Bulimia Its Incidence and Clinical Regina C. Casper, MD;

Importance in Patients With Anorexia Nervosa

Elke D. Eckert, MD; Katherine A.

\s=b\ Among the various eating patterns encountered in anorexia nervosa, the occurrence of bulimia (rapid consumption of large amounts of food in a short period of time) is a perplexing phenomenon, because its presence contradicts the common belief that patients with anorexia nervosa are always firm in their abstinence from food. We studied the eating habits of 105 hospitalized female patients within the context of a prospective

treatment study on anorexia nervosa: 53% had achieved weight loss by consistently fasting, whereas 47% periodically resorted to bulimia. The two groups were contrasted with regard to their

developmental and psychosocial history, clinical characteristics, and psychiatric symptomatology. Fasting patients were more introverted, more often denied hunger, and displayed little overt psychic distress. In contrast, bulimic patients were more extroverted, admitted more frequently to a strong appetite and tended to be older. Vomiting was frequent, and kleptomania almost exclusively present in bulimic patients, who manifested

greater anxiety, depression, guilt, interpersonal sensitivity, and had more somatic complaints. This association of bulimia with certain personality features and a distinct psychiatric symptomatology suggests that patients with bulimia form a subgroup among

(Arch

patients with anorexia nervosa. Gen Psychiatry 37:1030-1035, 1980) cardinal feature of anorexia

willful and that leads to loss. The of the illness is facilitated by denial of the emaciated condi¬ tion and abetted by a disturbed cognizance of body size. There is also an unusual ability to suppress or tolerate hunger feelings and to remain physically energetic and active despite the advanced weight loss.1

Thecovertly often triumphant pursuit of thinness life-threatening weight development

Accepted

nervosa

is

a

for publication Oct 2, 1979. From the Research Department, Illinois State Psychiatric Institute, Chicago (Drs Casper and Davis); the Department of Psychiatry, University of Illinois, Chicago (Dr Casper); the Department of Psychiatry, University of Minnesota, Minneapolis (Dr Eckert); the Department of Psychiatry, New York Hospital, New York (Dr Halmi); the Department of Psychiatry, Medical College of Virginia, Richmond (Dr Goldberg); and the Department of Psychiatry, University of Chicago (Dr Davis). Reprint requests to Research Department, Illinois State Psychiatric Institute, 1601 W Taylor St, Chicago, IL 60612 (Dr Casper).

Halmi, MD; Solomon C. Goldberg, PhD;

John M.

Davis,

MD

It is

commonly assumed that in patients with anorexia hunger is not sensed or successfully ignored, and it is less well known that in a certain proportion of patients, hunger sensations or a desire for food become so forceful and compelling that they cannot be curbed, and that these patients generally end up eating stupendous amounts of food (bulimia). Bulimia literally means "ox hunger" or a voracious appetite, but in recent years it has come more and more into use for binge-eating. nervosa,

See also

1036.

Researchers in anorexia nervosa have increasingly noticed this phenomenon. Several have proposed that this tendency to episodically overeat differentiates these patients with respect to psychodynamics,1 the severity of the illness,2-3 and prognosis.4·5 To our knowledge, the frequency of bulimia has seldom been reported, probably due to the difficulties encountered in obtaining this usually closely guarded information. Bulimia is by no means confined to anorexia nervosa, but in this article will only be discussed in this context. If bulimia then is conceived as a situation in which control over powerful hunger feelings is failing, one might expect that patients would be more conscious of their hunger and more often report it. One might also predict that "failure of control" extends to others areas, with bulimic patients conceivably displaying a different symptomatic picture. The aim of this investigation was (1) to determine the proportion of patients with anorexia nervosa who experi¬ ence bulimia in a prospective study comprising a large and well-selected patient sample, and (2) to further character¬ ize this patient population, using a variety of assessments, to determine whether the symptoms of bulimia represent an isolated occurrence or whether its association with other behaviors, feelings, thoughts, impulses, or attitudes might justify a distinction from fasting or abstaining patients with anorexia nervosa. PATIENTS AND METHODS

