Eating Behaviors 14 (2013) 192–198

Contents lists available at SciVerse ScienceDirect

Eating Behaviors

Disordered eating behaviors and sleep disturbances

Sandra Carvalho Bos ⁎, Maria João Soares, Mariana Marques, Berta Maia, Ana Telma Pereira, Vasco Nogueira,José Valente, António MacedoFaculty of Medicine, University of Coimbra, Portugal

⁎ Corresponding author at: Department of PsychologicaUniversity of Coimbra, Rua Larga, 3004–504 Coimbra, Pofax: +351 239 823170.

E-mail address: sbos@fmed.uc.pt (S.C. Bos).

1471-0153/$ – see front matter © 2013 Elsevier Ltd. Allhttp://dx.doi.org/10.1016/j.eatbeh.2013.01.012

a b s t r a c t

a r t i c l e i n f o

Article history:Received 20 July 2012Received in revised form 8 January 2013Accepted 30 January 2013Available online 21 February 2013

Keywords:Disordered eating behaviorsSleep disturbancesAdulthood

The aim of the present study was to investigate if disordered eating behaviors predicted the development ofsleep disturbances. A total of 870 students participated at baseline, 592 one year later (T1) and 305 two yearslater (T2). The Eating Attitudes Test-40 was used to assess global disordered eating behaviors, dietaryconcerns (DC), bulimic behaviors (BB) and social pressure to eat (SPE). Sleep disturbances were assessedby two items related to difficulties initiating sleep (DIS) and maintaining sleep (DMS). A sleep disturbanceindex (SDI) was calculated by summing DIS and DMS scores. Results revealed that global disordered eatingbehaviors at baseline predicted DIS, DMS and SDI at T1 and T2. Students with increased BB and SPE scoresat baseline were more likely to experience sleep onset and sleep maintenance difficulties in the long term.These results suggest that assessment and correction of eating behaviors might prevent sleep disturbances.

© 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Empirical and clinical findings indicate that eating disturbancesand sleep difficulties are associated. The continuous desire to loseweight, use of purgatives, dissatisfaction with body shape, feelingoverweight and the fear of becoming overweight have all been asso-ciated in a sample of college students with difficulties maintainingsleep and with the perception of not having had a restorative sleep(Seigel, Broman, & Hetta, 2004). Abnormal eating habits have alsobeen associatedwith less sleep in a community study upon 7812 collegestudents (Makino, Hashizume, Yasushi, Tsuboi, & Dennerstein, 2006). Aprevious cross-sectional study from our group (Soares et al., 2011) alsorevealed that global eating disorders, particularly bulimic behaviors(BB) and social pressure to eat (SPE), were associated with increaseddifficulties initiating and/or maintaining sleep among universitystudents.

Studies based on clinical samples show that patients suffering fromanorexia nervosa experience sleep alterations (Benca & Schenck, 2005;Lauer & Krieg, 2004). Sleep architecture is altered (decreased slowwave activity and REM sleep), sleep length is reduced and sleep effi-ciency is decreased in underweight anorectics (Delvenne, Kerkhofs,Appelboom-Fondu, Lucas, &Mendlewicz, 1992; Levy, Dixon, & Schmidt,1988; Nobili et al., 1999; Walsh, Goetz, Roose, Fingeroth, & Glassman,1985). It has also been documented that weight loss, starvationand malnutrition can all affect sleep. One of the earliest reports, theMinnesota experiment, showed that effects of semi-starvation included

l Medicine, Faculty of Medicine,rtugal. Tel.: +351 239 857759;

rights reserved.

a pre-occupation with food, binge eating, irritability, weakness, dryskin, hair loss and sleep disturbances (Garner, Vitousek, & Pike, 1997;Keys, 1948). After weight restoration following previous weight loss,sleep architecture and sleep continuity tended to improve (Lacey,Crisp, Kalucy, Hartmann, & Chien, 1975; Pieters, Theys, Vandereycken,Leroy, & Peuskens, 2004).

The association between disturbances of eating behaviors andsleep difficulties is also supported by the existence of the sleep-relatedeating disorder (International Classification of Sleep Disorders-2,ICSD-2; AASM, 2005) and by the proposed night eating syndrome(Allison et al., 2010; Howell, Shenck, & Crow, 2009). The former is char-acterized by recurrent episodes of involuntary eating and drinkingduring nocturnal arousals (most frequently from slow-wave sleep),which can have adverse health consequences (AASM, 2005). Patientsusually do not recall these episodes the next morning and purgativebehavior is not present. Episodes of nocturnal binging with high caloricfoods and bizarre substances are frequent. Medical problems arecommon, mainly as a consequence of weight gain (e.g. hypertension,diabetes mellitus and obstructive sleep apnea). Dental consequencescan also occur (Howell et al., 2009). The latter syndrome is character-ized by atypical sleep and eating patterns. Subjects may experienceevening hyperphagia (excessive wakeful eating after the last meal andbefore falling asleep), difficulties initiating and maintaining sleep, noc-turnal awakenings associated with food intake and morning anorexia(Howell et al., 2009). Because this syndrome is often associated withbinge eating or bulimia nervosa, it is not included as a separate categoryin the fourth edition of the Diagnostic and Statistical Manual of MentalDisorders Diagnostic (DSM-IV; APA, 1994) or the second edition of theInternational Classification of Sleep Disorders (ICSD-2; AASM, 2005)and its nosological status remains controversial (Allison et al., 2010;Striegel-Moore et al., 2006).

