Psychosomatics 39:371-378, August 1998
© 1998 The Academy of Psychosomatic Medine
The Evaluation of Eating and Weight Symptoms
A Comparison of Medically Ill and Eating Disorder Patients
Caroline P. Carney, M.D., and
William R. Yates, M.D.
Received May 30, 1997; revised September 11, 1997; accepted September 23, 1997. From the Departments of Psychiatry and Internal Medicine, The University of Iowa College of Medicine, Iowa City. Address reprint requests to Dr. Carney, Department of Psychiatry, The University of Iowa College of Medicine, Iowa City, IA 52242.
|
ABSTRACT
|
The purpose of this study was to determine the utility of hospitalization in a medical-psychiatry unit for eating disorder patients. A retrospective chart review of 48 patients given an International Classification of Disease-9 diagnosis of an eating disorder was conducted. Presenting symptoms, functional status using the Karnofsky Index, medical and psychiatric evaluation, diagnosis, treatment, and outcome were evaluated. Patients fell into 1 of 3 groups: those with an active eating disorder and comorbid medical complications (Eating Disorder [ED] positive [POS], n=25), those with a history of an eating disorder admitted for some other reason (ED history [HX], n=8), and those with eating or weight symptoms ultimately found to be related to a noneating disorder or primary medical process (ED negative [NEG], n=15). The ED POS patients were younger (28.1 vs. 49.1 years, P=0.0001) but had a lower functional status on admission compared with the ED NEG patients (Karnofsky score 51 vs. 72, P=0.0002). They were more likely to binge eat, restrain intake, and abuse laxatives (P=0.0001, P=0.024, P=0.037, respectively) but did not differ with respect to history of vomiting (P=0.113). The ED POS patients were more likely to be transferred to a general psychiatry or eating disorder unit. Overall length of stay was greater in this group (44.6 vs. 20.4 days, P=0.031). Initial evaluation of patients presenting with eating or weight symptoms may be difficult given similarities between the patients with primary eating disorders and those with other underlying medical causes. The medical-psychiatry unit provides comprehensive initial evaluation and treatment of patients with eating and weight symptoms.
Key Words: Eating Disorder • Eating • Weight • Anorexia Nervosa • Bulimia Nervosa • Medical-Psychiatry Unit
|
INTRODUCTION
|
The eating disorders, including anorexia nervosa (AN), bulimia nervosa (BN), and eating disorder not otherwise specified (ED NOS), have profound medical as well as psychiatric morbidity. Frequently, presentation to the medical system is the result of a medical complication associated with the abnormal eating behavior. Immediate danger may result from hypovolemia, hypokalemia, hypophosphatemia, fractured bones, and in the extreme, bone marrow failure, cardiac decompensation, arrhythmia, and shock.1,2 The purpose of this study is to evaluate the differences in presentation and hospital course of patients presenting with eating- and weight-related symptoms secondary to primary eating disorders (AN, BN, ED NOS) or to primary medical processes.
|
METHODS
|
A retrospective chart review was conducted for each admission on all patients with an International Classification of Disease-9 (ICD-9) coded eating disorder diagnosis (made by trained medical records abstractors) admitted to a medical-psychiatry unit at a large university hospital during a 6-year period. ICD-9 diagnoses were based on the diagnostic impressions documented in the patients' medical records by the ward clinicians. The codes were obtained from the hospital information system's database of discharge diagnostic codes. Patient confidentiality was maintained by deleting names and assigning all patients a study code number. Access to the medical records was limited to the study authors. The project was reviewed and approved by the institutional review board.
