International Journal of (2007) 31, 1722–1730 & 2007 Nature Publishing Group All rights reserved 0307-0565/07 $30.00 www.nature.com/ijo ORIGINAL ARTICLE Night eating syndrome and nocturnal snacking: association with obesity, and psychological distress

SL Colles, JB Dixon and PE O’Brien

Centre for Obesity Research and Education (CORE), The Alfred Hospital, Monash University, Melbourne, Victoria, Australia

Objective: Night eating syndrome (NES) is characterized by a time-delayed pattern of eating relative to , where most food is consumed in the evening and night. This study aimed to investigate the clinical significance of NES and nocturnal snacking by exploring the relationship between NES and (1) obesity, (2) binge (BED) and (3) psychological distress. Subjects: One hundred and eighty bariatric surgery candidates, 93 members of a non-surgical weight loss support group and 158 general community respondents (81 males/350 females, mean age: 45.8713.3 years, mean body mass index (BMI): 34.8710.8 and BMI range: 17.7–66.7). Methods: NES diagnosis required within the previous 3 months: (1) no appetite for breakfast, (2) consumption of X50% of daily energy after 1900 hours and (3) sleep difficulties X3 nights/week. Nocturnal snacking (awakening to eat) was recorded. Validated questionnaires assessed BED, symptoms of depression, appearance dissatisfaction (AD) and mental health-related quality of life (MHQoL). NES and binge eating (BE) (X1 episode/week) were confirmed by interview. Results: NES criteria were met by 11.1% of the total cohort. Across all groups, BE (P ¼ 0.001), BMI (P ¼ 0.003) and male gender (P ¼ 0.013) explained 10% of NES variance. Individuals with co-morbid NES and BE reported similarly elevated psychological distress as other binge eaters. NES alone was not associated with psychological distress. Those with NES who consumed nocturnal snacks reported poorer MHQoL (P ¼ 0.007) and greater depressive symptoms (P ¼ 0.039) and hunger (P ¼ 0.013) than others with NES. Low MHQoL (P ¼ 0.007) and male gender (P ¼ 0.022) explained 27% of the variance in the nocturnal snacking group. Discussion: In this study, NES was positively associated with BMI, BE and male gender. Elevated psychological distress was only apparent in those who consumed nocturnal snacks. Further characterization and understanding of the clinical significance of NES and nocturnal snacking is required. International Journal of Obesity (2007) 31, 1722–1730; doi:10.1038/sj.ijo.0803664; published online 19 June 2007

Keywords: night eating syndrome; nocturnal eating; ; eating disorder;

Introduction pattern of eating relative to sleep, where most food is consumed late in the day and into the evening and night.2 The escalating worldwide prevalence of obesity has drawn Features of the syndrome have generally included morning attention to the association between ‘non-normative’ eating , evening hyperphagia and ,3–10 and more patterns, weight gain and obesity. Night eating syndrome recent research criteria have stipulated the consumption of (NES) is such a condition, observed most frequently among nocturnal snacks, where individuals wake from sleep to groups of and obese individuals. First character- eat.11–15 Individuals with NES, who report nocturnal snack- ized by Stunkard et al.1 in 1955, awareness of NES as a ing may constitute a group with more severe symptoms than behavioral entity has only emerged in the last decade. those who do not wake to eat.16 At present, NES is not Individuals with NES are characterized by a time-delayed formally listed as an eating or sleep disorder, and no agreed diagnostic criteria exist. Correspondence: SL Colles, Centre for Obesity Research and Education NES prevalence estimates are low in community samples (CORE), The Alfred Hospital, Monash University, 23-99 Commercial Road, (range: 0.4–1.6%),5,17,18 but markedly higher among groups Melbourne 3004, Australia. of obese persons seeking medical or surgical weight loss E-mail: [email protected] 19 Received 16 October 2006; revised 6 March 2007; accepted 21 May 2007; (range: 6–64%). Despite this, only one cross-sectional study published online 19 June 2007 has shown a positive relationship between NES and body Night eating syndrome and nocturnal eating SL Colles et al 1723 mass index (BMI).20 Symptoms of depression are more Australia, and the remainder resided in the neighboring consistently associated with NES,1,4,7 but may be influenced states of South Australia and New South Wales. by the concurrence of binge eating disorder (BED). BED involves regular episodes of excessive, uncontrolled over- Bariatric surgery candidates. Consecutive, eligible persons eating, and is strongly associated with psychological accepted into the bariatric surgery program at The Centre for distress.21 In obese populations, BED has been linked with Bariatric Surgery, The Avenue Hospital, Melbourne, Australia. NES 6,9,10,22 and nocturnal snacking;23,24 however, the Two hundred and thirty of 240 subjects agreed to participate nature of the association is uncertain. It is currently unclear and were provided with a survey pack and consent form. Of whether NES, as a discrete condition, is associated with these, 180 completed surveys and consent forms were emotional distress, impairment or disability, and thereby returned, a response rate of 78%. represents an eating disorder of clinical significance.21,25,26 NES could simply constitute a variant of normal eating Weight loss support group respondents. This sample provided behavior that may be linked to weight gain and obesity. data from obese persons who were not seeking bariatric The aim of this study was to investigate the clinical surgery. Subjects were recruited from ‘Take Off Weight significance of NES by exploring the relationship between Naturally’ (TOWN), a weight loss company that consists of NES and (1) BMI, (2) BED and (3) psychological distress. over 130 support groups within Victoria. Research flyers were Three groups differing in body weight and treatment-seeking disseminated throughout the groups and interested indivi- status (bariatric surgery candidates, weight loss support duals were asked to contact the research center to arrange for group members and general community respondents) were the survey pack to be sent out. In total, 158 survey packs recruited. Associations between NES and binge eating, were distributed, and 93 completed surveys were returned, symptoms of depression, body image/appearance dissatisfac- which represents a response rate of 59%. Names and contact tion (AD), mental health-related quality of life (MHQoL) and phone numbers were volunteered by 71% of respondents. eating behavior were investigated. It was hypothesized that: General community respondents. Subjects were randomly (1) NES prevalence would increase with increasing BMI; recruited through flyers placed on notice boards in the (2) NES would be associated with higher psychological general community, flyers on notice boards in two large distress than non-NES (and non-BED) after controlling metropolitan hospitals, and through survey distribution at for any distress related to age, gender and BMI; a large Australian university. Flyers invited any interested (3) Individuals reporting co-morbid NES and BED would individuals, who were not actively seeking weight loss to display higher levels of psychological distress than those contact the research centre to arrange for the survey pack to with either NES or BED; and be sent out. Of 260 distributed packs, 158 completed surveys (4) The subgroup of NES who woke to consume nocturnal were received; a response rate of 61%. Names and contact snacks would show greater overlap with BED and higher phone numbers were volunteered by 72% of respondents. psychological distress than NES who did not wake to eat.