193S.C. Bos et al. / Eating Behaviors 14 (2013) 192–198

It has been observed that regularly restricting sleep duration, whichoccursmore frequently inmodern societies, can interferewith themet-abolic and nutritional status/balance of the body (Crispim et al., 2007).Eating behavior is regulated by a complex interplay of peripheral endo-crine stimuli and central neurotransmitter systems, as well as by circa-dian rhythms and by environmental cues (Passani & Blandina, 2011).Levels of leptin decrease and of ghrelin increasewhich, in turn, enhanceappetite, food intake and body mass index (Crispim et al., 2007). Forexample, Bjortvatn et al. (2007) observed that short sleep durationswere associated with elevated body mass index (BMI) and increasedobesity in a sample of 8860 subjects.

The nature of the relationship and biological mechanisms involvedin the association between eating and sleep problems is not completelyclear. However, it is likely that the hypothalamus plays a central role inthe interface between the sleep–wake cycle, feeding behavior andarousal/stress reactivity. For instance, the lateral hypothalamic area(LHA), which produces the peptides orexins A and B, plays a key rolein the regulation of both ingestive behavior and the sleep–wake cycle.Several lines of evidence support the role of orexins as modulators ofresponses to stress (Winsky-Sommerer et al., 2004). Thus, severalhypothalamic systems, such as the orexinergic and the histaminergicsystem, located in the tuberomammillary nucleus (TMN), are at theheart of different homeostatic functions, illustrating the intricate rela-tionship between wakefulness, stress reactivity and feeding.

Although disordered eating habits have been associated with sleepdisturbances, to our knowledge there is no study investigating whetherthese abnormal habits and bodymass index (BMI) could predict subse-quent sleep disturbance. Therefore, the aim of the present longitudinalstudy was to investigate this issue in a non-clinical sample of universitystudents. Results will enable a better understanding of the effects ofabnormal eating behavior upon sleep to be obtained, andmight contrib-ute to the rationale that acting upon these dysfunctional behaviorswould enable sleep disturbances, which are known to affect anindividual's daytime mood, work activities and well-being, to beprevented.

2. Methods

This research was reviewed and approved by the Medical EthicsReview Committee of the University Hospital of Coimbra (Macedoet al., 2007).

2.1. Procedure

Students were approached and experimental measures were com-pleted as described previously (Azevedo et al., 2010; Soares et al.,2011). Data were collected at three time-points separated by intervalsof one academic year (2000/2001, 2001/2002, 2002/2003), whenexaminations were not taking place. Baseline was considered the firststage, Time 1 (T1) the 1st follow-up year and Time 2 (T2) the 2ndfollow-up year. All participation was voluntary and confidentialitywas ensured.Most students returned the questionnaires but individualswere free not to return a questionnaire (when their participation wasdeemed to have ended) without fear of any form of reprisal.

2.2. Participants

A total of 870 undergraduate students (Mean age, M=19.59 years;SD=1.61; Range=17–25) participated in the study; 544 were females(M=19.5 years, SD=1.56) and 326 males (M=19.8 years, SD=1.68). The vast majority was single (N=861; 99.0%) and 81.1% (n=706) were Medicine students. Of the students assessed at baseline,65% (n=592) and 48.5% (n=305) completed the same measures oneyear (T1) and two years later (T2), respectively. Gender distribution,course allocation and marital status were similar between the threeassessments.

2.3. Measures

2.3.1. Eating disordered attitudes/behaviorsThe Portuguese version of the Eating Attitudes Test-40 (EAT, Garner

& Garfinkel, 1979; Soares, Macedo, Gomes, & Azevedo, 2004) was usedto measure the severity of behavioral eating disorders, with particularinterest in three specific dimensions: (1)Diet Concerns (DC), (2) BulimicBehavior (BB) and (3) Social Pressure to Eat (SPE).

The Portuguese EAT-40 was derived from factor analysis based ona sample of 596 female university students. A principal componentanalysis with varimax rotation (orthogonal method of factor rotation)was applied. A three factor solution was obtained similar to the originalEAT-40 factor structure byGarner, Olmsted, Bohr, and Garfinkel (1982).The first dimension, DC, included 17 items, the second dimension, BB,integrated 10 items and the third dimension, SPE, incorporated 6items. Items with factor loadings lower or equal to .3 in all dimensions(7 items) were excluded. As a result, the Portuguese EAT-40 versionincludes a total of 33 items. Each item is scored on a 6-point Likert-type scale ranging from “never” (0) to “always” (5).

Based on total EAT scores and DC, BB and SPE scores, differentsub-groups were formed. Subjects with total EAT scores 1 standard de-viation above themean formed the groupwith high total EAT score andsubjects with scores 1 standard deviation below the mean formed thegroupwith low total EAT score. Similar methods were used for formingsub-groups for each of the EAT dimensions (DC, BB, SPE).

2.3.2. Body mass index (BMI)BMI (kg/m2) was calculated using the formula: body weight in

kilograms divided by height in meters squared. Height and weightwere self-reported by students at baseline, T1 and T2.

2.3.3. Sleep disturbancesTwo items were used to assess sleep. These items were: (1) “I have

difficulty in falling asleep”, used to assess difficulties initiating sleep(DIS); (2) “I wake up many times during the night”, used to assessdifficulties maintaining sleep (DMS). Each item was scored on a6-point scale ranging from 0 (never) to 5 (always). An overall sleepdisturbance score index (SDI) was calculated by summing the scoresfrom the two individual items (possible range, 0–10) where a higherscore indicated greater subjective sleep disturbance. Based on thisscore, the following groups were formed: (i) good sleepers: studentsreporting that they “never” or “rarely” had DIS/DMS at baseline and atfollow-up (n=74; 24.3%); (ii) persistent sleep difficulties group:students reporting that they “often”, “very often” or “always” hadDIS/DMS at baseline and at follow-up (n=23; 7.5%); (iii) onset sleepdifficulties group: students who did not report sleep complaints atbaseline (never/rarely/sometimes) but suffered from sleep complaints(DIS/DMS) either at T1 (1-year) or T2 (2-years) of the follow-up (n=32; 10.5%); (iv) remission sleep difficulties group: students who reportedsleep complaints at baseline but which decreased (never/rarely/sometimes responses) either at T1 or T2 (n=22; 7.2%).