The population was not age- or gender-limited. One author (C.P.C.) reviewed the charts and classified patients based on history, reported physical examination findings, laboratory and radiology tests, and other diagnoses made by the joint internal medicine and psychiatric ward team. Eating disorder diagnoses based on DSM-III-R criteria were made by both the medical-psychiatry unit clinicians during the admission and by the author (C.P.C.).3 Eating disorder behaviors such as binge eating and the various forms of purging were recorded if the admitting physician had documented the presence of such behaviors in the admission history taken from the patient. Discrepancies between the ICD-9–coded diagnosis and the reviewer's retrospective diagnoses were noted. The reviewer's diagnosis was assumed to be correct if the ICD-9 diagnosis contradicted a more specific DSM-III-R diagnosis. Three categories were constructed based on chief complaint, history of present illness, past history, and hospital course. Group I Eating Disorder Positive (ED POS) was composed of the group of subjects admitted for treatment of an active eating disorder. Group II Eating Disorder Negative (ED NEG) were admitted for eating and weight symptoms but were ultimately found to have a cause other than a DSM-III-R diagnosis of an eating disorder. Group III Eating Disorder History (ED HX) had a history of an eating disorder but were being admitted for a nonrelated problem. Karnofsky scores were determined as a measure of functional capacity at the time of admission to the hospital.4 The Karnofsky Performance Status Scale is a numerical scale for quantifying patients' status relative to the degree of independence in carrying out normal activities and self-care. The scale ranges from 0, which denotes "dead," to 100, which denotes "no complaints, no evidence of disease." It has been used in populations of terminally and chronically ill patients. Overall length of stay on the medical-psychiatry unit and in the hospital if the patient was transferred to another ward was assessed by using computerized hospital records. A descriptive analysis was performed and statistical analyses were conducted with Statview Version 4.0.5 Group means for continuous variables were calculated by using a one-factor analysis of variance, while means for categorical variables were calculated using chi-square analysis.
|
RESULTS
|
Fifty patients, for a total of 54 admissions, met criteria for review. Each admission was recorded separately. Charts were not available on two patients; therefore, only basic demographic data could be obtained on those subjects, and they were not included in the study population (overall N=48). Age ranged from 16 to 87 years, and mean age was different between the groups (ED POS=28.3, EDHX=25.1, ED NEG=49.1, P<0.0001). Marital status was similar between the groups, except that those patients without eating disorders were more likely to be widowed (P=0.02). Seven subjects were male. Three of these males had an active eating disorder (12% of all ED POS patients). Calculation of body mass index was not possible, as few patients had admission height recorded.
All patient admissions fell into 1 of 3 groups: 1) those with an active eating disorder and comorbid medical complication (ED POS, n=25 admissions in 19); 2) those with a history of an eating disorder but admitted for some other nonrelated reason (EDHX, n=8); and 3) those with eating andweight symptoms, but not a DSM-III-R eating disorder (ED NEG, n=15). Charts were not available on the two remaining patients. Of the 48 patients available for review, AN was diagnosed in 17 subjects, BN in 1, and ED NOS in 1. These 19 patients were responsible for the total of 25 admissions described earlier. See Table 1 for continuous variables.
Symptoms specific for eating disorders were evaluated. Those patients with eating disorders were more likely to have a history of binge eating, dietary restraint, and laxative abuse (P=0.0001, P=0.024, P=0.037, respectively). There was no difference in history of vomiting (P=0.113), although the study's power to detect this outcome was low. Table 2 reflects the percent of patients in each group engaging in weight-control behaviors.
A trend toward a history of previous psychiatric hospitalization was noted in the ED POS vs. ED NEG patients (4.44 vs. 1, P=0.066) However, no difference was noted in the number of prior medical hospitalizations (1.06 vs. 2.8). Table 3 describes the medical findings at presentation in those ED POS and ED NEG patients. Eleven ED POS and only three ED NEG patients presented with hypokalemia. Ten ED POS and nine ED NEG had evidence of volume depletion. Hyponatremia was present in five ED POS, but not present in any of the ED NEG patients.
Continuous variables are described in Table 4. Karnofsky scores were significantly different between the groups (ED POS=72, ED HX=85, ED NEG=51.1, P=0.0002), with the ED NEG group having the lowest, or least functional, score. Age accounted for only 12% of the variance noted in functional status at admission. Despite significant difference in functional status, weight on admission was not different (ED POS 46.1 kg, ED HX 55.4 kg, ED NEG 51.3 kg, P=0.255).
