On the role of sadness in the psychopathology of anorexia nervosa

Eva Naumann a,n, Brunna Tuschen-Caffier a, Ulrich Voderholzer b, Jennifer Svaldi a

a University of Freiburg, Department of Clinical Psychology and Psychotherapy, Engelbergerstrasse 41, 79106 Freiburg, Germanyb Schoen Clinic Roseneck, Prien, Germany

a r t i c l e i n f o

Article history:Received 5 July 2013Received in revised form14 December 2013Accepted 27 December 2013Available online 4 January 2014

Keywords:Eating disordersAnorexia nervosaNegative emotionSadnessDietary restrictionDesire to exercise

a b s t r a c t

Recent models on the development and maintenance of eating disorders propose negative emotions tobe important precursors for the occurrence of eating disorder symptomatology. In fact, previous researchon bulimia nervosa (BN) and binge eating disorder provides evidence that negative emotions are anantecedent condition for binge eating. However, there is a lack of research examining the influence ofnegative emotions on restrictive eating and exercising in individuals with anorexia nervosa (AN). In anexperimental study, women with AN (n¼39) and BN (n¼34) as well as a non-eating disordered controlgroup (CG; n¼34) watched a sadness-inducing film clip. Before and after the film clip participants ratedtheir current desire to engage in dietary restriction (DTR) and desire to exercise (DTE). Main resultsreveal that DTR significantly increased after the film clip in women with AN only, while DTE decreasedover time in all groups. Results are in line with the notion that negative emotions have a prominentinfluence on the core eating pathology in AN.

& 2014 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

According to the Diagnostic and Statistical Manual of MentalDisorders (DSM-5; American Psychiatric Association (APA), 2013)anorexia nervosa (AN) is an eating disorder characterized by aphobic fear of gaining weight along with an unwillingness tomaintain the minimal healthy body weight. Typical behaviors toachieve or maintain underweight include caloric restriction whichsometimes even culminates in chronic self-imposed starvation.The severe self-starvation is one reason for the increased morbid-ity and mortality associated with AN (Fichter et al., 2006; Mitchelland Crow, 2006; Berkman et al., 2007). With a lifetime prevalenceof up to 80% extreme physical activity is another common weight-reducing behavior in individuals suffering from AN (Davis et al.,1997). While a regular amount of exercise has a beneficial effect onmood and physical health (Blumenthal et al., 2007; Deslandeset al., 2009; Archer, 2012; Morris et al., 2012), research shows thatexcessive exercise in AN is positively associated with eatingpsychopathology, hospitalization periods and rates of relaps(Strober et al., 1997; Solenberger, 2001; Peñas-Lledó et al., 2002;Carter et al., 2004; Bewell-Weiss and Carter, 2010). As such, thereis a need for research identifying factors that trigger restriction offood intake and excessive exercise in AN.

Models on the maintenance of eating disorders highlight theimportance of negative emotions in the occurrence of eating

disorder symptomatology (Stice, 2001; Fairburn et al., 2003). Assuch, a great body of research exists showing that negativefeelings trigger binge eating and purging behavior in bulimianervosa (BN) as well as binge attacks in binge eating disorder(BED) (Alpers and Tuschen-Caffier, 2001; Hilbert and Tuschen-Caffier, 2007; Smyth et al., 2007). For example, after an insecurityand sadness-inducing guided imagery task ratings of hunger anddesire to binge increased in bulimic patients, whereas no changeswere found in healthy controls (Tuschen-Caffier and Voegele,1999).

In contrast to BN and BED, research on the role of negativeemotions in the maintenance of anorectic behavior is still sparse.Besides, on a theoretical level it remains controversial whethereating pathology in AN can be influenced by intense mood states.Some theoretical accounts suggest that mood intolerance andmaladaptive emotion regulation behavior is rather atypical inindividuals with AN compared to other eating disorders (Fairburnet al., 2003). On the other hand, there are recent models on theonset and maintenance of AN that emphasize the emotion regula-tion functioning of anorectic behavior (Haynos and Fruzzetti, 2011).That is, maladaptive behaviors as dietary restriction and excessiveexercise may be triggered by adverse emotional states and mighttherefore serve as a dysfunctional form of emotion regulation.Empirical evidence supports this notion showing that individualswith AN have more difficulties tolerating and regulating negativeemotions compared to healthy controls (Harrison et al., 2009;Wildes et al., 2010; Svaldi et al., 2012). Moreover, one study(Merwin et al., 2010) reports significant positive correlations ofnon-acceptance of emotional responses with dietary restriction in

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/psychres

Psychiatry Research

0165-1781/$ - see front matter & 2014 Elsevier Ireland Ltd. All rights reserved.http://dx.doi.org/10.1016/j.psychres.2013.12.043

n Corresponding author. Tel.: þ49 761 2033016; fax: þ49 761 2033022.E-mail address: eva.naumann@psychologie.uni-freiburg.de (E. Naumann).