2.4. Data analyses and statistics

The Statistical Package for Social Sciences (SPSS) was applied(versions 12.0–14.0). Kruskal–Wallis tests, Student's t tests and MannWhitney U tests were used to explore differences between groupswith respect to disordered eating behavior and sleep disturbances.Pearson correlation analyses were conducted and Cohen criterion wasadopted for size-effect interpretation of coefficient correlations: 0.1=small; 0.3=medium; and 0.5=large (Cohen, 1992).

Hierarchical and logistic regression analyseswere applied to investi-gate which factors predicted sleep difficulties over time. The analyseswere carried out for the whole sample and not separately for malesand females, as a previous cross-sectional study (Soares et al., 2011)

Table 1.2Difficulties initiating and maintaining sleep, sleep disturbance index and body massindex at baseline, Time 1 and Time 2 in groups with low or high bulimic behavior atbaseline.

Bulimic behavior at baseline in two groups p

Low group High group

n; mean (SD) n; mean (SD)

Difficulties initiating sleepBaseline 74; 1.31 (0.94) 82; 2.05 (1.06) b0.001T1 74; 1.39 (1.10) 82; 2.01 (1.20) b0.001T2 37; 1.62 (1.04) 51; 2.12 (1.37) 0.067

Difficulties maintaining sleepBaseline 74; 0.82 (0.71) 82; 1.45 (0.93) b0.001T1 74; 1.11 (0.84) 82; 1.48 (0.86) 0.008T2 37; 1.08 (0.89) 51; 1.57 (0.88) 0.003

Sleep disturbance indexBaseline 74; 2.13 (1.45) 82; 3.50 (1.72) b0.001T1 74; 2.50 (1.75) 82; 3.49 (1.81) 0.001T2 37; 1.70 (1.39) 51; 2.49 (1.38) 0.010

Body mass indexBaseline 73; 21.51 (2.28) 82; 21.30 (2.24) 0.562T1 73; 21.69 (2.47) 80; 21.10 (2.04) 0.110T2 37; 21.70 (2.68) 50; 20.80 (1.89) 0.085

Notes: SD=standard deviation; T1: Time 1; T2: Time 2.

194 S.C. Bos et al. / Eating Behaviors 14 (2013) 192–198

had obtained similar results for both genders. More details of the proce-dure used are included in the relevant part of the Results section.

3. Results

3.1. Descriptive data

The mean EAT total score was 54.6 (SD=16.1) at baseline, 54.0(SD=15.8) at T1 and 54.1 (SD=17.3) at T2. Mean values for SDIwere 2.9 (SD=1.6) at baseline, 2.9 (SD=1.6) at T1 and 2.1 (SD=2.4) at T2. For BMI, the mean values were 21.2 (SD=2.3) at baseline,21.3 (SD=2.3) at T1 and 21.4 (SD=2.4) at T2.

DIS was reported by 18.4% of subjects at baseline (response options“often”, “very often” and “always”), 16.6% at T1 and 18.8% at T2. DMS(response options assessed in the same way) was reported by 5.2% ofsubjects at baseline, 5.4% at T1 and 6.8% at T2. When exploring theprevalence of difficulties initiating and/or maintaining sleep together,results showed that 19.9% (n=173) reported one or both sleep difficul-ties at baseline, 17.7% (n=105) at T1 and 21% (n=64) at T2. Both sleepdifficulties (initiating and maintaining sleep) were reported by 7.5% ofstudents at all three stages of the study.

3.2. Sleep difficulties and BMI scores by disordered eating behaviorgroups

The group with more severe global disorders of eating behavior atbaseline revealed more DIS, DMS and SDI compared to the group withless eating disorders at baseline and stages T1 and T2 of the study.However, no significant differences between these groups wereobserved with respect to BMI values (Table 1.1).

The group with more DC showed more overall sleep disturbancesthan the group with lower values in this EAT dimension at baseline(M=3.17, S=1.78 vs. 2.62, SD=1.57, t=−2.19, p=0.03). No otherstatistically significant differences were observed between the DCgroups.

In comparison to the group with less BB, the group with moresevere BB experienced more DIS, DMS and SDI at baseline, T1 andT2. Marginally significant differences were observed between the

Table 1.1Difficulties in initiating and maintaining sleep, sleep disturbance index and body massindex at baseline, Time 1 and Time 2 in groups with low or high total scores for disordersof eating behavior at baseline.