There was also no significant difference in the length of treatment with intravenous fluids (P=0.331) or enteral nutrition (P=0.574) between the groups. The likelihood of undergoing an invasive gastrointestinal procedure or radiological procedure was also not different (P=0.415 and P=0.127, respectively). Eleven ED POS patients underwent electrocardiogram testing. Of these, seven had normal sinus rhythm, three had sinus bradycardia, and only one had sinus tachycardia. Nine ED NEG patients had electrocardiograms, all of which revealed normal sinus rhythm. Gastrointestinal diagnostic procedures in the eating disorder group included abdominal X ray (n=7), abdominal ultrasound (n=5), endoscopic retrograde cholangiopancreatography (n=3), and upper gastrointestinal (GI) contrast study (n=1). Similarly, six ED NEG patients underwent upper GI contrast studies, and seven had abdominal ultrasounds. Length of stay (LOS) range (1–87 days) on the medical-psychiatry unit was similar for the ED POS and ED NEG groups (14.1 vs. 14.8 days), but total hospitalization was greater for the ED POS group (44.6 vs. 20.4 days, P=0.03). Mean weight at discharge from the medical-psychiatry unit was unchanged from admission weight in all groups (ED POS 46.1–44.9 kg, ED HX 55.4–48.4 kg, ED NEG 51.3–48.9 kg). No differences were noted between groups. There was also no difference in the rate of discharge directly to home from the medical-psychiatry unit between the ED POS (44%) and ED NEG (43%) groups. Sixty-two percent of the ED HX group were discharged directly home, but this was not significantly different from the other groups. No deaths were noted during the period of review.
Comorbid psychiatric diagnoses in both groups were also recorded. In those patients with eating disorders, eight polysubstance abuse/dependence and five alcohol abuse/dependence diagnoses were made. The diagnosis of major depressive disorder was made five times in the ED POS group. In the ED NEG group, six diagnoses of major depressive disorder were made, as well as five each of alcohol abuse/dependence and eight polysubstance abuse/dependence.
|
DISCUSSION
|
The medical complications of the eating disorders are numerous and can affect all body systems. Nearly 40% of bulimic inpatients had previously reported significant medical complications, and 70% required some type of medical treatment.6 Eighty-four percent of the patients with anorexia with bulimic features had severe medical complications, and 10% with restricting anorexia or anorexia alternating with binge/purge cycles required intensive care treatment.6
Mortality rates in eating disorders range from 2% to 19%. Sudden death may be attributable to electrolyte abnormalities and concomitant cardiac arrhythmias.7 Long-term outcome has been systematically studied, and suicide has been found to be a significant factor in mortality.8–10 Specific associations of increased mortality included being in the lowest weight group at the time of presentation (<35 kg) and having recurrent hospital admissions for eating problems.9
As this study indicates, persons presenting with eating and weight symptoms may be complex to diagnose and treat. The initial intent of this retrospective review had been to assess the level of medical morbidity of eating disorder patients presenting to a medical-psychiatry unit. The medical-psychiatry setting was chosen because patients with psychiatric symptoms complicated by medical symptoms are generally first admitted to the medical-psychiatry unit at this university hospital. However, at the time of the review of the medical records, it was realized that a portion of the group of patients given an ICD-9 diagnosis of an eating disorder by medical records abstractors actually had anorexia associated with a primary medical illness. Thus, because both groups presented with eating and weight symptoms, we chose to compare the course and early outcome of each group. This comparison revealed that, although on admission the groups appeared similar, functional status and overall LOS were significantly different, likely reflecting the role of psychopathology on a medical presentation of symptoms. A prior study stressed that the duration of hospitalization in a general medical unit be limited, so that rapid nutritional and weight restoration could become the main goal of treatment.9,11 Some authors have suggested that the site of treatment may be less important than the philosophy and commitment of the medical staff.12 This study suggests that patients presenting with eating and weight symptoms have significant medical signs and symptoms for which intensive medical evaluation may be necessary to determine etiology.