Psychiatry Research 215 (2014) 711–717

women with AN. In qualitative studies AN patients report that mainreasons for restriction of food intake include the inability to tolerateadverse emotions, attempts to control negative emotions and effortsto provide positive feelings like pride and security (Serpell et al.,1999; Dignon et al., 2006; Nordbø et al., 2006; Federici and Kaplan,2008). Additionally, an ecological momentary assessment (EMA)study (Engel et al., 2005) revealed a prominent correlation betweenstressful events and affect lability with restrictive behavior andrituals in AN. Further evidence for the reinforcing influence ofnegative mood on anorectic eating behavior stems from biopsycho-logical studies showing that dietary restraint can have a decreasingeffect on the plasma tryptophan availability (Kaye, 2008). Theplasma tryptophan modulates brain 5-HT functional activity, whichin turn, is thought to have a positive influence on mood (Frank et al.,2001).

Regarding excessive exercise, there is indirect evidence of aclose linkage between negative emotions and extreme physicalactivity in individuals with AN. For example, Peñas-Lledó et al.(2002) found that subjects with AN subtyped as high excessiveexercisers were characterized by high levels of depression andanxiety compared to those subtyped as low excessive exercisers.Likewise, an EMA study on patients with AN and BN showed thatindividuals with an increased desire to be physically active aredisposed to suffer from a chronically negative affect (Vansteelandtet al., 2007).

Notwithstanding, the previously mentioned studies do not allowto draw conclusions about cause and effect. In a very recent studythough Wildes et al. (2012) experimentally tested the effects ofnegative mood on eating disorder symptoms in patients with AN.Specifically, AN participants either watched a negative emotion-inducing film clip or a neutral film clip. Prior to and after the filmclip, participants had to rate several self-constructed items on Likertscales ranging from one (agree not at all) to five (completely agree).The items focused on maladaptive thoughts about eating, shape andweight (e.g., I feel fat), as well as urges to engage in eating disorderbehaviors typically associated with AN (e.g., I want to restrict).While no changes were found in AN participants allocated to theneutral condition, those allocated to the negative mood conditionself-reported a significant increase in eating disorder symptomsfollowing the negative emotion induction. However, as this studysummed up different kinds of eating disorder pathology (e.g., urgeto restrict, state body image disturbance) it remains unclear whichof the specific anorectic symptoms were actually triggered by theinduced adverse emotions. This is especially important as previousstudies were able to show that negative mood leads to an over-estimation of the own body size as well as higher levels of bodydissatisfaction (Plies and Florin, 1992; Baker et al., 1995). Therefore,it is possible that only changes in the items measuring body imagedisturbances were responsible for the results of Wildes et al. (2012).

In light of the research just mentioned, the aim of the presentstudy was to experimentally test the influence of negative emotionson the urge to engage in dietary restriction (DTR) and exercise (DTE)in patients with AN. Based on former studies identifying sadness tobe a common pre-binge emotion in BN and BED (Cooper andBowskill, 1986; Chua et al., 2004), a sadness-inducing film clip wasused as negative emotional stimulus. In addition to an AN group, weincluded both a non-eating disordered control group and a group offemales with BN. The former was included to test for possible effectsof mood on eating pathology. The latter was included in order toprovide information about differential effects between the two eatingdisorders, as previous studies reported high rates of syndrome shift(Agras et al., 2000) and similarities (Norman and Herzog, 1983)between AN and BN. With regard to the extensive overlap in thediagnostic criteria particularly between AN of bulimic subtype and BN(American Psychiatric Association (APA), 2013), analyses comparingAN restrictive and bulimic subtypes were also included to further

explore whether the affect regulation model fits both AN subtypes.In line with current models on the onset and maintenance ofAN (Haynos and Fruzzetti, 2011), we predicted that an induction ofsadness would lead to an increase of DTR and DTE only in individualswith AN.

2. Method

2.1. Participants

The sample consisted of women with a DSM-IV-TR (American PsychiatricAssociation (APA), 2000) diagnosis of AN (n¼39) and BN (n¼34) recruited froman inpatient clinic and via advertisements and announcements in the local media.The non-eating disordered control group (n¼34) was recruited via ads in the localmedia. Twenty-nine (74.35%) AN subjects were classified as restrictive subtype and10 (26.31%) as bulimic subtype.