Measures/materials Methods A cover sheet on the weight loss support group and general community surveys requested respondent’s age, home post All participants were recruited between August 2004 and code and contact phone number (both optional), height, January 2006. Inclusion required an age between 18–70 years. weight and date this weight was last checked. Although Subjects were excluded if they had undergone previous based solely on self-report, 87% of subjects from the support bariatric surgery. Six individuals were also excluded due to group and general community stated that they had weighed 27 night-shift work, as was one student with a pattern of late- themselves within the previous month. The surgical group night studying and eating. The primary data were obtained via consented to have demographic and anthropometric infor- self-report surveys. Of 648 distributed, 431 eligible surveys mation obtained from clinic notes. All survey packs were were returned, representing an overall response rate of 66.5%. otherwise identical and consisted of a questionnaire on NES, The study was approved by the Monash University and five validated surveys listed below. Standing Committee on Ethics in Research involving humans, and was conducted in accordance with the Helsinki Eating disorder diagnoses and eating behavior. A self-report Declaration of 1975 as revised in 1983. All subjects were survey screened for NES diagnostic criteria based on those of informed regarding the nature of the questionnaires and Stunkard et al.10 in 1996. The survey informed subjects that consented to study involvement. the questions related to past 3 months only. Six questions requiring a yes/no response were listed: Subjects The majority of all respondents were Caucasians. Ninety (1) Do you usually have no appetite for breakfast? three percent were residents of the state of Victoria, (2) Do you skip breakfast on 3 or more days of the week?