Total EAT score at baseline in two groups p

Low group High group

n; mean (SD) n; mean (SD)

Difficulties initiating sleepBaseline 57; 1.40 (1.00) 54; 2.00 (1.10) 0.003T1 57; 1.39 (0.96) 54; 2.06 (1.24) 0.002T2 29; 1.31 (0.71) 31; 1.90 (1.35) 0.037

Difficulties maintaining sleepBaseline 57; 0.91 (0.66) 54; 1.28 (0.90) 0.016T1 57; 1.00 (0.65) 54; 1.41 (0.81) 0.005T2 29; 1.00 (0.65) 31; 1.52 (0.96) 0.014

Sleep disturbance indexBaseline 57; 2.32 (1.38) 54; 3.28 (1.76) 0.002T1 57; 2.39 (1.37) 54; 3.46 (1.77) 0.001T2 29; 1.55 (1.27) 31; 2.32 (1.47) 0.034

Body mass indexBaseline 56; 20.45 (2.27) 54; 21.08 (2.44) 0.159T1 57; 20.67 (2.26) 52; 20.85 (2.24) 0.678T2 29; 20.34 (1.89) 31; 20.57 (1.60) 0.622

Notes: EAT=disorders of eating behavior test; SD=standard deviation; T1: Time 1;T2: Time 2.

groups for DIS at T2. BMI was not significantly different betweenthe BB groups at any of the three times of testing (Table 1.2).

For SPE, the groupwith higher scores in this dimension experiencedmore DIS, DMS and SDI at baseline, T1 and T2 than the other group (lowSPE). The group with more severe SPE also revealed lower BMI valuescompared to the other group at all stages of the study (Table 1.3).

When exploring in more detail associations between individualitems in the EAT score and sleep variables, “Preoccupied with a desireto be thinner” and “I avoid eating when I am hungry” were DC itemsat baseline thatweremost consistently associatedwith sleep difficultiesat T1 and T2; “I have the impulse of vomiting aftermeals” and “Give too

Table 1.3Difficulties initiating and maintaining sleep, sleep disturbance index and body massindex (baseline, Time 1 and Time 2) in low and high groups for social pressure to eatat baseline.

Social pressure to eat in two groups at baseline p

Low group High group

n; Mean (SD) n; Mean (SD)

Difficulties initiating sleepBaseline 100; 1.33 (0.95) 92; 2.00 (1.11) b0.001T1 100; 1.32 (0.96) 92; 1.95 (1.22) b0.001T2 52; 1.42 (1.00) 53; 2.00 (1.18) 0.008

Difficulties maintaining sleepBaseline 100; 1.01 (0.67) 92; 1.41 (0.96) 0.001T1 100; 1.10 (0.78) 92; 1.51 (0.97) 0.002T2 52; 1.10 (0.72) 53; 1.55 (0.97) 0.011

Sleep disturbance indexBaseline 100; 2.34 (1.33) 92; 3.41 (1.79) b0.001T1 100; 2.42 (1.56) 92; 3.46 (1.85) b0.001T2 52; 1.77 (1.29) 53; 2.57 (1.49) 0.006

Body mass indexBaseline 99; 22.87 (2.49) 90; 19.53 (1.89) b0.001T1 98; 23.10 (2.53) 89; 19.73 (1.98) b0.001T2 52; 22.99 (2.49) 52; 19.80 (1.95) b0.001

Notes: SD=standard deviation; T1: Time 1; T2: Time 2.

Table 2Correlations between EAT items and sleep variables at Time 1 and Time 2.

Difficultiesinitiating sleep

Difficultiesmaintainingsleep

Sleepdisturbanceindex

T1 T2 T1 T2 T1 T2

Dietary concerns items — baselinePreoccupied with a desire to bethinner

.094⁎ .158⁎⁎ – .130⁎ .092⁎ .132⁎

I avoid eatingwhen I am hungry – .142⁎ .104⁎ .142⁎ .093⁎ .167⁎⁎

Bulimic behavior items — baselineI have the impulse of vomitingafter meals

.149⁎⁎ .186⁎⁎ .133⁎⁎ .236⁎⁎ .167⁎⁎ .223⁎⁎

Give too much time andthought to food

.149⁎⁎ .120⁎ .085⁎ .128⁎ .138⁎⁎ .138⁎

Social pressure to eat items—baselineFeel that others pressure me toeat

.209⁎⁎ .180⁎⁎ .161⁎⁎ .208⁎⁎ .216⁎⁎ .216⁎⁎

Feel that others would preferme to eat

.198⁎⁎ .132⁎ .189⁎⁎ – .226⁎⁎ .113⁎

Notes: Cohen criterion for size effect interpretation of coefficient correlations: 0.1=small;0.3=medium; .0.5=large; T1: Time 1; T2: Time 2.⁎ p≤0.05.

⁎⁎ p≤0.01.

195S.C. Bos et al. / Eating Behaviors 14 (2013) 192–198

much time and thought to food” were the BB items most consistentlyassociatedwith sleep difficulties at T1 and T2 and “Feel that others pres-sure me to eat”was the item from the SPE dimension persistently asso-ciated with sleep difficulties at both follow-up times of assessment(T1 and T2) — Table 2.

Table 3Total EAT score, dietary concerns, bulimic behaviors, social pressure to eat and body mass i

Baseline 1 2

Good sleepers Onset sd

n=74 n=32

Total EAT scoreM (SD) 47.75 (14.33) 59.19 (15.22)Md 45.50 60.00P25–75 38.25–54.50 48.50–66.5Range 25–90 35–93

Dietary concernsM (SD) 16.59 (9.05) 21.28 (11.36)Md 16.00 22.00P25–75 10.00–21.00 12.50–26.25Range 0–51 0–48

Bulimic behaviorsM (SD) 4.84 (2.98) 7.34 (4.15)Md 4.00 7.50P25–75 3.00–7.00 4.00–9.75Range 0–14 1–20

Social pressure to eatM (SD) 4.44 (2.47) 6.03 (3.43)Md 4.00 6.50P25–75 2.5–6.0 3.25–8.00Range 1–14 1–13

Body mass indexM (SD) 21.36 (2.46) 21.09 (1.83)Md 20.95 21.05P25–75 19.75–23.10 19.55–22.67Range 15.8–29.4 17.9–24.5

sd=sleep difficulties M=mean; SD=standard deviation, Md=median; P=percentile.⁎ p≤0.05.