Exclusion criteria for all participants included the presence of current sub-stance abuse or addiction (except sustained full remission), bipolar disorder, pastpsychosis, schizophrenia (currently symptomatic) and current suicidal ideation.Eating disorders were diagnosed by means of the Eating Disorder ExaminationInterview (EDE; Fairburn and Cooper, 1993; Hilbert et al., 2004). All other diagnoseswere established by means of the Structured Clinical Interview (SCID) for DSM-IVAxis I and the borderline personality section of the Axis II interview (First et al.,1997; Wittchen et al., 1997). Studies show that the prevalence of eating disordersis high in adolescent females (e.g., Hudson et al., 2007; Le Grange et al., 2012).Therefore, to increase the ecological validity of the study the lower age limit forparticipants was 16 years.

2.2. Materials

2.2.1. Induction of sadnessIn order to induce sadness a scene from the movie “The Lion King” (length:

2 min, 11 s) was selected in which a young lion finds his father dead. As part of a setof standardized emotional film stimuli, this movie scene from “The Lion King” hasbeen shown to reliably elicit sadness with little influence on other negativeemotions (Rottenberg et al., 2007). The following instruction adapted from Gross(1998) was presented prior to the clip: “We will now be showing you a short filmclip. It is important to us that you watch the film clip carefully. However, if you findthe film to be too distressing, please call the experimenter.” To intensify sadness,a freeze image was presented after the film clip for 15 s, in which the young lionsnuggles up to his dead father.

2.2.2. Manipulation check itemsOn the day of the experiment, participants were instructed to eat 1–2 h prior to

the experiment. However, during the hour before the experiment no food intakewas allowed to control for hunger.

To check whether participants really had comparable initial hunger levels,participants rated how hungry they were on a 100 mm Visual Analog Scale (VAS)anchored from “not at all” to “very much” at the beginning of the experiment.

To ensure that the film clip induced sadness, emotional states were rated by theparticipants before and after the film-clip on 100 mm VASs anchored from “not atall” to “very much”. State sadness was assessed by the item “At the moment I feelsad”. Current positive emotion was assessed by the following item: “At the momentI feel happy”.

2.2.3. Self-reported desire to engage in dietary restriction (DTR) and desire to exercise(DTE)

Like the manipulation check items, items on DTR and DTE were presented viacomputer on 100 mm VASs anchored from “not at all” to “very much”. Items werepresented prior to (baseline) and after watching the film clip (post-film).

Current DTR was assessed using the following self-constructed item: Atthe moment I do not want to eat anything. The item “At the moment I would liketo exercise” was used to measure the current DTE.

2.2.4. QuestionnairesInternal consistencies of our study sample for the questionnaires were

calculated for the entire sample. The following questionnaires were administered:(1) The Eating Disorder Examination Questionnaire (EDE-Q; Fairburn and Beglin,1994) is a 36-item self-report measure that assesses the severity of eatingpathology with four subscales (restraint eating, eating concern, weight concernand shape concern) and with a global score. The global score and the subscalesshow high internal consistency, stability and validity (Fairburn and Beglin, 1994;Hilbert et al., 2007). Internal consistencies in our study ranged from α¼0.91 (EDE-Qweight concern) to α¼0.98 (EDE-Q global score). (2) The current version ofthe Beck Depression Inventory-II (BDI Beck et al., 1996; Hautzinger et al., 2007) isa self-report measure that assesses severity of depression over the last two weeks.

E. Naumann et al. / Psychiatry Research 215 (2014) 711–717712

It consists of 21 items with a scale ranging from 0 to 3 and shows high internalconsistency (α¼0.84), test-retest reliability (rtt¼0.75) and discriminant validity(Kuehner et al., 2007). The internal consistency in our study was α¼0.96.

2.3. Procedure

Diagnostics and experimental sessions were conducted on separate days withno more than one week in-between. Participants were informed that they will bepresented with a number of film clips and pictures in the context of differentexperiments revolving around their emotional reactions. All subjects gave theirinformed consent before taking part in the experiment. In the case of under agedindividuals written informed consent of one parental unit was required in addition.The study was approved by the local ethic committee.