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1724 (3) Do you usually eat the majority of your food intake, that The Multidimensional Body Self Relations Questionnaire is greater than half of the calories that you would eat (MBSRQ)34 was used to calculate an AD score, as a measure of over a 24 h period, after 1900 hours? body image distress. The appearance orientation (AO) (4) Do you have trouble getting to sleep and/or staying subscale assesses the importance an individual places on asleep on 3 or more days of the week? physical appearance and presentation. Appearance evalua- (5) Have you experienced awakenings during the night on tion (AE) provides a measure of how an individual self- at least 3 nights of the week over the last 3 months? assesses their own appearance and attractiveness. The level (6) When you awaken during the night, do you find yourself of AD is determined by calculating the difference between frequently consuming snacks? the AO and AE.35 The Medical Outcomes Trust Short Form-36 (SF-36),36,37 a NES diagnosis required that persons within the previous widely used and validated survey of general health and 3-month period usually: (1) had no appetite for breakfast, (2) outcomes, was used to assess health-related QoL. The SF-36 consumed 50% or more of total energy intake after 1900 consists of eight domains individually weighted into physi- hours and (3) had trouble getting to sleep or staying asleep cal and mental components, which can be combined to on three or more nights of the week. calculate physical component summary (PCS) and mental The Questionnaire on Eating and Weight Patterns – component summary (MCS) scores. A lower MCS indicates Revised (QEWP-R),28,29 a 28-iem instrument to assess BED poorer self-rated psychological health and more social criteria as outlined in the Diagnostic and Statistical Manual disability due to emotional problems. The SF-36 MCS was of Mental Disorders, 4th edition (DSM-IV),21 was used to used as a measure of psychological distress. screen for characteristics of binge eating. The QEWP-R also collects data on weight and shape concerns and symptoms of bulimia nervosa. Data analyses Following completion of the NES questionnaire and the Descriptive statistics were used to calculate the mean7s.d. QEWP-R, all surgical candidates participated in a short for continuous variables when the total study group was semistructured interview, and community respondents divided according to recruitment origin, and to define ‘NES and support group members who reported characteristics of only,’ ‘BE only’ and ‘Co-morbid NES and BE.’ Two control binge eating or night eating underwent a semistructured groups matched for age, gender, BMI and ‘recruitment phone interview. The purpose of the semistructured clinical origin’ to ‘NES only’ and ‘BE only’ were derived from persons and phone interviews was to verify survey responses. The in the ‘no NES or BE’ group. These matched control groups interview moved systematically through both surveys, were also presented as mean7s.d. Binary logistic regression repeating all questions. Subjects were provided with fuller identified factors independently predictive of NES. Factors descriptions of difficult concepts such as the experience of entered into the model included ‘recruitment origin,’ loss of control. During the interview, a brief diet history by gender, BMI, age, depression score, SF-36 MCS and PCS, AD an experienced dietitian (SLC) determined whether subjects score and BE. The difference in mean values for ‘NES only,’ usually ate greater than half their dietary energy after 1900 ‘BE only’ and their matched controls, and ‘NES only,’ ‘BE hours. only’ and ‘Co-morbid NES and BE,’ was assessed using Eating behaviors and cognitions were further assessed independent t-tests, and w2 analysis for gender. Character- using The Three Factor Eating Questionnaire (TFEQ).30 This istics of the subgroups of NES who did and did not consume widely used tool contains 51 items that measure three nocturnal snacks were also compared using independent dimensions of human eating behavior: (1) cognitive t-tests, and w2 analyses for gender and BE. Finally, within all dietary restraint (deliberate restriction of food intake; intent NES and within the total cohort, binary logistic regression to diet); (2) disinhibition of eating (the inability to resist and linear regression identified factors predictive of noctur- social, emotional or external eating cues); (3) subjective nal snacking. Factors entered into the models included feelings of hunger. Scores increase with increasing eating ‘recruitment origin,’ gender, BMI, age, depression score, pathology. SF-36 MCS and AD score. SPSS version 12.0.1 was used for statistical analysis. Psychological health and quality of life. Symptoms of depres- sive illness were assessed using the revised Beck Depression Inventory (BDI),31 a 21-item self-report instrument that Results assesses traits of a major depressive episode. The BDI has been used widely in subjects ranging from normal weight Non-responders to obese. The validity and internal consistency of the BDI There was no difference in gender, age or BMI between are well documented,31,32 although weight ranges in study participants and non-responders in the surgical group. populations were not specified. Within a possible range of 0 Owing to the anonymity of the surveys disseminated to and 63, a score of 0–9 was considered ‘normal’; 10–16 ‘mild’; the community groups, differences between responders and 17–29 ‘moderate’; 30–63 ‘severe depression’.33 non-responders could not be assessed.

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1725 Table 1 Descriptive characteristics and comparison of the three original recruitment groups

Group 1: general community Group 2: weight loss support Group 3: bariatric surgery P-value a respondents group candidates n 158 (36%) 93 (22%) 180 (42%) Male/female 34/124 8/85 39/141 0.018 Mean age 41.3713.5a 55.1712.4b 44.8711.2c 0.000ab 0.026ac 0.000bc Mean BMI 24.875.1a 32.777.3b 44.576.8c 0.000ab 0.000ac 0.000bc NES 9 (5.7%) 4 (4.3%) 35 (19.4%) 0.000 NES+snacksb 2 (1.3%) 1 (1.1%) 13 (7.2%) 0.005 Nocturnal snacksb 6 (3.8%) 4 (4.3%) 20 (7.0%) 0.016 BEb 3 (1.9%) 5 (5.4%) 44 (24.4%) 0.000 BDI depression score 5.575.5a 9.477.3b 16.879.0c 0.000ab 0.000ac 0.000bc MBSRQ-AD scorec 0.1471.0a 1.070.88b 1.871.0c 0.000ab 0.000ac 0.000bc SF-36 MCS 49.776.5a 49.876.9b 46.378.2c NSab 0.000ac 0.001bc SF-36 PCS 53.278.1a 46.2710.9b 36.979.5c 0.000ab 0.000ac 0.000bc TFEQ restraint 8.574.9a 12.673.9b 8.273.9c 0.000ab NSac 0.000bc TFEQ disinhibition 5.773.6a 9.074.0b 11.773.3c 0.000ab 0.000ac 0.000bc TFEQ hunger 4.473.2a 6.073.4b 8.873.6c 0.001ab 0.000ac 0.000bc

Abbreviations: ANOVA, analysis of variance; BDI, Beck Depression Inventory; BMI, body mass index; BE, binge eaters; MCS, mental health component score; NES, night eating syndrome; NS, not significant; PCS, physical component score; SF-36, The Medical Outcomes Trust Short Form-36; TFEQ, three-factor eating questionnaire. Statistical analysis using ANOVA with Tukey post-hoc analysis for continuous variables and presented as mean7s.d., and w2 for categorical variables and presented as n (% of original recruitment group). aPaired letters indicate to which variables each P-value belong. bResults presented as number of subjects (percentage of each recruitment group). cMBRSQ-AD; Multidimensional Body Self Relations Questionnaire – Appearance Dissatisfaction. This score is determined by calculating the difference between the appearance orientation and appearance evaluation subscales.