One hundred five female patients who satisfied the diagnostic criteria for anorexia nervosa* outlined in Table 1 and who were in need of hospitalization were admitted to participate in a controlled National Institute of Mental Health-sponsored treatment study at

Downloaded From: http://archpsyc.jamanetwork.com/ by a STANFORD Univ Med Center User on 07/03/2015

Table 1 .—Criteria for

Diagnosis

of Anorexia Nervosa

between 10 and 40 years and onset of illness between ages of 10 and 30 years Loss of at least 25% of original body weight and/or 15% below normal weight for age and height (normal weights obtained from Metropolitan Life Insurance Policy Scales and Iowa Growth Charts for Children) Demonstration of a distorted attitude and behavior toward eat¬ ing, food, or weight that was represented by any of the follow¬

Age

_

ing:

Denial of illness, with a failure to recognize nutritional needs Apparent enjoyment in losing weight A desired body image of extreme thinness Unusual hoarding or handling of food At least one of the following manifestations: Lanugo (downy pelage) Bradycardia (persistent resting pulse rate of 60 beats per min¬ ute or less) Hypothermia (36.1 °C) Episodes of bulimia (compulsive overeating) Vomiting (may be self-induced) Periods of overactivity

Amenorrhea of at least three months' duration unless illness oc¬ curs before onset of menses No known medical illness that could account for the anorexia and weight loss No other major, psychiatric disorder, such as major affective dis¬ order or schizophrenia

physical discomfort, social interruption, or sleep. The statistical significance of the difference between the bulimic group and the fasting group was assessed by the 2 test for categorical data and by the unpaired t test for continuous variables. An estimate of the frequency of bulimic episodes was obtained and specified from "less than once a week," to "about once a week," to "several times a week," to "at least every day." Vomiting behavior was recorded in the psychiatric and social history form and also as part of the psychiatric rating scale that reflected the patient's behavior on the ward during the pretreatment week. Presence or absence and frequency of bulimia, or presence of vomiting was correlated (Pearson's product moment correlation) against every other vari¬ able recorded in the assessment measures during the pretreat¬ ment period. RESULTS Characteristics and Clinical Features

Bulimia is quite common in anorexia nervosa (Table 2). Approximately half (47%) of the patients reported episodes of overeating. In contrast, the remaining patients (53%) consistently fasted. Two thirds of the bulimic patients admitted to binge-eating more than once a week, 16% to at least every day. Even though bulimic and fasting patients resembled each other in many

respects,

a

number of

important differences emerged. Bulimia patients were significantly (P < .05) older, most of them more than 18 years old, compared to the fasting group, half of whom between the ages of 12 and 17 years. Patients were not found to differ with respect to their age at onset of the were

either the University of Iowa, the University of Minnesota, or the Illinois State Psychiatric Institute hospitals. Male patients were excluded from the study because their small number usually precludes a valid analysis. All patients, and their parents in the case of minors, gave informed written consent after a full explanation of the study protocol. Treatment results have been published elsewhere." During the initial seven-day assessment period, depending on the patient's physical condition, either food or a liquid nutrient was offered; patients were closely supervised and in frequent contact with the nurses and the attending