Presentation software (Neurobs, Inc., Albany, California, USA) was used for thepresentation of the instructions, stimulus delivery and to record participants0

subjective ratings. On arrival at the experimental session, participants were lead toa quiet laboratory room, seated in front of a 17-inch monitor and were givenheadphones. After instructing participants on the experiment, the laboratoryassistant left the participant alone in the room and started the experiment. Atfirst, participants rated their current emotional states, DTR, and DTE (baseline).Then, the instruction to carefully watch the following film clip was presented (seeSection 2.2.1). To ensure that participants had enough time to read and understandthis instruction, they had to press a button in order to continue with theexperiment. Then, the film clip and the still image of the film clip were presented.Afterwards, subjective ratings (post-film) were re-assessed. After this segment,participants made a pause and then continued with other experimental assign-ments unrelated to this paper.

2.4. Statistical analyses

Socio-demographic characteristics and manipulation check on baseline hungerwere analyzed by means of univariate analyses of variance (ANOVAs). Manipulationcheck on the sadness induction and hypotheses were tested by means of repeatedmeasures ANOVAs.1 If the assumption of sphericity was not met (MauchlysSphericity Test: po0.05), degrees of freedom for dependent variables werecorrected conservatively by Greenhouse-Geisser. Being exceedingly robust againstviolation of normality (Tabachnick and Fidell, 2007), ANOVAs were also adopted forvariables deviating from normal distribution. Where indicated, Scheffé post-hoctests were conducted to identify pairwise differences among the groups. Effectsizes of the main effects and interactions are reported by partial eta squared (η2),whereby values larger than 0.01 refer to small, 0.06 to moderate, and 0.14 to largeeffect sizes (Cohen, 1988). Effect sizes for post-hoc analyses are reported by Cohen0sd (Cohen, 1988), whereby values larger than 0.20 refer to small, 0.50 to moderate,and 0.80 to large effect sizes. All tests of significance were two tailed.

3. Results

3.1. Socio-demographics and questionnaires

Groups did not differ significantly in age, marital status andeducational level. As expected, Body Mass Index (BMI¼weight/height²) in the AN group was significantly lower compared to theother two groups, which did not differ from one another in termsof weight. Regarding eating pathology and depression, there wereno significant differences between the two eating disorder sam-ples, while both groups had significantly higher scores on therespective scales compared to the CG (see Table 1 for means (M),standard deviations (S.D.), frequencies and statistics).

3.2. Manipulation check

With regard to the initial hunger, an ANOVA revealed that therewere no significant differences in hunger levels at the baselinemeasurement between the groups, F (2, 104)¼0.17, p¼0.84 (ANgroup: M¼10.20, S.D.¼17.20; BN group: M¼12.38, S.D.¼21.45;CG: M¼10.00, S.D.¼17.23).

A 3 (Group: AN, BN, CG)�2 (Time: baseline, post-film) repeatedmeasures ANOVA on sadness revealed a significant main effect oftime, F (1, 104)¼196.92, po0.001, η2¼0.65, whereby in all groups

subjective sadness significantly increased after watching the filmclip (baseline: M¼28.85, S.D.¼30.90; post-film: M¼71.79, S.D.¼27.41). In addition, the main effect of group was significant, F(2, 104)¼23.83, po0.001, η2¼0.31. Scheffe0s post-hoc analysesshowed that there were no significant differences in subjectivesadness between AN (baseline: M¼34.65, S.D.¼32.77; post-film:M¼78.10, S.D.¼25.20) and BN (baseline: M¼44.20, S.D.¼29.91;post-film: M¼81.38, S.D.¼18.41), p¼0.42, while both eating dis-order groups reported significantly higher state sadness at thebaseline and the post-film measurement compared to the CG(baseline: M¼6.85, S.D.¼12.40; post-film: M¼54.97, S.D.¼30.15),pso0.001, ds41.33. There was no significant group� time interac-tion, F (2, 104)¼1.03, p¼0.36.

A repeated measures group (AN, BN, CG)� time (baseline, post-film) ANOVA conducted on positive emotion yield a significant maineffect of time, F (1, 104)¼182.16, po0.001, η2¼0.64, and group, F (2,104)¼15.97, po0.001, η2¼0.24, while there was no significantgroup� time interaction, F (2, 104)¼2.94, p¼0.06. Thereby, positiveemotion decreased significantly in all groups (baseline:M¼43.84, S.D.¼25.15; post-film: M¼16.15, S.D.¼22.13). Scheffe0s post-hoc testsrevealed no significant differences between AN (baseline: M¼42.31,S.D.¼26.38; post-film: M¼16.15, S.D.¼22.13) and BN (baseline:M¼37.54, S.D.¼25.94; post-film: M¼15.27, S.D.¼18.08) in positiveemotions, p¼0.42, while both eating disorder groups had signifi-cantly lower state positive emotion at the baseline and the post-filmassessment in comparison to the CG, (baseline: M¼69.22,S.D.¼17.30; post-film: M¼37.80, S.D.¼18.64), pso0.001, ds41.00.