Participant description 4.4% of the total cohort. No subject met criteria for bulimia Data were obtained from persons of a wide BMI range. The nervosa. final groups comprised the surgery candidates (n ¼ 180, BMI range: 31.9–66.7), weight loss support group members (n ¼ 93, BMI range: 21.3–60.2) and community respondents Correlates and characteristics of NES (n ¼ 158, BMI range: 17.7–45.5). Table 1 lists descriptive The effect of age, gender, BMI and ‘recruitment group’ on the features of each group. presence of NES was assessed in a binary logistic regression model. In the analysis, BMI (Po0.001) and male gender (P ¼ 0.034) explained a significant proportion of the variance Prevalence of eating pathology in NES diagnosis (r2 ¼ 0.049). Figure 1 illustrates NES Based on responses to the self-report questionnaire, 62 prevalence according to BMI category. w2 Analysis showed persons met NES criteria; however, following the confirma- there was a significant difference in NES prevalence across tory clinical or phone interview, this number reduced to a five BMI categories, w2(4, n ¼ 48) ¼ 22.71, Po0.001. final group of 48. Rates of NES were significantly different A second binary logistic regression model explored the between groups at 19.4% in the surgery candidates, 4.3% association between NES and psychological distress while in the support group and 5.7% in the community, w2(2, controlling for BMI and gender. Factors entered included BDI n ¼ 428) ¼ 11.33, Po0.001 (Table 1). A subgroup of NES also depression score, SF-36 MCS, AD score and binge eating. BE reported nocturnal snacking (NES þ snacks). This occurred in were almost seven times more likely to manifest NES (odds 7.2, 1.1 and 1.3% of each recruitment group, respectively, ratio: 6.9; 95% confidence interval (CI): 3.5–13.7). Overall, and was also statistically different between groups, w2(2, BE status (P ¼ 0.001), BMI (P ¼ 0.003) and male gender n ¼ 428) ¼ 5.43, P ¼ 0.005. In particular, rates of NES and (P ¼ 0.013) explained 10% of the variance in NES diagnosis. nocturnal snacking were higher among the surgical candi- w2 Analysis confirmed the strong association between dates than the support and community groups. NES and BE, w2(1, n ¼ 431) ¼ 38.56, Po0.001, and showed a When binge eaters who reported one binge episode per positive trend toward NES and male gender, w2(1, week (n ¼ 14) were compared to those reporting two or n ¼ 431) ¼ 3.81, P ¼ 0.051. more binges per week (n ¼ 38); both groups showed similar The group containing ‘NES only’ (n ¼ 29) was matched for demographic, psychological and behavioral characteristics age, gender, BMI and ‘recruitment origin’ to a comparison (data not shown). These two groups were therefore com- group derived from persons in the ‘No NES or BE’ category. bined and collectively termed binge eaters (BE). Prevalence The ‘BE only’ (n ¼ 33) were also compared to a matched of BE varied between recruitment groups, at 24.4% in the group of subjects without NES or BE. Table 2 demonstrates surgery candidates, 5.4% in the support group and 1.9% there was no statistical difference between the ‘NES only’ among the community respondents, w2(2, n ¼ 428) ¼ 25.14, group and matched controls for all psychological or Po0.001. Co-existing NES and binge eating was present in behavioral variables. In contrast, BE showed significantly

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1726

20%

16%

12%

8%

4%

0% Normal Overweight Class I Class II Class III Weight BMI 25-29.99 Obesity Obesity Obesity BMI 18-24.99 BMI 30-34.99 BMI 35-39.99 BMI 40+

Figure 1 Distribution of all subjects with NES, according to BMI category. Across five BMI categories the prevalence of NES increases as BMI increases, and is statistically different between groups (Po0.001). Statistical analysis using w2.

Table 2 Comparison of ‘NES only’ and ‘BE only’ with control groups matched for age, gender, BMI and recruitment group, derived from ‘No NES or BE’

NES Only (a) Control Group P-value BE Only (b) Control Group P-value Co-morbid NES & BE (c) P-value (a)&(c)/(b)&(c)a

n 29 29 33 33 19 (4%) Male/female 11/18 11/18 NS 5/28 5/28 NS 3/16 0.099/NS Mean age 47.4710.7 47.0711.3 NS 42.278.6 42.378.4 NS 42.177.8 0.071/NS Mean BMI 40.2710.4 39.279.0 NS 43.778.7 42.877.7 NS 43.777.3 NS/NS Current weight 117730.8 110.2729.1 NS 119727.0 116722.4 NS 120724.5 NS/NS BDI score 13.9711.1 10.376.9 NS 21.579.8 14.078.8 0.002 20.879.9 0.033/NS AD score 1.171.5 1.271.2 NS 2.170.9 1.671.1 NS 2.070.90 0.032/NS Weight/shape 2.670.84 2.470.8 NS 3.370.7 2.970.8 0.030 3.370.73 0.005/NS SF-36 PCS 40.879.5 43.4711.5 NS 39.8711.0 39.7710.9 NS 34.578.9 0.031/0.089 SF-36 MCS 47.1710.2 48.876.8 NS 42.077.4 47.677.8 0.004 43.977.4 NS/NS Restraint 8.174.4 8.374.4 NS 8.074.1 8.174.7 NS 9.072.9 NS/NS Disinhibition 9.973.7 9.674.0 NS 14.071.7 9.973.8 o0.001 14.471.4 o0.001/NS Hunger 7.573.7 7.573.7 NS 10.573.1 7.973.7 0.004 10.872.5 o0.001/NS

