Sleep. 14(5):419-431

© 1991 Association of Professional Sleep Societies

Sleep-Related Eating Disorders: Polysomnographic Correlates of a Heterogeneous Syndrome Distinct from Daytime Eating Disorders *Carlos H. Schenck, *Thomas D. Hurwitz, tScott R. Bundlie and tMark W. Mahowald *Minnesota Regional Sleep Disorders Center, Department of Psychiatry, and tDepartment of Neurology, Hennepin County Medical Center and the University of Minnesota Medical School, Minneapolis, Minnesota, U.S.A.

Summary: Over a 5-yr period, 19 adults presented to our sleep disorders center with histories of involuntary, nocturnal, sleep-related eating that usually occurred with other problematic nocturnal behaviors. Mean age (±SD) at presentation was 37.4 (±9.1) yr (range 18-54); 73.7% of the patients (n = 14) were female. Mean age of sleeprelated eating onset was 24.7 (±12.9) yr (range 5-44). Eating occurred from sleep nightly in 57.9% (n = II) of patients. Chief complaints included excessive weight gain, concerns about choking while eating or about starting fires from cooking and sleep disruption. Extensive polysomnographic studies, clinical evaluations and treatment outcome data identified three etiologic categories for the sleep-related eating: (a) sleepwalking (SW), 84.2% (n = 16); (b) periodic movements of sleep (PMS), 10.5% (n = 2) and (c) triazolam abuse (0.75 mg hs), 5.3% (n = I). DSM-III Axis I psychiatric disorders (affective, anxiety) were present in 47.4% (n = 9) of the patients, and only two patients had a daytime eating disorder (anorexia nervosa), each in remission for 3-7 yr. Nearly half of all patients fulfilled established criteria for being overweight, based on the body mass index. Onset of sleep-related eating was linked directly to the onset of SW, PMS, triazolam abuse, nicotine abstinence, chronic autoimmune hepatitis, narcolepsy, encephalitis or acute stress. In the SW group, 72.7% (8/ II) of patients had nocturnal eating and other SW behavior suppressed by clonazepam (n = 7) and/or bromocriptine (n = 2) treatment. Both patients with PMS likewise responded to treatment with combinations of carbidopaiL-dopa, codeine and clonazepam. Thus, sleep-related eating disorders can generally be controlled with treatment of the underlying sleep disorder. Key Words: Sleep-related (nocturnal) eating disorders- Eating disorders- Periodic movements of sleep-Sleepwalking- Triazolam abuse-Affective disorders-Sleep-related injury-Narcolepsy-Clonazepam-Bromocriptine-HLA-DR2negative narcolepsy-Carbidopa/L-dopa-Chronic autoimmune hepatitis-Body mass index.

Nocturnal eating disorders have rarely been mentioned in the scientific literature (1-5), and only one case report (2) and an abstract on three cases (4) have included polysomnographic (PSG) data. The "Nighteating Syndrome" was described in 1955: of 25 patients with treatment-refractory obesity, 80% had "nocturnal hyperphagia, insomnia and morning anorexia" (1). Their mean age was 36.1 yr and 92% were female. A nightly eating disorder related to arousals from REM sleep has been documented in a 37 year old man Accepted for publication March 1991. Address correspondence and reprint requests to Dr. Carlos H. Schenck, Minnesota Regional Sleep Disorders Center, Hennepin County Medical Center, Department of Psychiatry (844), 70 I Park Avenue South, Minneapolis, MN 55415, U.S.A.

during 6 PSG studies (2). He had consumed odd substances such as sunflower oil, and he attributed this problem to "a sort of impulsion" (2). Treatment was not attempted. There was no history of daytime eating disorder. Bulimia nervosa masquerading as sleepwalking (SW) has been reported in a 32-yr-old woman who hadmultiple nightly awakenings and would "eat anythingI can find ... the mess next morning is unbelievable; ... a disgrace, I feel ashamed and I hate myself' (3). She had been teased for being fat and had fantasies of chopping offparts of her hips and thighs. Behavioral therapy for bulimia nervosa rapidly eliminated the nocturnal eating. An atypical form of SW has been described in two women and one man, aged 25-45 yr, who had multiple

419 Downloaded from by guest on 03 January 2018



nightly episodes of sleep-related binge-eating with subsequent amnesia (4). PSG studies documented complex behaviors arising from NREM sleep in all three patients, two of whom were overweight despite strict daytime diets. The third patient was of normal weight but formerly was "anorectic." Treatment issues were not addressed. Sleep-related overeating in a 36-yr-old woman was reported, in a letter, to be controlled by a phobic-stimulus behavioral treatment strategy (5). We now present data from a series of 19 patients with sleep-related eating who underwent clinical evaluations, PSG studies and, in most cases, therapeutic trials. Our findings will be placed within current knowledge on sleep behavior disorders, and also will be applied to diverse clinical disciplines.