This study reveals several specific considerations, perhaps the most important of which is making the proper diagnosis. Diagnostic clarification of underlying psychiatric illness early in the course of hospitalization may result in greater likelihood of problem-based evaluation and treatment. Supervised observation of the actual pattern of eating may clarify between the classic eating disorders and other diagnoses such as conversion disorder.13 With proper diagnosis, the early evaluation of a physical manifestation may be better performed. The results suggest that those persons with an active eating disorder, compared with those with eating or weight symptoms from a medical problem, were more likely to report a history of or currently be engaging in binge eating, dietary restraint, laxative abuse, and exercise, whereas both groups reported a similar frequency of vomiting.
Notably, the ED POS patients in this study group were the most functionally impaired, despite being of a younger age, and had a significantly longer total length of hospital stay when compared with other patients with prominent eating or weight symptoms admitted during the same time period. However, the groups were of similar weight at both admission and discharge. Although it might seem that the patients should have had greater restoration of weight during the stay on the medical-psychiatry unit, failure to restore weight may be explained by the complexity associated with dual diagnoses, and the underlying severity of the eating disorder itself.
Frequently, weight restoration may be lengthy, especially if the admission weight is extremely low. The classic goal of restoration is to achieve a weight gain of 1–1.5 kilograms per week, gradually increasing the caloric intake each week.14 Rapid refeeding may result in fluid retention and lead to cardiac failure. Review of charts in the ED POS population suggested that measures such as postmeal observation and restriction of access to the restroom were not consistently applied. Given the severity of the patients' physical status at admission and interruptions in feeding secondary to preparation requirements for many diagnostic procedures, failure to gain weight was not surprising. If patients are identified with an eating disorder early in the course of hospitalization, certain eating protocols could be established earlier, thus ensuring earlier weight gain. Hypokalemia has been reported to occur in 14% of patients with bulimia nervosa.15 More recently, others reported that in bulimic outpatients, only 4.6% were hypokalemic, and most of these patients purged using laxatives either alone or with concurrent vomiting.16 Forty-four percent of the medical-psychiatry unit population in this study presented with hypokalemia, perhaps relating to disease severity. Additionally, hyponatremia has been reported in about 5% of bulimic patients but occurred in 20% of this population.17 An occurrence of central pontine myelinolysis, a complication of too rapid reversal of hyponatremia, has been reported in a patient with AN with bulimic features.18 A complication such as this likely can be avoided with a medical team more accustomed to treating volume and electrolyte abnormalities.
Unlike other populations, this group did not have gross electrocardiogram changes.19 In Keys's classic work on starvation, subjects at weight loss nadir (24%) below baseline experienced bradycardia as well as decreases in electrocardiogram amplitude, decreased heart size, and diminished stroke volume and hypotension. During nutritional rehabilitation, heart size normalized, but cardiac function did not return to normal in all patients until 32 weeks of weight restoration had taken place.20 Only one of the ED POS patients in the current study had a cardiac complication, biopsy-proven cardiomyopathy considered secondary to her use of syrup of ipecac, a known cause of cardiomyopathy.21–23 Only two persons in the ED POS group had orthostatic hypotension, a condition that has been reported in up to 87% of anorexic persons.19,24,25
Many GI diseases, including inflammatory bowel disease, peptic ulcer disease, achalasis, and motility disorders, mimic eating disorders.26–30 Given that all groups of patients in the study population presented with eating or weight symptoms, it is not surprising that there was not a significant difference in the number of procedures used to work up a GI etiology. GI diagnoses at admission in the ED POS group included pancreatitis (n=1), gastroparesis (n=1), Mallory-Weiss tear (n=1), and pancreatic pseudocyst (n=1).