Abbreviations: AD, appearance dissatisfaction; BDI, Beck Depression Inventory; BMI, body mass index; BE, binge eaters; MCS, mental health component score; NES, night eating syndrome; NS, not significant; PCS, physical component score. ‘NES only’ were similar to their matched controls, while ‘BE only’ were distinguished from their matched control on several psychological and eating-related measures. Comparison of ‘Co-morbid NES & BE’ with ‘NES only’ and ‘BE only’ ((a)&(c)/ (b)&(c)) showed psychological and eating-related measures were lower among ‘NES only’ than ‘BE only’ or ‘Co-morbid NES & BE’. Statistical analysis using independent t-tests for continuous variables and presented as mean7s.d., and w2 for categorical variables. aP-values for (a)&(c); ‘NES only’ and ‘Co-morbid NES & BE, (b)&(c);‘BE only’ and ‘Co-morbid NES & BE.

higher scores for symptoms of depression, SF-36 MCS, weight Table 2 illustrates that weight-related variables were similar and shape concern, dietary disinhibition and hunger, between the three eating disordered groups. Yet importantly, compared to matched controls without binge eating. the ‘Co-morbid NES and BE’ and ‘BE only’ groups scored Characteristics of subjects with ‘Co-morbid NES and BE’ similarly high on all psychological variables. Comparison of (n ¼ 19) were also assessed (Table 2). In total, 40% of those the ‘Co-morbid NES and BE’ and ‘NES only’ groups revealed with NES manifested binge eating, and a similar proportion significantly lower psychological distress among those with of BE also manifested NES (37%). Weight-related variables, ‘NES only’. Symptoms of depression measured by the revised measures of psychological distress and eating behaviors BDI, AD score, dietary disinhibition and hunger, and of ‘Co-morbid NES and BE’ were compared to ‘NES only’ importance of weight and shape were all significantly lower (n ¼ 29) and ‘BE only’ (n ¼ 33) using independent t-tests. in the ‘NES only’ group.

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1727 Comparison of groups who did and did not consume nocturnal Discussion snacks Characteristics of the subgroup of NES who consumed This study compared NES in a large cohort of persons nocturnal snacks (NES þ snack; n ¼ 16) were compared with ranging widely in BMI and treatment-seeking status. Char- NES who did not report this behavior (NES-no snack; n ¼ 32). acteristics of NES were contrasted with non-NES and with Mean BMI was similar in both groups, t(46) ¼À0.92, persons manifesting BE and co-morbid NES and BE. The P ¼ 0.362. Gender distribution was not statistically different, clinical significance of nocturnal snacking was also explored. w2(1, n ¼ 48) ¼ 2.47, P ¼ 0.116, however; males comprised Of primary interest were differences in BMI and markers of 44% of ‘NES þ snacks’ compared to 22% of the ‘NES-no psychological distress. Importantly, regardless of weight snack’ group. Binge eating was not more prevalent in the control endeavors, NES prevalence was positively associated ‘NES þ snacks’ (43.8%) than ‘NES-no snack’ group (37.5%), with BMI. Until now, this association has been generally w2(1, n ¼ 48) ¼ 0.17, P ¼ 0.676, therefore BE were not ex- accepted due to consistently higher NES prevalence esti- cluded from subsequent analysis of psychological traits. The mates in cohorts of overweight and obese when compared to ‘NES þ snack’ group showed significantly higher symptoms the general community. Only one cross-sectional study has of depression on the revised BDI, t(43) ¼À2.13, P ¼ 0.039, a directly supported the positive relationship between NES and lower SF-36 MCS score, t(43) ¼ 2.83, P ¼ 0.007 and greater BMI,20 while the majority have shown no connection.3,5–8 TFEQ hunger score, t(46) ¼À2.59, P ¼ 0.013. In a binary The lack of association may be the result of a small BMI range logistic regression model, low SF-36 MCS (P ¼ 0.007) and within homogenous populations and inadequate power to male gender (P ¼ 0.022) explained 27% of the variance in the detect group differences. The current data, collected speci- ‘NES þ snacks’ group. fically to provide BMI values across a broad spectrum, clearly Finally, within the total cohort (n ¼ 431), the character- show a strong, independent association between NES istics of 30 individuals who reported the consumption of and BMI. nocturnal snacks were explored. Of this group, n ¼ 16/30 also Binge eating behavior was also closely linked with NES. reported full NES criteria, that is the ‘NES þ snacks’ group, Most strikingly, BE were almost seven times more likely to and n ¼ 14/30 were not diagnosed with NES. Significantly manifest NES than non-BE. Co-morbid NES and BE was 4% more males comprised the small group of nocturnal snack- in the total cohort. Yet in the NES group, 40% reported binge ers, w2(1, n ¼ 431) ¼ 4.467, P ¼ 0.035 with 12.3% of all males eating, and among the BE group 37% also reported NES. reporting nocturnal snacking compared to 5.7% of females. These findings represent similar rates of co-morbid NES and In the binary logistic regression model, gender, (P ¼ 0.002), BE to other obese study groups;6,10,22 however, higher8 and the revised BDI score, (P ¼ 0.019) and SF-36 MCS score, lower9,10 rates have been reported. Overlap between noctur- (P ¼ 0.038) explained 7.3% of the variance in nocturnal nal snacking and BED has also been reported within obese snacking behavior. The average score on the revised BDI for populations.23,38–40 While BED and NES have been studied the nocturnal snackers was 19.3711.4 compared to and described as separate entities,13,41 these data highlight 10.378.5 in the remainder of the cohort. Table 3 presents that co-occurrence of the two conditions is common. the results of a linear regression analysis to assess factors Similarities may exist in the consumption of an objectively driving the three measures of psychological distress. Noctur- large amount of food for the circumstances, and binge nal snacking within the total cohort was positively asso- eating, which can extend over several hours, is common in ciated with BDI score and negatively associated with the late afternoon and evening. While a perceived lack of MHQoL. Again, NES was not associated with any psycholo- control in binge eating is essential, night eating behavior gical measure. may also be under poor self-control.42