psychiatrist. Following admission, a social worker, using a structured psy¬ chiatric and social history form containing 290 items, collected information from the patient and her family concerning events and attitudes present, preceding, and since the onset of illness and obtained a careful medical and developmental history including psychosexual development. The patient's personality characteris¬ tics, attitudes, and symptoms were assessed by means of a variety of measures during the seven-day assessment period preceding the study protocol: (1) the Slade Anorectic Behavior Scale," which requires two nurses to observe the presence or absence of anorectic behavior such as exercising, delayed meal times, concealing food; (2) the Hopkins Symptom Check List,10 self-rated by the patient and rated by two nurses, which consists of 61 items that provide six factor scores; and (3) the Raskin Mood Scale," containing 52 statements providing nine clinical scales describing various feel¬ ing states, which was self-rated by the patient and separately by two nurses; (4) an anorectic attitude scale,12 consisting of 63 attitude statements representing eight attitude categories typical of anorexia nervosa: (a) denial of illness, (b) loss of appetite, (c) interpersonal control, (d) thin body ideal, (e) hypothermia, (fi hyperactivity, (g) psychosexual immaturity, and (h) independence seeking, all answered by the patient; (5) a psychiatric rating scale containing 14 items, such as denial of illness, thin body ideal, fear of becoming a compulsive eater, psychosexual immaturity, com¬ pleted by the attending psychiatrist; (6) the Minnesota Multiphas¬ ic Personality Inventory (MMPI); and (7) a body size estimation task,11'4 administered by a research nurse, which involves a visual estimation of different body parts and areas that can be compared to the actual sizes; both measures give a body image perception index. Patients

were divided according to whether they had or had not experienced bulimia, as noted in the psychiatric and social history form. Bulimia was defined as uncontrollable rapid ingestion of large amounts of food over a short period of time, terminated by

illness. The mean highest body weight (56 kg) and the mean lowest body weight (33 kg) ever attained was essentially the same for both groups.

Dieting Habits In both groups, patients made a conscious decision to restrict their food intake; they began dieting intentionally and did so in response to events in their lives. Sixty-nine percent of patients with bulimia readily admitted to a strong appetite and a greater desire to eat meat (Table 3), while significantly fewer of the fasting patients, 46%, conceded hunger feelings (P < .02). This is particularly pertinent because it indicates that those who can success¬ fully deny or ignore hunger are those who can lose weight solely by abstaining from food, while those who experience a strong appetite would find it more difficult to entirely refrain from eating. A majority (57%) of the bulimic patients admitted to vomiting after meals, whereas vomit¬ ing was relatively rare (18%) in the consistently fasting group. Anorectic Behavior

Here the groups showed slight, but revealing differ¬ ences. Fewer bulimic patients said they liked to cook—most likely because they had a harder time resisting the temp¬ tation to eat and thus avoided contact with food—in comparison with fasting patients, 79% of whom said they enjoyed cooking and thereby admitted that food was on their mind. The fasting patients thus disclosed a desire for food, which belies the assertion that they had no appe¬

tite.

Kleptomania

Interestingly, compulsive stealing was reported almost exclusively in patients with bulimia (P < .01). Relevant to this is the observation that bulimic patients scored signif¬ icantly higher on the psychopathic deviate scale of the MMPI (Table 3). Developmental Psychosocial and Sexual History The mean birth weight of 3.2 kg was similar in the two groups. Six percent of bulimic patients had been over-

Downloaded From: http://archpsyc.jamanetwork.com/ by a STANFORD Univ Med Center User on 07/03/2015

Table 2.—Characteristics and Clinical Features of Patients With Anorexia Nervosa No.

(%)

of Patients4

Bulimia

Age, yr 12-17

18-36

Height,

cm

145-160 160-170 >

Table 2.—Characteristics and Clinical Features of Patients With Anorexia Nervosa (cont'd) No.

Bulimia

Fasting Group (N = 56)

Group (N = 49) 20.2 ± 4.7f 14(29) 35(71)

27(48) 29(52)

159.7 ± 5.9 33 15

162.9 18 27

51.7 ± 5.0

53.3

Group (N = 49)

19.7 ± 5.5 < ±

.05

7.0
16 32(65) Precipitating factors Anorexia nervosa starting with voluntary dieting 46(94) Events in patient's life that preceded or coincided with this episode 39(80) Dieting habits No appetite 15(31)

Highest

ever

Strong appetite Frequency of bulimia

Less than once a week About once a week Several times a week

At least every day Vomiting after meals Yes

No Use of purgatives Anorectic behavior Likes cooking Admires a thin person Strange food habit score

34(69)

±

5.5

56.2 ± 10.4 33.8 ± 5.5

30(53) 26(46)