METHODS During a 5-yr period (1985-1990), 19 adults (0.5% of all adult referrals) presented clinically to our center for evaluation of involuntary, sleep-related eating, which was a primary complaint in 73.7% (14/19) and a secondary complaint in 26.3% (5/19). Most patients also reported histories of complex and injurious noneating nocturnal behaviors occurring from perceived sleep. The primary complaint group considered nocturnal eating to be distinctly problematic, whereas the secondary complaint group considered it to be only inappropriate. The presenting complaint for the latter group was sleep-related injury. Each patient completed a standard, comprehensive questionnaire covering lifetime sleep/wake, medical, neurologic and psychiatric history and a review of systems. Past medical and psychiatric records were reviewed and family members usually were interviewed. Patients had undergone a medical evaluation prior to referral. A clinical sleep disorders interview was conducted on all 19 patients, a psychiatric examination [utilizing a diagnostic checklist for DSM-III (6) Axis 1 disorders] on 18 patients (14 by CH.S., 4 by T.D.H.), and a neurological examination on 13 patients. The final two patients in our series also completed the structured Diagnostic Interview Schedule, computerized version (c-DIS) (7), utilized in detecting any cument DSM-III Axis 1 diagnosis. Patients were encouraged to complete the following psychometric tests after their interviews: Minnesota Multiphasic Personality Inventory (MMPI), SymptomChecklist 90 (SCL-90), Beck Depression Inventory (BDI) and Zung Self-Rating Anxiety Scale (SAS). Overnight sleep laboratory evaluation, utilizing standard recording and scoring methods (8), was conducted on all 19 patients during their customary hours for sleeping. PSG monitoring included an electro-ocuSleep, Vol. 14, No.5, 1991 Downloaded from by guest on 03 January 2018

logram, nine-channel EEG montage with a paper speed of 15 mm/sec (periodically increased to 30 mm/sec), electromyogram (EMG; chin, bilateral anterior tibialis muscle and, in most cases, bilateral extensor digitorum muscle), EKG and nasal-oral airflow detection with thermistors. Full respiratory monitoring for sleep apnea was employed whenever indicated by history or by findings during PSG study. Audiovisual recording of behavior accompanied all PSG studies. Certified technicians made written notations of observed behaviors. Urine toxicology screens were ordered on 68.4% (13/19) of patients, as indicated by past history or as a clinical precaution. Patients did not bring food to the PSG studies, nor was food available in the sleep lab; patients were informed that food was available in the hospital cafeteria throughout the night. A Multiple Sleep Latency Test (MSLT) (9) was performed on 84.2% (16/19) of patients, given their reports of sleep disruption, daytime fatigue or sleepiness.

RESULTS Clinical data Table 1 presents the main clinical features of this predominantly female group, which was 94.7% (18/ 19) Caucasian (patient #19 was black); 52.6% (10/19) of the patients were married, 26.3% (5/19) were separated or divorced, 73.7% (14/19) were employed, and 57.9% (11/19) were college educated. Although sleeprelated eating was generally long-standing, the other problematic nocturnal behaviors had even longer durations-by a considerable margin-in 52.6% (10/19) of the cases. In the other 47.4% (9/19) of cases, eating was the first manifestation of a nocturnal behavior disorder (i.e. parasomnia); in 31.6% (6/19), sleep-related eating was virtually the only type of problematic nocturnal behavior (patients #2, 4, 14, 15, 17, 18). Twenty-one percent (4/19) of the patients (#6, 7, 9, 16) were also included in a study on sleep-related injury (10). SW onset in this group began during adulthood in 25.0% (4/16) of cases. The PSG findings responsible for the diagnostic assignments in Table 1 will be presented in a subsequent section. Gastroenterologic disorders affected 21.0% (4/19) of patients, but only in the case of patient #2-who had autoimmune hepatitis-was onset linked (at age 5 yr) to the onset of sleep-related eating. Peptic ulcer disease (PUD), in all of the three cases, began years after the onset of sleep-related eating, and symptomatic relief of PUD had no beneficial effect on nocturnal eating. Patient # 13 had gastric bypass surgery for morbid obesity, and the subsequent 62.3-kg weight loss was maintained for the 4 yr before referral. No patient in our