Table 3 Factors within the total cohort associated with the three measures of psychological distress

Symptoms of depression (revised BDI) Mental health-related QoL (SF-36 MCS) Appearance dissatisfaction (MBSRQ)

Age NS b ¼ 0.11; P ¼ 0.019 NS Female gender b ¼À0.16; Po0.001 b ¼ 0.13; P ¼ 0.004 b ¼À0.27; Po0.001 BMI b ¼ 0.43; Po0.001 b ¼À0.10; P ¼ 0.042 b ¼ 0.49; Po0.001 NES NS NS NS Nocturnal snacks b ¼ 0.18; Po0.001 b ¼À0.19; Po0.001 NS BE b ¼ 0.26; Po0.001 b ¼À0.20; Po0.001 b ¼ 0.15; Po0.001 Total variancea (%) 39 14 36

Abbreviations: BDI, Beck Depression Inventory; BE, binge eaters; BMI, body mass index; MBSRQ, The Multidimensional Body Self Relations Questionnaire; NES, night eating syndrome; NS, not significant; QoL; quality of life; SF-36 MCS, The Medical Outcomes Trust Short Form-36 Mental Health Component score. Night eating syndrome was not associated with any marker of psychopathology, while nocturnal snacking was associated with higher symptoms of depression and poorer mental health-related QoL. Binge eating was strongly related to all three measures. Statistical analysis using linear regression. b-andP-values are provided for significant associations. aCombined R2-value.

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1728 Yet despite some behavioral similarities between sleep-related disorder, such as or restless legs binge eating and night eating, associated levels of psycho- syndrome,40 are possible correlates. Waking to use the logical distress appear markedly different. The group with bathroom,42 or insomnia, where eating acts as a ‘time NES scored low on all psychological measures, and was killer’40 are other factors that may lead to nocturnal comparable to persons without NES. This was in sharp snacking. The identification and treatment of low mood contrast to the BE group who yielded significantly higher could assist to alleviate nocturnal snacking, and reduce the symptoms of depression and AD, more weight and shape associated risk of weight gain. concern, dietary disinhibition and hunger, and lower Finally, lower evening levels may contribute to MHQoL than matched non-bingeing controls. Furthermore, nighttime hunger and stimulate nocturnal snacking. Low when NES and binge eating co-occurred, binge eating was circulating leptin has been observed in normal weight the factor clearly associated with elevated psychological and obese NES, compared to weight-matched non-NES,11 distress. although similar leptin levels have also been reported.12 The A number of studies have examined the link between NES present study did not specifically measure evening hunger, (diagnosed according to various criteria) and associated but found elevated hunger ratings in nocturnal snackers psychological disturbance. However, few studies have con- using a general measure of self-reported hunger. trolled for the influence of binge eating. Those that have A strength of the current study is the inclusion of a large controlled for binge eating have reported similar levels of cohort of subjects comprising a broad BMI range, recruited psychological functioning between NES and non-NES in a from geographically similar locations within the same normal weight sample of Black females18 and an obese timeframe. BMI was derived from clinic measurements of sample.8 Another study, which notably diagnosed NES by weight and height in the obese surgical candidates, and consumption of X25% total energy after the evening meal self-report in the community and weight loss support and/or awakenings to eat X3 times in a week (nocturnal groups. Although self-report weights tend toward under- snacking), did report a positive association between NES estimation, particularly as body weight increases,48 the and symptoms of depression in obese NES compared to majority of our overweight and obese respondents were weight-matched controls.13 A high level of psychological weighed manually. Furthermore, self-reported weights in a disturbance (and concurrent binge eating) have also been general population have shown adequate sensitivity and reported among nocturnal eaters referred to a sleep clinic for good specificity when compared to actual weight.49 polysomnography.40 Although the recruitment methods differed slightly and The present study was the first to examine differences the three original groups possessed distinct characteristics, between subgroups who did and did not consume nocturnal our methods of statistical analyses controlled for possible snacks. Frequent nocturnal snackers reported higher symp- confounders, in particular differences in BMI and the toms of depression and hunger, and lower MHQoL compared presence of binge eating. to the NES who did not wake to eat. This association between Another potential limitation of this study was the collec- nocturnal snacking and psychological distress provides tion of data by self-report questionnaires. To minimize this clinical significance to the behavioral feature of waking to weakness, the surveys selected had been validated previously eat, and supports the proposal that nocturnal snackers are a within a range of population groups. Furthermore, the group with more severe impairment.16 This also highlights QEWP-R and NES survey were used as a screening tool,50 the hitherto lack of distinction regarding the status of NES as and research criteria for BED and NES were verified by either an eating disorder or variant of normal eating behavior. NES clinical or phone interview. An on-going limitation of as defined in the present study appeared to be an extension research involving NES is the lack of formally validated of normal eating behavior linked with weight gain and diagnostic criteria and assessment methods. This study obesity. Further examination is required and should also employed the most commonly applied criteria10 and consider the level of control over nocturnal eating and the validated self-reported behavior in an interview. Nocturnal time-delayed pattern of eating, and the link between NES snacking, which is emerging as an important component of and stress.11,43,44 NES, was also considered. Nocturnal snackers also tended to be male. Although NES as defined in this study showed a strong positive gender differences in nocturnal snackers have not previously association with obesity, while frequent nocturnal snacking been assessed, one other study of morbidly obese treatment conferred an elevated risk of psychological disturbance. seekers has observed an increased risk of NES in males.20 In These findings highlight two clinically significant relation- contrast, similar gender distributions between NES and ships, and importantly, provide a step toward differentiating non-NES have been found in surgical populations,6,45 and a variant of normal eating behavior from obese3,7 and general samples.3,7,17 Although limited by a associated with emotional distress or impairment. The high relatively small male sample, our findings suggest that degree of overlap between NES and binge eating is also depressed men are more likely to engage in nocturnal noteworthy and merits additional study. Finally, we suggest snacking behavior. While there are no clear explanations, that male gender may be a risk factor for NES and nocturnal work-related stress,46 obstructive sleep apnea40,47 or another snacking. Awareness of NES and nocturnal snacking and the