Eating disorders may coexist with inflammatory bowel disease (IBD). Grybowski has reported three cases of IBD in females with an established eating disorder.29 Two patients who were lactose intolerant used milk ingestion as a purgative. Fear of weight gain in patients with underlying IBD may complicate treatment. Patients may refuse steroids, become "intolerant" of oral feeding, use milk products to induce diarrhea, and may continue vomiting, as this will not directly affect the lower GI tract.
A well-established link between infection with Helicobacter pylori and peptic ulcer disease exists. One study comparing adolescents with eating disorders to a group of well-nourished control subjects and a group of patients with known H. Pylori infection showed no increased incidence of the infection in eating disorder patients.30 A high degree of suspicion for the presence of an eating disorder should exist for patients with atypical ulcers or gastritis symptoms that do not respond to usual treatments with antacids, H2 blockers, or proton pump inhibitors. An efficient workup, including imaging the GI tract to exclude the presence of GI pathology, may make the diagnosis and treatment of an eating disorder more effective.
Three of 25 patients (12%) diagnosed with ED POS were male. Prevalence rates of males with eating disorders range from 5% to 15%.31 Nearly half of the males presenting with eating or weight symptoms (43%) were given a diagnosis of AN. Although the sample size is extremely small, it could be suggested that males presenting with low body weight and eating symptoms should have a thorough history to include the use of food restriction or methods of purging including extreme exercise. Previously reported factors associated with higher risk of eating disorders in males include dieting to avoid teasing, to improve athletic performance, to avoid an illness from which the father may suffer, and to improve body image for a homosexual partner.31
This study is limited by the small number of patients in each group, reflecting a small degree of power. It is possible that greater differences could exist if the sample population was larger. For instance, no underlying electrocardiogram abnormalities were detected, as is usually seen with severe medical complications associated with eating disorders. A population bias may also exist, as patients were enrolled based on an ICD-9 diagnosis of AN, BN, or ED NOS. Therefore, other patients presenting with eating or weight symptoms may have been excluded.
Although Karnofsky scores were determined based on chart review of preadmission functioning and severity of disease at presentation, underlying medical complications were not ranked on a scale of no compromise to severely compromised. Other measures of global assessment of functioning and quality of life might be included in further work.
This study further confirms, however, that significant medical morbidity is associated with eating disorders, particularly AN with or without bulimic features. To facilitate expedient treatment, a standardized assessment and treatment plan might be used in patients presenting with eating and weight symptoms. A future review of the effectiveness of such a plan implemented for use in a medical-psychiatry unit is in order.
|
REFERENCES
|
-
Pomeroy C, Mitchell JD: Medical complications and management of eating disorders. Psychiatric Annals 1989; 19:488–493
-
Comerci GD: Medical complications of anorexia nervosa and bulimia nervosa. Med Clin N Am 1990; 74:1293–1310[Medline]
-
American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC, American Psychiatric Association, 1987, pp. 65–71
-
Karnofsky DA, Burchenal JH: The clinical evaluation of chemotherapeutic agents in cancer, in Evaluation of Chemotherapeutic Agents in Cancer, edited by MacLeod CM. New York, Columbia University Press, 1949, pp. 191–205
-