⁎⁎ p≤0.01.

3.3. Disordered eating behaviors and BMI scores by sleep groups

The groups with persistent, remission or onset sleep difficultiesrevealed more disordered eating behavior (total EAT score), BB andSPE at baseline than the group of good sleepers (Table 3). The groupwith sleep onset difficulties showed higher DC scores at baseline thandid the good sleepers group (M=21.28, SD=11.36 vs. M=16.59,SD=9.05, pb0.05). No significant differences were observed betweenthe sleep groups and BMI at baseline.

3.4. Regression analyses

Multiple regression analyses were conducted in order to explorewhether BMI and EAT dimensions at baseline predicted DIS, DMS orSDI at T1 and T2. Hierarchical regressions were conducted as follows:BMI was included in the model, followed by the three EAT dimensions(DC, BB and SPE) entered as a block. This procedure was repeated foreach of the dependent variables (DIS, DMS, SDI) and independentlyfor T1 and T2. The results are summarized in Tables 4.1–4.3.

BMI was found not to be a significant predictor of difficulties initiat-ing or maintaining sleep or overall sleep disturbances at T1 or T2.

Global disorders of eating behavior predicted: (i) difficulties initi-ating sleep at T1 (R2=0.065, pb0.001) and T2 (R2=0.060, pb0.001);(ii) difficulties maintaining sleep at T1 (R2=0.028, p=0.001) and T2(R2=0.044, pb0.004) and (iii) overall sleep difficulties at T1 (R2=0.061, pb0.001) and T2 (R2=0.054, p=0.002).

When exploring in more detail the EAT dimensions that predictedsleep difficulties, it was observed that BB and SPE were the dimensionsthat significantly explained difficulties initiating and maintaining sleepand overall sleep disturbances at T1. At T2, both BB and SPE explained

ndex compared between four different sleep groups.

3 4 Kruskal Wallis(Mann Whitney U)

Remission sd Persistent sd

n=22 n=23

61.29 (16.81) 65.22 (26.37) 0.002⁎⁎

61.00 58.50 (1b2⁎⁎, 3⁎⁎, 4⁎⁎)47.25–73.50 43.75–80.5037–96 39–133

21.64 (11.58) 23.17 (14.87) 0.035⁎

21.50 20.00 (1b2⁎)12.75–29.00 12.00–30.005–51 4–62

7.54 (4.29) 8.91 (5.37) b0.001⁎⁎

7.50 8.00 (1b2⁎⁎, 3⁎⁎, 4⁎⁎)4.00–9.25 4.00–13.001–18 1–21

7.27 (3.18) 6.87 (3.02) b0.001⁎⁎

6.50 6.00 (1b2⁎, 3⁎⁎, 4⁎⁎)5.50–9.00 4.00–9.002–14 2–13

21.41 (2.10) 20.75 (2.33) 0.85021.65 20.9019.55–23.07 19.60–21.9017.5–24.5 16.0–26.0

Table 4.1Hierarchical regression analyses: body mass index and EAT dimensions as predictors of difficulties initiating sleep.

Baseline Difficulties initiating sleep

Time 1 Time 2

β R2 change F p Β R2 change F p

Body mass index −0.048 0.002 1.346 0.246 −0.075 0.006 1.686 0.195Dietary concerns −0.023 0.620 0.048 0.479Bulimic behaviors 0.187 b0.001 0.151 0.031Social pressure to eat 0.147 0.001 0.123 0.038

0.065 13.263 b0.001 0.060 6.293 b0.001

Significant results in bold.

196 S.C. Bos et al. / Eating Behaviors 14 (2013) 192–198

difficulties initiating sleep but SPE was the only EAT dimension thatexplained difficulties maintaining sleep and overall sleep disturbances.

Logistic regression analysis showed that higher BB values at baselinewere associated with an increased likelihood of experiencing onset ofsleep difficulties (β=0.206; OR=1.2; 95%CI=1.03–1.47, p=0.025)or persistent sleep difficulties (β=0.276; OR=1.3; 95%CI=1.07–1.62, p=0.009). Similar results were obtained for the SPE dimension(group with onset of sleep difficulties vs. good sleepers group: β=0.303, OR=1.4, 95%CI=1.09–1.68, p=0.006; group with persistentsleep difficulties vs. good sleepers group: β=0.388, OR=1.5, 95%CI=1.10–1.98, p=0.009).

4. Discussion

The prevalence of sleep disturbances found in our study sample iswithin the range (16–21%) reported by Ohayon (2002). Subjects whoreported difficulties initiating and/or maintaining sleep were 19.9% atbaseline, 17.7% at T1 and 21.0% at T2. Persistent sleep difficulties at allstages of the study (either difficulties initiating or maintaining sleep,or both) were reported by 7.5% of subjects. This prevalence is alsoclose to the range (6%) reported by Ohayon (2002).

One important finding of the present study is that subjects withmore severe global disorders of eating at baseline reported significantlymore sleep difficulties in initiating and maintaining sleep and in overallsleep disturbances at all stages of the study, when compared to studentswith less severe disorders. These results are in accord with communitybased studies which reveal that eating disorders amongst college stu-dents are associated with sleep disturbances, particularly with the per-ception of not having had a restorative sleep (Seigel et al., 2004), withless sleeping hours (Makino et al., 2006) and with difficulties initiatingand/or maintaining sleep (Soares et al., 2011). The main strength of thepresent follow-up study is that it gives novel information indicating thatdisordered eating habits are consistently associated either with currentor subsequent difficulties in initiating and/or maintaining sleep. To ourknowledge, this is the first report in the literature documenting theseobservations in a longitudinal study.