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1729 risks they impose on weight gain and psychological distress 16 Schenck CH, Mahowald MW. Review of nocturnal sleep-related are still little known in general practice. Future work should eating disorders. Int J Eat Disord 1994; 15: 343–356. further define NES features of clinical importance to guide 17 Rand C, Kuldau JM. Eating patterns in normal weight individuals: bulimia, restrained eating, and the night eating syndrome. Int J the development of agreed diagnostic criteria, and develop Eat Disord 1986; 5: 75–84. targeted intervention strategies. 18 Striegel-Moore RH, Dohm FA, Hook JM, Schreiber GB, Crawford PB, Daniels SR. Night eating syndrome in young adult women: prevalence and correlates. Int J Eat Disord 2005; 37: 200–206. 19 de Zwaan M, Burgard M, Schenck CH, Mitchell JE. Night time eating: a review of the literature. Eur Eat Disorders Rev Acknowledgements 2003; 11:7–24. 20 Aronoff NJ, Geliebter A, Zammit G. Gender and body mass index We thank all of the study participants for their time and as related to the night-eating syndrome in obese outpatients. involvement, and the staff at the Centre for Obesity Research J Am Diet Assoc 2001; 101: 102–104. 21 American Psychiatric Association. Diagnostic and Statistical and Education (CORE) and The Centre for Bariatric Surgery Manual of Mental Disorders: DSM-IV. 4th ed. Washington, DC, in Windsor, Victoria for their on-going support and 1994. assistance. 22 Adami GF, Meneghelli A, Scopinaro N. Night eating and binge eating disorder in obese patients. Int J Eat Disord 1999; 25: 335–338. 23 Greeno CG, Wing RR, Marcus MD. Nocturnal eating in binge References eating disorder and matched-weight controls. Int J Eat Disord 1995; 18: 343–349. 24 Grilo CM, Masheb RM. Night-time eating in men and women 1 Stunkard AJ, Grace WJ, Wolff HG. The night-eating syndrome: a with binge eating disorder. Behav Res Ther 2004; 42: 397–407. pattern of food intake among certain obese patients. Am J Med 25 Tanofsky-Kraff M, Yanovski SZ. Eating disorder or disordered 1955; 19: 78–86. eating? Non-normative eating patterns in obese individuals. Obes 2 O’Reardon JP, Peshek A, Allison KC. Night eating syndrome: Res 2004; 12: 1361–1366. diagnosis, epidemiology and management. CNS Drugs 2005; 19: 26 Colles SL, Dixon JB. Night eating syndrome: impact on bariatric 997–1008. surgery. Obes Surg 2006; 16: 811–820. 3 Ceru-Bjork C, Andersson I, Rossner S. Night eating and nocturnal 27 Lennernas M, Hambraeus L, Akerstedt T. Shift related dietary eating-two different or similar syndromes among obese patients? Appetite Int J Obes Relat Metab Disord 2001; 25: 365–372. intake in day and shift workers. 1995; 25: 253–265. 4 Kuldau JM, Rand C. The night eating syndrome and bulimia in 28 Spitzer RL, Devlin M, Walsh BT, Hasin D, Wing R, Marcus MD the morbidly obese. Int J Eat Disord 1986; 5: 143–148. et al. Binge eating disorder: a multisite field trial of the diagnostic 5 Rand CS, Macgregor AM, Stunkard AJ. The night eating syndrome criteria. Int J Eat Disord 1992; 11: 191–203. in the general population and among postoperative obesity 29 Spitzer RL, Yanovski S, Wadden T, Wing R, Marcus MD, Stunkard surgery patients. Int J Eat Disord 1997; 22: 65–69. A et al. Binge eating disorder: its further validation in a multisite 6 Adami GF, Campostano A, Marinari GM, Ravera G, Scopinaro N. study. Int J Eat Disord 1993; 13: 137–153. Night eating in obesity: a descriptive study. Nutrition 2002; 18: 30 Stunkard AJ, Messick S. The three-factor eating questionnaire to 587–589. measure dietary restraint, disinhibition and hunger. J Psychosom 7 Gluck ME, Geliebter A, Satov T. Night eating syndrome is Res 1985; 29: 71–83. associated with depression, low self-esteem, reduced daytime 31 Beck AT, Steer RA. Manual for the Beck Depression Inventory. hunger, and less weight loss in obese outpatients. Obes Res 2001; Psychological Corporation: San Antonio, TX, 1993. 9: 264–267. 32 Richter P, Werner J, Heerlein A, Kraus A, Sauer H. On the validity 8 Napolitano MA, Head S, Babyak MA, Blumenthal JA. Binge eating of the Beck Depression Inventory. A review. Psychopathology 1998; disorder and night eating syndrome: psychological and behavior- 31: 160–168. al characteristics. Int J Eat Disord 2001; 30: 193–203. 33 Smarr KL. Measures of depression and depressive symptoms. 9 Powers PS, Perez A, Boyd F, Rosemurgy A. Eating pathology before Arthritis Rheum 2003; 49: S134–S146. and after bariatric surgery: a prospective study. Int J Eat Disord 34 Cash T. Multidimentional Body-Self Relations Questionnaire 02/94 1999; 25: 293–300. ed. Old Dominion University: Norfolk, VA, 1994. 10 Stunkard A, Berkowitz R, Wadden T, Tanrikut C, Reiss E, Young L. 35 Dixon JB, Dixon ME, O’Brien PE. Body image: appearance Binge eating disorder and the night-eating syndrome. Int J Obes orientation and evaluation in the severely obese. Changes with Relat Metab Disord 1996; 20: 1–6. weight loss. Obes Surg 2002; 12: 65–71. 11 Birketvedt GS, Florholmen J, Sundsfjord J, Osterud B, Dinges D, 36 Ware J. SF-36 Health Survey: Manual and Interpretation Guide. The Bilker W et al. Behavioral and neuroendocrine characteristics of Health institute, New England Medical Center: Boston, 1997. the night-eating syndrome. JAMA 1999; 282: 657–663. 37 Ware J. SF-36 Physical and Mental Health Summary Scales: A User’s 12 Allison KC, Ahima RS, O’Reardon JP, Dinges DF, Sharma V, Manual. The Health Institute, New England Medical Center: Cummings DE et al. Neuroendocrine profiles associated with Boston, 1994. energy intake, sleep, and stress in the night eating syndrome. 38 Hsu LK, Betancourt S, Sullivan SP. Eating disturbances before and J Clin Endocrinol Metab 2005; 90: 6214–6217. after vertical banded gastroplasty: a pilot study. Int J Eat Disord 13 Allison KC, Grilo CM, Masheb RM, Stunkard AJ. Binge eating 1996; 19: 23–34. disorder and night eating syndrome: a comparative study of 39 Hsu LK, Sullivan SP, Benotti PN. Eating disturbances and outcome disordered eating. J Consult Clin Psychol 2005; 73: 1107–1115. of gastric bypass surgery: a pilot study. Int J Eat Disord 1997; 21: 14 O’Reardon JP, Ringel BL, Dinges DF, Allison KC, Rogers NL, 385–390. Martino NS et al. Circadian eating and sleeping patterns in the 40 Manni R, Ratti MT, Tartara A. Nocturnal eating: prevalence and night eating syndrome. Obes Res 2004; 12: 1789–1796. features in 120 insomniac referrals. Sleep 1997; 20: 734–738. 15 O’Reardon JP, Stunkard AJ, Allison KC. Clinical trial of sertraline 41 Stunkard AJ, Allison KC. Two forms of disordered eating in in the treatment of night eating syndrome. Int J Eat Disord 2004; obesity: binge eating and night eating. Int J Obes Relat Metab 35: 16–26. Disord 2003; 27: 1–12.

