8.1 1.7 4.3 1.8
4.6a
4.7
0.7
1.5
44d
35a
44
1.3
1.6
3.9k’
2.l’ 9.0c 90d
s.o
healthy comparison subjects (p< 0.005). male healthy comparison subjects (p< 0.05). male healthy comparison subjects (p< 0.01). male healthy comparison subjects (p< 0.001).
disorder, one met the criteria for simple phobia, and one met the criteria for agoraphobia with panic disorden. Among the male patients with obsessive-compu!sive disorder, 10 (37%) met the DSM-III-R criteria for depression, two (7%) met the criteria for social phobia and generalized anxiety disorder, one (4%) had social phobia, and one met the criteria for generalized anxiety disorder. Further, two men (7%) had histories of agoraphobia with panic disorder, and three ( 1 1 % ) had histories of depression. Sixty comparison subjects (25 male and 35 female) were recruited from local advertisements to complete a standardized battery of psychological tests including the Eating Disorder Inventory. All of the comparison subjects had been medication free for at least 3 weeks before entering the study. None of them had current or past significant medical on psychiatric disondens, as detenmined by a structured interview. All potential cornpanison subjects also completed the SCL-90-R (12), the Maudsley Obsessive-Compulsive Inventory (13), and the Beck Depression Inventory and were administered the Hamilton Rating Scale for Depression to exclude those with significant obsessive-compulsive disorder or depressive symptoms. A body mass index was also calculated for each of the comparison subjects. Data on these subjects are summarized in table 1. Thirty-two female patients meeting the DSM-III-R criteria for anorexia nervosa (N=10) on bulimia nervosa (N=22) were administered the Eating Disorder Inventony at the time of their admission to the eating disorder unit at NIMH. Both groups of eating disorder patients had been free of psychotropic medication for at least 1 month before their admission. At the time the inventory was administered, the patients with anorexia nenvosa were at 60%-70% of their calculated average body weight (14), and the bulimic patients were all within calculated normal weight limits. Table 1 presents data on these eating disorder patients. The scores on the Eating Disorder Inventory were first analyzed by using multivaniate analysis of variance
1554
Group Healthy
Sub jects Only
(MANOVA) to determine whether there were group or sex differences or any interaction of these factors when all variables were examined simultaneously. The criterion statistic in all MANOVA procedures was Wi!ks’s lambda. Subsequently, data from the obsessivecompulsive disorder and comparison groups were cornpared by using a two-way analysis of variance (ANOVA) testing the main effects of group and gender and any interaction of the main effects for each subscale of the Eating Disorder Inventory. The effect of rnedication (obsessive-compulsive disorder group only) was evaluated in the same manner. The effect of age was evaluated by using an analysis of covaniance procedure in which age was entered as the covaniate in determining group and gender effects for each subscale of the Eating Disorder Inventory. All of our eating disorder patients were female, so only the female obsessive-cornpulsive disorder patients and comparison subjects were included in the comparative data analyses involving those patients. Again, a MANOVA was conducted to determine whether group differences existed when all variables were considered simultaneously. Subsequent one-way ANOVAs on each variable were accompanied by t tests with the Bonfenmoni connection (alpha=0.0S) comparing female obsessive-compulsive disorder patients, healthy comparison subjects, anonexic patients, and bulirnic patients. All data we report are expressed as mean± SD.
RESULTS To ascertain whether there were any significant effects when all variables were considered simultaneously, we conducted a MANOVA comparing group (obsessive-compulsive disorder patients and healthy comparison subjects), gender (male and female), and Group by Gender effects. This revealed a significant effect for both group and gender (F=12.34, df=8, 108, p