When exploring in more detail which EAT dimensions predictedlong-term difficulties in initiating and/or maintaining sleep and per-sistent insomnia, it was observed that BB and SPE were the significantpredictors.

Table 4.2Hierarchical regression analyses: body mass index and EAT dimensions as predictors of dif

Baseline Difficulties maintaining sleep

Time 1

β R2 change F

Body mass index 0.028 0.001 0.472Dietary concerns −0.023Bulimic behaviors 0.104Social pressure to eat 0.117

0.028 5.457

Significant results in bold.

The fact that BB was associated with sleep difficulties could berelated to the abnormal eating patterns or the presence of binge eatingepisodes, particularly in the late evening/early night, both of which areknown to compromise sleep. In fact, subjects experiencing a “nighteating syndrome” often complain about difficulties in initiating ormaintaining sleep, or both, as a result of consuming a large amount offood in the late evening (taking in more than 50% of their total dailycalories after 7 pm). These subjects also usually reportmorning anorexiaand may wake up during the night to snack (Lauer & Krieg, 2004).Nolan and Geliebter (2012) recently observed that university students(n=246) with moderate and full symptoms of the “night eating syn-drome” also referred lower sleep quality.

Furthermore, sleep difficulties can contribute to cumulative sleeploss, which has been shown to be associated with increased levels ofappetite and hunger (Crispim et al., 2007; Crispim, Zalcman, Padilha,Dáttilo, & Túlio de Mello, 2008), so re-enforcing BB. This vicious cycleis difficult to break and might be an explanation for the persistentlong-term sleep difficulties experienced by students with severe BB.Moreover, elevated psychological distress that students might beexperiencing (due to examinations, for example) can precede or rein-force nocturnal snacking (Colles, Dixon, & O'Brien, 2007),which is likelyto interfere with sleep. In fact, stress-related problems, particularlysocial anxiety and psychological distress, are the main reasons for stu-dents to seek help (psychological treatment and counseling) in theoutpatient clinic at Coimbra University (Pereira, 2004).

The explanation for the association between SPE and sleep diffi-culties is not so straightforward but might be explained, in part, bythe individual's BMI. Students who reported more SPE also had alower BMI (M=19.5, SD=1.89), which is considered in the literatureclose to the value (BMI=18.5) for being underweight (WHO, 1995).A low BMI has been associated with fasting attitudes/behaviors whichare known to influence the continuity and architecture of the mainsleep period (Delvenne et al., 1992; Garner et al., 1997; Levy et al.,1988; Nobili et al., 1999; Walsh et al., 1985).

Another (but not mutually exclusive) explanation could be relatedto personality traits. Perfectionists, particularly subjects with highlevels of socially prescribed perfectionism (SPP), that is, individualswho perceive that others have excessively high standards and expec-tations relating to themselves (Hewitt & Flett, 1991), often experi-ence more psychological distress and negative affect (Molnar, Reker,

ficulties maintaining sleep.

Time 2

p β R2 change F p

0.492 −0.040 0.002 0.489 0.4850.629 0.065 0.3400.032 0.096 0.1730.006 0.122 0.0410.001 0.044 4.539 0.004

Table 4.3Hierarchical regression analyses: body mass index EAT dimensions as predictors of sleep disturbance index.

Baseline Sleep disturbance index

Time 1 Time 2

β R2 change F p β R2 change F p

Body mass index −0.017 0.000 0.159 0.690 −0.062 0.004 1.161 0.282Dietary concerns −0.026 0.577 0.036 0.593Bulimic behaviors 0.173 b0.001 0.114 0.101Social pressure to eat 0.154 b0.001 0.154 0.010

0.061 12.531 b0.001 0.054 5.645 0.002

Significant results in bold.

197S.C. Bos et al. / Eating Behaviors 14 (2013) 192–198

Culp, Sadava, & DeCourville, 2006). If such distress is associated withnegative coping responses, for example, rumination (Flett, Hewitt,Blankstein, & Gray, 1998), this combination may lead to a cognitivehyperarousal thatmakes the subjectmore prone to react to life stressorssuch as the pressure of family or peers by eating. This cognitivehyperarousal (associated with worry and rumination) is one possiblemechanism for explaining insomnia (Perlis, Smith, & Pigeon, 2005). Infact, the study of Azevedo et al. (2010) revealed that SPPwas the perfec-tionist dimension which indicated prospective sleep difficulties; andthe study of Soares et al. (2009) showed that students with high levelsof SPP also had more SPE one year and two years after baseline. Thus, itcould be that SPP is a personality trait that makes the subject moreprone to experience sleep difficulties (associated with rumination,psychological distress and negative coping strategies)when confrontedby significant others to eat. This social pressure may lead to the onset ofeating disorders as was found in the longitudinal study of the McNightInvestigators (2003), which was based on data from 1358 femalestudents who were followed over a 4-year period.

Our results do not support an association between DC and sleep. Thefact that DC are nowadays frequent (almost “normative”) among thegeneral population, particularly among university students, and notexclusively found in subjects with abnormal eating habits, may explainthis result.