International Journal of Obesity Night eating syndrome and nocturnal eating SL Colles et al 1730 42 de Zwaan M, Roerig DB, Crosby RD, Karaz S, Mitchell JE. 46 Dahlgren A, Kecklund G, Akerstedt T. Different levels of work- Nighttime eating: a descriptive study. Int J Eat Disord 2006; 39: related stress and the effects on sleep, fatigue and cortisol. Scand J 224–232. Work Environ Health 2005; 31: 277–285. 43 Birketvedt GS, Sundsfjord J, Florholmen JR. Hypothalamic– 47 Dixon JB, Schachter LM, O’Brien PE. Predicting and pituitary–adrenal axis in the night eating syndrome. Am J Physiol excessive day sleepiness in the severely obese: indicators for Endocrinol Metab 2002; 282: E366–9. polysomnography. Chest 2003; 123: 1134–1141. 44 Pawlow LA, O’Neil PM, Malcolm RJ. Night eating syndrome: 48 Rowland ML. Self-reported weight and height. Am J Clin Nutr effects of brief relaxation training on stress, mood, hunger, 1990; 52: 1125–1133. and eating patterns. Int J Obes Relat Metab Disord 2003; 27: 49 Nieto-Garcia FJ, Bush TL, Keyl PM. Body mass definitions of 970–978. obesity: sensitivity and specificity using self-reported weight and 45 Rand CS, Kuldau JM. Morbid obesity: a comparison between a height. Epidemiology 1990; 1: 146–152. general population and obesity surgery patients. Int J Obes Relat 50 Wadden TA, Phelan S. Assessment of quality of life in obese Metab Disord 1993; 17: 657–661. individuals. Obes Res 2002; 10 (Suppl 1): 50S–57S.

International Journal of Obesity