SLEEP-RELATED EATING DISORDERS TABLE 1. Clinical characteristics of 19 patients with sleep-related eating disorders

Diagnostic category

Patient number"


Sleepwalking (SW) (n = 16) Definite 1 (P) Female (n = 12) 2 (P) Female

Age of nocturAge of nal beAge at sleep- havior presen- eating disor- Gastroenterotation onset der on- logic disor(yr) set (yr) dersb (yr)

18 23

9 5

9 5




4 (P) Female




5 (P) Female




6 (S) Female




Chronic autoimmune hepatitis

Neurologic disordersb

Psychiatric disorders b•c

Alcoholsubstance abuse b

Eating disordersb

Narcolepsy (a)


3 (S)


Narcolepsy (a)

Adjustment disorder (r) Generalized anxiety (a) Agoraphobia (r), major depression

AN (r)


Encephalitis (r)

7 (P) Male




Major depression (r), dysthymia (a), generalized anxiety (a) Bipolar disorder

Alcohol (r)


Generalized anxiety (a)

8 (P) Female




9 (S)





Migraine headaches

10 (P) Female




Mixed seizure disorder (a)

11 (P) Female 12 (S) Female

42 47


22 5

13 (S)


Indeterminate' 25 38 41



Schizoaffective disorder (a)d

AN (r) Nicotine (a) Amphetamine (r)

Probable (n = 4)


14 (P) Male 15 (P) Male 16 (P) Female

40 42 42

Periodic movements of sleep (PMS) (n = 2) 17 (P) Male 49 18 (P) Female 54 Triazolam toxicity (n = 1) 19 (P) Female 45

Bipolar disorder


5 38 29

PUD (r)

34 31

PUD (r)





Alcohol (r)


Closed head injury (r)

PUD (r)

Nicotine (r) Major depression

Triazolam (a)


Sex Group statistics Total group (n = 19) 73.7% female (14/19) SWsubgroup(n= 16) 75.0% female (12/16)

Age at presentation (yr)

Age of sleepeating onset (yr)

Age of nocturnal behavior disorder onset (yr)

37.4 ± 9.11 (range 18-54)

24.7 ± 12.91 (range 5-44) (n = 17)

15.6 ± 14.Qf (range 2-44)

35.1 ± 7.91 (range 18-47)

21.6 ± 11.61 (range 7-41) (n = 14)

11. 7 ± 11.2f (range 2-38)

a (P), primary complaint of sleep-related eating [73.7% (14/19) of patients]; (S), secondary complaint of sleep-related eating [26.3% (5/ 19) of patients]. b (a), "active" disorder; (r), disorder in remission; AN, anorexia nervosa; PUD, peptic ulcer disease. Psychiatric disorders were identified by utilizing DSM-I1I Axis 1 diagnostic criteria. d Patients #7 and 10 were disabled by their psychiatric disorders. , The onset of sleep-related eating in patients #9 and 13 occurred many years after the onset of their other nocturnal behavior disorders. fMean ± SD. C

Sleep, Vol. 14, No.5, 1991 Downloaded from by guest on 03 January 2018



study recalled experiencing hunger or gastric reflux symptoms either at bedtime or immediately prior to an episode of nocturnal eating. Neurologic disorders affected 31.6% (6/19) of patients and comprised five diverse etiologies. Narcolepsy onset in patient #2 occurred at age 21 yr together with a pathologic pattern of nocturnal eating that she called "unconscious eating", which was sloppy, dangerous and associated with undesired weight gain. This pattern was superimposed on a long-standing pattern of "conscious" nocturnal eating beginning at age 5 yr. In patient #3, narcolepsy onset was linked with the onset of nocturnal eating in childhood. The onset of encephalitis triggered the onset of SW (with eating during SW) in patient #6. Closed head injury, without known sequelae, in patient #16 occurred 23 yr before the onset of sleep-related eating. In patients #9 and 10, migraine headaches (which had been active for 13 yr) or seizure disorder emerged years after the onset of sleep-related eating. The neurologic exams performed at our center were unremarkable"