One of the limitations of the present study is that depression andanxiety have not been controlled for even though, in a normal sample,the possibility that a mood or anxiety disorder could be acting as themain factor explaining the observed associations is not straightforward.However, we recognize that it is important to disentangle the effects ofdepression and anxiety symptoms, from disordered eating habits andsleep difficulties. One possible hypothesis is that abnormal eatingattitudes may precede sleep disturbances and these two conditionscombined can precede a more serious psychiatric disorder such asDepression, Anxiety or an Eating Disorder. Future studies are neededto investigate this issue. Another limitation of the study is that objectivemeasurements of neither BMI nor sleep were performed. Furtherstudies involving such objective measurements are required.

In conclusion, the main finding of the present study is that disor-dered eating habits, particularly BB and SPE, predicted later difficul-ties in initiating and/or maintaining sleep. These two factors alsopredicted the onset and persistency of sleep disturbances with time.Increased SPE was associated with a lower BMI. These findings indi-cate that assessing eating disorders, particularly BB and SPE, in collegestudents can be extremely apposite to preventing future occurrenceand persistence of sleep disturbances, which are known to affectnegatively human well-being and daytime functioning (Drake, Roehrs,& Roth, 2003).

Role of funding sourceFinancial supportwas obtained from thePortuguese Foundation for Science and Tech-

nology (Project title: Perfectionism and Obsessive–Compulsive Spectrum Disorders;Grant nr. 37569/PSI).

The involvement of the funding sponsor in the study design, collection, analysis,interpretation of data, writing the manuscript or in the decision to submit the manu-script for publication was null.

ContributorsAuthors António Macedo, Maria João Soares and José Valente designed the study

and wrote the protocol; authors Sandra Carvalho Bos and António Macedo conductedliterature searches and provided summaries of previous research studies; authorsMaria João Soares and Sandra Carvalho Bos conducted the statistical analysis; authorSandra Carvalho Bos wrote the first draft of the manuscript; and all authors, SandraCarvalho Bos, Maria Joao Soares, Mariana Marques, Telma Pereira, Berta Maia, VascoNogueira, Jose Valente, Antonio Macedo contributed to and have approved the finalmanuscript.

Conflict of interestThe authors declare no conflict of interest.

AcknowledgmentsThe data for this report were drawn from a research on Perfectionism and Obsessive–

Compulsive Spectrum Disorders, funded by the Portuguese Foundation for Science andTechnology (Grant: nr. 37569/PSI). The co-operation of Professors and Students is grate-fully acknowledged.

References

Allison, K. C., Lundgren, J. D., O'Reardon, J. P., Geliebter, A., Gluck, M. E., Vinai, P., et al.(2010). Proposed diagnostic criteria for night eating syndrome. International Journalof Eating Disorders, 43, 241–247.

American Academy of Sleep Medicine—AASM (2005). International classification ofsleep disorders. Diagnostic and coding manual (2nd ed.). Westchester, IL: AmericanAcademy of Sleep Medicine.

American Psychiatric Association—APA (1994). Diagnostic and statistical manual of mentaldisorders, DSM-IV (4th ed.). Washington, DC: American Psychiatric Association.

Azevedo, M. H., Bos, S. C., Soares, M. J., Marques, M., Pereira, A. T., Maia, B., et al. (2010).Longitudinal study on perfectionism and sleep disturbance. The World Journal ofBiological Psychiatry, 11, 476–485.

Benca, R. M., & Schenck, C. H. (2005). Sleep and eating disorders. InM. H. Kryger, T. Roth, &W. C. Dement (Eds.), Principles and practice of sleepmedicine (pp. 1337–1344) (4th ed.).Philadelphia: Elsevier Saunders.

Bjortvatn, B., Sagen, I. M., Oyane, N., Waage, S., Fetveit, A., Pallesen, S., et al. (2007). Theassociation between sleep duration, body mass index and metabolic measures inthe Hordaland Health Study. Journal of Sleep Research, 16, 66–76.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.Colles, A. H., Dixon, J. B., & O'Brien, P. E. (2007). Night eating syndrome and nocturnal

snacking: association with obesity, binge eating and psychological distress. Interna-tional Journal of Obesity, 31, 1722–1730.

Crispim, C. A., Zalcman, I., Dáttilo, M., Padilha, H. G., Edwards, B., Waterhouse, J., et al.(2007). The influence of sleep and sleep loss upon food intake and metabolism.Nutrition Research Reviews, 20, 195–212.

Crispim, A. C., Zalcman, I., Padilha, H. G., Dáttilo, M., & Túlio de Mello, M. (2008). Sono easpectos nutricionais. In S. Tufik (Ed.), Medicina e Biologia do Sono (pp. 406–415).São Paulo: Manole.

Delvenne, V., Kerkhofs, M., Appelboom-Fondu, J., Lucas, F., & Mendlewicz, J. (1992).Sleep polygraphic variables in anorexia nervosa and depression: A comparativestudy in adolescents. Journal of Affective Disorders, 25, 167–172.

Drake, C. L., Roehrs, T., & Roth, T. (2003). Insomnia causes, consequences and therapeutics:An overview. Depression and Anxiety, 18, 163–176.

Flett, G. L., Hewitt, P. L., Blankstein, K. R., & Gray, L. (1998). Psychological distress andthe frequency of perfectionistic thinking. Journal of Personality and Social Psychology,75, 1363–1381.

Garner, D. M., & Garfinkel, P. E. (1979). The eating attitudes test: An index of symptomsof anorexia nervosa. Psychological Medicine, 9, 273–279.

Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982). The eating attitudestest: Psychometric features and clinical correlates. Psychological Medicine, 12,871–878.

Garner, D. M., Vitousek, K. M., & Pike, K. M. (1997). Cognitive-behavioral therapy foranorexia nervosa. In D. M. Garner, & P. E. Garfinkel (Eds.), Handbook of treatmentfor eating disorders (pp. 94–144) (2nd ed.). New York: The Guilford Press.