Abrams Concepts, Inc.: Stat View. Berkeley, CA, Abrams Concepts, Inc., 1992
-
Hall RCW, Hoffman RS, Beresford TP, et al: Physical illness encountered in patients with eating disorders. Psychosomatics 1989; 30:174–191[Abstract/Free Full Text]
-
Isner JM, Roberts WC, Heymsfield SB, et al: Anorexia nervosa and sudden death. Ann Intern Med 1985; 102:49–52
-
Isager T, Brinch M, Kreiner S, et al: Death and relapse in anorexia nervosa: survival analysis of 151 cases. J Psychiatr Res 1985; 19:515–521[Medline]
-
Patton GC: Mortality in eating disorders. Psychol Med 1988; 18:947–951[Medline]
-
Theander S: Outcome and prognosis in anorexia nervosa and bulimia: some results of previous investigations, compared with those of a Swedish long-term study. J Psychiatr Res 1985; 19:493–508[Medline]
-
Croner S, Larsson J, Schildt B. Severe anorexia nervosa treated with total parenteral nutrition. Acta Paediatr Scanda 1985; 74:230–236
-
Garfinkel PE, Garner DM: Anorexia Nervosa: A Multidimensional Perspective. New York, Brunner/Mazel, 1982
-
Garfinkel PE, Kaplan AS, Garner DM: The differentiation of vomiting/weight loss as a conversion disorder from anorexia nervosa. Am J Psychiatry 1983; 140:1019–1022[Abstract/Free Full Text]
-
Kennedy SH, Shapiro C: Medical management of the hospitalized patient, in Medical Issues and the Eating Disorders: The Interface, edited by Kaplan AS, Garfinkel PE. New York, Brunner/Mazel, 1993, pp. 213–238
-
Mitchell JE, Specker S, deZwann M: Comorbidity and medical complications of bulimia nervosa. J Clin Psych 1991; 52 (Oct suppl):13–20
-
Greenfeld D, Mickley D, Quinlan D, et al: Hypokalemia in outpatients with eating disorders. Am J Psychiatry 1995; 152:60–63[Abstract/Free Full Text]
-
Mitchell JE, Pyle RL, Eckert ED, et al: Electrolyte and other physiological abnormalities in patients with bulimia. Psychol Med 1983; 13:273–278[Medline]
-
Greenberg WM, Shah PJ, Vakharia M: Anorexia nervosa/bulimia and central pontine myelinolysis. Gen Hosp Psychiatry 1992; 14:357–358[Medline]
-
Palla B, Litt IF: Medical complications of eating disorders in adolescents. Pediatrics 1988; 81:613–623[Abstract/Free Full Text]
-
Keys A: Cardiovascular effects of malnutrition and starvation. Modern Concepts of Cardiovascular Disease 1948
-
Manno BR, Manno JE: Toxicology of ipecac: a review. Clin Toxicol 1977; 10:221–242[Medline]
-
Isner JM: Effects of ipecac on the heart. N Engl J Med 1986; 313:1253–1255
-
Tolstoi LG: Ipecac-induced toxicity in eating disorders. Int J Eat Disord 1990; 9:371–375
-
Brotman AW, Rigotti N, Herzog DB: Medical complications of eating disorders: outpatient evaluation and management. Compr Psychiatry 1985; 26:258–272[Medline]
-
Jacobs MB, Schneider JA: Medical complications of bulimia: a prospective evaluation. QJM 1985; 214:177–182
-
McClain CJ, Humphries LL, Hill KK, et al: Gastrointestinal and nutritional aspects of eating disorders. J Am Coll Nutr 1993; 12:466–474[Abstract]
-
Kamal N, Chami T, Andersen AE, et al: Delayed gastrointestinal transit times in anorexia nervosa and bulimia nervosa. Gastroenterology 1991; 101:1320–1324[Medline]
-
Waldholtz BD, Andersen AE: Gastrointestinal symptoms in anorexia nervosa: a prospective study. Gastroenterology 1990; 98:1415–1419[Medline]
-
Gryboski JD: Eating disorders in inflammatory bowel disease. Am J Gastroenterol 1993; 88:293–296[Medline]
-
Sherman P, Leslie K, Goldberg E, et al: Helicobacter pylori infection in adolescents with eating disorders and dyspeptic symptoms. J Pediatr 1993; 122:824–826[Medline]
-
Andersen AE: Eating disorders in males, in Eating Disorders and Obesity: A Comprehensive Handbook, edited by Bromwell K, Fairburn C. New York, Guilford, 1995
Get information about faster international access.
Privacy Policy
Copyright © 1998
Academy of Psychosomatic Medicine.
All rights reserved.
Home
| Search
| Current Issue
| Past Issues
| Subscribe
| All APPI Journals
| Help
| Contact Us
|
TOP
ABSTRACT
INTRODUCTION