The BDI, SCL-90 and SAS were completed by 52.663.1% (10-12/19) of patients. The mean BDI score was 8.7 [±7.9 (SD)], the mean SCL-90 global severity score was 0.85 (±0.81) and the mean SAS index score was 44.6 (± 10.4). These scores were not elevated, apart from a borderline-elevated SCL-90 score. A comparison of mean scores between patients with and those without active Axis 1 disorders revealed significant differences, utilizing the non parametric Mann-Whitney Utest: BDI, 19.0 (±1.4) (n = 3) vs. 5.3 (±5.6) (n = 7) (U = 0, exact 2-tailed p = 0.02); SCL-90, 1.79 (±0.80) (n = 3) vs. 0.45 (±0.37) (n = 8) (U = 1, exact 2-tailed p = 0.03); SAS, 57.0 (±3.5) (n = 4) vs. 39.2 (±7.5) (n = 8) (U = 0, exact 2-tailed p = 0.017). Alcohol or substance abuse (apart from nicotine) affected 21.0% (4/19) of patients, with amphetamine abstinence lasting 22 yr and alcohol abstinence lasting 5-6 yr. Onset oftriazolam abuse immediately resulted in nocturnal eating, which subsequently ceased upon discontinuation oftriazolam. Nicotine abuse in patient #11 was incorporated within a ritual of nocturnal eating, as she would enter the kitchen to smoke and then Psychiatric data eat. For patient # 17, abstinence from nicotine abuse 5 Psychiatric disorders affected 47.4% (9/19) of pa- yr previously had precipitated the onset of multiple tients, with 31.6% (6/19) having affective disorders and nightly episodes of postarousal eating. Urine toxicol21.0% (4/19) anxiety disorders. The one patient (# 15) ogy screens at the time of the PSG studies were negative that did not undergo a psychiatric exam had an MMPI for any illicit substance. Daytime eating disorders affected only 10.5% (2/19) interpreted as normal, and was not depressed, anxious of patients, with remissions lasting 3-7 yr prior to reor psychotic during the neurologic exam. Patients #2 and 3, who completed the c-DIS, were not found to ferral. have any current Axis 1 disorder. Of patients with psychiatric disorders, 66.7% (6/9) had previously b(~en Nocturnal eating data hospitalized during relapses; five patients were hosTable 2 provides the core features of sleep-related pitalized on multiple occasions. eating in this study. The Appendix contains case dePsychiatric disorders began years after the onset of scriptions illustrating the heterogeneity of this synsleep-related eating in 77.8% (7/9) ofthe patients hav- drome. The majority of patients binged on high-calorie ing psychiatric disorders; in patients #10 and 13, psy- foods, and often prepared entire meals; the exceptions chiatric disorders antedated the onset of sleep-related were those who usually ate modest snacks such as cold eating. In the SW group, 50.0% (8/16) had psychiatric cereal. Impaired judgment and sloppiness were comdisorders (six active, with two disabled). mon, as patients ate raw or cooked food with their A direct temporal association between psychiatric hands, poured food on themselves, attempted to drink disorder and sleep-related eating could not be estab- ammonia cleaning solution, dropped food on the floor lished in any case. Nevertheless, nocturnal eating in or took items out of the freezer and scattered them patient #11 originated during stressful circumstances, around the house. They also indiscriminately put large with the birth of her first child while being married to quantities of sugar or salt on food, and ate butter or an alcoholic husband. sugar by the spoonful. MMPIs were completed by 68.4% (13/19) of paThe impulsive consumption of very hot beverages tients. For the eight patients without active psychiatric or oatmeal led to scalding injuries. Several patients disorders, the MMPIs were interpreted as normal in lacerated their digits while cutting food. Frenzied runfive, as mildly abnormal in two and as moderately ning to the kitchen resulted in collisions with furniture, depressed in one. For the five patients with active psy- doors and walls. Disinhibited eating extended to pechiatric disorders, the MMPI interpretation corrobo- culiar concoctions having ingredients that could be nurated the clinical diagnosis in three, was considered tritive or nonnutritive (e.g. cigarettes). There was no normal in one case of generalized anxiety disorder and report of alcoholic beverages being consumed, nor was there any report of sexual disinhibition. was ruled invalid in one. Sleep, Vol. 14, No.5, 1991 Downloaded from by guest on 03 January 2018