Hewitt, P. L., & Flett, G. L. (1991). Perfectionism in the self and social contexts: Conceptu-alization, assessment, and association with psychopathology. Journal of Personalityand Social Psychology, 60, 456–470.

198 S.C. Bos et al. / Eating Behaviors 14 (2013) 192–198

Howell, M. J., Shenck, C. H., & Crow, S. J. (2009). A review of nighttime eating disorders.Sleep Medicine Reviews, 13, 23–34.

Investigators, The McKnight (2003). Risk factors for the onset of eating disorders inadolescent girls: Results of the McKnight longitudinal risk factor study. The AmericanJournal of Psychiatry, 160, 248–254.

Keys, A. (1948). Caloric undernutrition and starvation, with notes on protein deficiency.JAMA, 16, 500–511.

Lacey, J. H., Crisp, A. H., Kalucy, R. S., Hartmann, M. K., & Chien, C. N. (1975). Weightgain and the sleeping electroencephalogram: study of 10 patients with anorexianervosa. British Medical Journal, 6, 556–558.

Lauer, C. J., & Krieg, J. C. (2004). Sleep in eating disorders. Sleep Medicine Reviews, 8,109–118.

Levy, A. B., Dixon, K. N., & Schmidt, H. (1988). Sleep architecture in anorexia nervosaand bulimia. Biological Psychiatry, 23, 99–101.

Macedo, A., Soares, M. J., Azevedo, M. H., Gomes, A., Pereira, A. T., Maia, B., et al. (2007).Perfectionism and eating attitudes in Portuguese university students. EuropeanEating Disorders Review, 15, 296–304.

Makino, M., Hashizume, M., Yasushi, M., Tsuboi, K., & Dennerstein, L. (2006). Factorsassociated with abnormal eating attitudes among female college students in Japan.Archive of Women's Mental Health, 9, 203–208.

Molnar, D. S., Reker, D. L., Culp, N. A., Sadava, S. W., & DeCourville, N. H. (2006). Amediated model of perfectionism, affect, and physical health. Journal of Research inPersonality, 40, 482–500.

Nobili, L., Baglietto, M. G., De Carli, F., Savoini, M., Schiavi, G., Zanotto, E., et al. (1999). Aquantified analysis of sleep electroencephalography in anorectic adolescents.Biological Psychiatry, 45, 771–775.

Nolan, L. J., & Geliebter, A. (2012). Night eating is associated with emotional and externaleating in college students. Eating Behaviors, 13, 202–206.

Ohayon, M. M. (2002). Epidemiology of insomnia: What we know and what we stillneed to learn. Sleep Medicine Reviews, 6, 97–111.

Passani, M. B., & Blandina, P. (2011). Histamine receptors in the CNS as targets for thera-peutic intervention. Trends in Pharmacological Sciences, 32, 242–249.

Pereira, A. M. S. (2004). Concepções e práticas metodológicas em diferentes paradigmasde investigação. In L. Oliveira, A. Pereira, & R. Santiago (Eds.), Investigação emEducação. Abordagens conceptuais e práticas (pp. 47–58). Porto: Porto Editora.

Perlis, M. L., Smith, M. T., & Pigeon, W. R. (2005). Etiology and pathophysiology ofinsomnia. In M. Kryger, T. Roth, & W. C. Dement (Eds.), Principles and practice ofsleep medicine (pp. 714–725). Philadelphia: Saunders Company.

Pieters, G., Theys, P., Vandereycken, W., Leroy, B., & Peuskens, J. (2004). Sleep variablesin anorexia nervosa: Evolution with weight restoration. International Journal ofEating Disorders, 35, 342–347.

Seigel, K., Broman, J. -E., & Hetta, J. (2004). Problemas de sueño y síntomas detranstornos de la conducta alimentaria (TCA) en mujeres jóvenes. European Journalof Psychiatry, 18, 5–13.

Soares, M. J., Macedo, A., Bos, S. C., Maia, B., Marques, M., Pereira, A. T., et al. (2011).Sleep disturbances, body mass index and eating behavior in undergraduate students.Journal of Sleep Research, 20, 479–486.

Soares, M. J., Macedo, A., Bos, S. C., Marques,M., Maia, B., Pereira, A. T., et al. (2009). Perfec-tionism and eating attitudes in Portuguese students: A longitudinal study. EuropeanEating Disorders Review, 7, 390–398.

Soares, M. J., Macedo, A. F., Gomes, A. A., & Azevedo, M. H. P. (2004). A VersãoPortuguesa do Teste de Atitudes Alimentares-40. Psiquiatria. Clínica, 18, 11–24.

Striegel-Moore, R. H., Franko, D. L., May, A., Ach, E., Thompson, D., & Hook, J. M. (2006).Should night eating syndrome be included in the DSM? International Journal of EatingDisorders, 39, 544–549.

Walsh, B. T., Goetz, R., Roose, S. P., Fingeroth, S., & Glassman, A. H. (1985).EEG-monitored sleep in anorexia nervosa and bulimia. Biological Psychiatry, 20,947–956.

Winsky-Sommerer, R., Yamanaka, A., Diano, S., Borok, E., Roberts, A. J., Sakurai, T., et al.(2004). Interaction between the corticotropin-releasing factor systemand hypocretins(orexins): A novel circuit mediating stress response. Journal of Neuroscience, 24(50),11439–11448 (15).

World Health Organization,WHO (WHO, 1995). Physical status: The use and interpretationof anthropometry. Report of a WHO Expert Committee. Geneva: World HealthOrganization.