SLEEP-RELATED EATING DISORDERS TABLE 2. Characteristics of sleep-related eating in 19 patients %


1) Preferential consumption of high-colorie foods



Sweets and pasta especially

2) 3) 4) 5) 6)

Binging Elaborate food preparation Sloppy food preparation/consumption Injuries from cooking/eating Idiosyncratic concoctions

68.4 63.1 57.9 31.6 26.3

13 12 11 6 5

7) Dreamlike mentation associated with sleep-related eating 8) Amnesia for sleep-related eating a) Partial! in tenni tten t b) Complete c) No amnesia: full awareness 9) Frequency of sleep-related eating a) Nightly b) Variable (with high-frequency clustering) 10) Eating distribution a) Throughout the night b) First half of the night c) > 2 hr after sleep onset 11) Abdominal distention upon terminal awakening from sleep 12) Weight gain from sleep-related eatinga 13) Patient concerns about sleep-related eating: a) Weight gain b) Starting fires from cooking c) Injury d) Loss of control e) Choking while eating



Never any purging Entire hot/cold meals Eating with hands, spilling food Scalds, bums, lacerations, ecchymoses Cat food/salt sandwiches; odd mixtures in blenders Dreams were food-related and congruent with patient's actions

57.9 31.6 10.5

11 6 2

57.9 42.1


78.9 10.55 10.55 63.1

15 2 2 12



68.4 36.8 31.6 31.6 21.0

13 7 6 6 4



During a (mean ± SD) 11.4 ± 7.3 yr


Often resulting in abstention from breakfast For n = 4, the history was indeterminate

a Weight gain from sleep-related eating, kg, mean ± SD: 12.3 ± 7.9; range 5.9-30.4; data for n = 9, because weight gain was indeterminate in 2 patients. Weight at presentation, kg, mean ± SD: 75.7 ± 15.0. Height at presentation, cm, mean ± SD: 167.5 ± 11.2.

Dreamlike mental imagery could accompany such activity, as with one patient who carried lettuce around the house while dreaming of finding a safe place for it. Another patient dressed up for a dinner party and then ate while dreaming that the guests had arrived. Full awareness of nocturnal eating was reported by both patients with PMS. Nevertheless, these 2 patients-along with 11 other patients with a varied range of recall for nocturnal eating-were convinced of the automatic, involuntary nature of their eating. A majority of patients ate nightly, with extremes of six times nightly for 20 yr (patient #11) and of eight times nightly for 5 yr (patient # 17). Furthermore, eating occurred throughout the night in most patients. The distension caused by excessive nocturnal eating, along with lack of hunger, resulted in a general abstention from breakfast. None of the patients overate during the daytime, except patient #13, and most ate less during the daytime than they ate before the onset of nocturnal eating, in an effort to halt their progressive weight gain. All instances of daytime dieting were a response to nocturnal overeating with weight gain. Nearly three-quarters of the patients gained weight from sleep-related eating. According to criteria espoused by the National Institutes of Health (11), 44.4%

(8/18) of our patients were overweight (> 20% in excess of desirable weight), as determined by the body mass index [BMI; weight (kg)/height (m)2], with cut-off scores of 27.3 for females and 27.8 for males (patient #13, with BMI of 31.1, was excluded from this analysis because of incomplete remission of obesity after gastric bypass surgery). Of the overweight group, 16.7% (3/18) were severely overweight, with BMI scores exceeding the cut-off scores (12) of32.3 for females and 31.1 for males. The mean BMI for the entire group (i.e. n = 18) was 26.6 (±4.7), with the range being 20.1-36.0. No patient was underweight (BMI

Sleep-related eating disorders: polysomnographic correlates of a heterogeneous syndrome distinct from daytime eating disorders.

Over a 5-yr period, 19 adults presented to our sleep disorders center with histories of involuntary, nocturnal, sleep-related eating that usually occu...
986KB Sizes 0 Downloads 0 Views