3.3. Desire to engage in dietary restriction (DTR)

A 3 (group: AN, BN, CG)�2 (time: baseline, post-film) repeatedmeasures ANOVA conducted on DTR yield a significant main effectof group, F (2, 104)¼12.60, po0.001, η2¼0.20, whereby the AN andBN group did not significantly differ in their DTR levels, p¼0.27,while both eating disorder groups had significantly higher DTRscores compared to the CG, pso0.007, ds40.78. There also was asignificant main effect of time, F (1, 104)¼11.38, p¼0.001, η2¼0.10,and a significant group� time interaction, F (2, 104)¼6.09, p¼0.003,η2¼0.11. To further explore the significant main effect of time and thesignificant group� time interaction, follow-up paired t-tests on theDTR baseline and DTR post-film scores were conducted separatelyfor group. Follow-up paired t-tests showed that only in the AN groupthere was a significant increase in the DTR after the sadness-inducingfilm clip, t¼�5.03, d.f.¼38, po0.001, d¼�0.72. There was nosignificant change in the DTR from prior to after the film clip inthe BN group, t¼�0.769, d.f.¼33, p¼0.45, and the CG, t¼�0.23, d.f.¼33, p¼0.82 (see Fig. 1a for Ms and S.D.s).

In order to check for possible differences in DTR changebetween the two AN subtypes, a 2 (subtype: restrictive AN, bulimicAN)�2 (time: baseline, post-film) repeated ANOVA was con-ducted on DTR. The results revealed no significant main effect ofsubtype, F (1, 37)¼3.24, p¼0.08, and no significant subtype� timeinteraction, F (1, 37)¼0.14, p¼0.71. There was a significant maineffect of time, F (1, 37)¼20.48, po0.001, η2¼0.36, showing thatboth the restrictive AN group (baseline: M¼46.62, S.D.¼37.22;post-film: M¼69.62, S.D.¼31.78) and the bulimic AN group (base-line: M¼63.60, S.D.¼31.22; post-film: M¼90.70, S.D.¼16.47)showed an increase in DTR in response to a sadness-inducingfilm clip.

3.4. Desire to exercise (DTE)

A group (AN, BN, CG)� time (baseline, post-film) repeatedmeasures ANOVA conducted on DTE revealed a significant maineffect of group, F (2, 104)¼3.37, p¼0.03, η2¼0.06, whereby womenwith AN reported significantly higher DTE scores than the CG1 The results of ANCOVAs with BDI as a covariate yielded comparable results.

E. Naumann et al. / Psychiatry Research 215 (2014) 711–717 713

throughout the experiment, p¼0.03, d¼0.62, while there were nodifferences between the eating disorder groups or the CG and theBN group, po0.34. In addition, there was a significant main effect oftime, F (1, 104)¼8.59, p¼0.004, η2¼0.08. Thereby, all groupsshowed a significant decrease in DTE after watching the film-clip.There was no significant group� time interaction, F (2, 104)¼0.98,p¼0.34. (see Fig. 1b for Ms and S.D.s).

4. Discussion

Current models of AN propose that negative emotions play animportant role in the maintenance of anorectic behavior (Haynosand Fruzzetti, 2011). While much efforts have been invested toanalyze the relation between negative emotions and binge eatingin BN and BED (Alpers and Tuschen-Caffier, 2001; Hilbert andTuschen-Caffier, 2007; Smyth et al., 2007), there is a lack of

research on the influence of adverse feelings in the occurrenceof eating disorder symptoms in AN. Therefore, the present studywas designed to experimentally investigate the effects of sadnesson the DTR and DTE as two common weight-reducing behaviors inAN. For this purpose, a sadness-inducing film clip was presented towomen with AN, BN and a non-eating disordered control group.Prior to and after the film clip participants rated their currentemotions as well as their DTR and DTE on VASs. It was predictedthat only in females with AN the DTR and the DTE would increaseafter inducing sadness.

With regard to the main hypothesis of the study, the resultsreveal that compared to the BN and non-eating disordered CG, theDTR significantly increased after the sadness-inducing film clip inwomen with AN. In line with recent models on the developmentand maintenance of anorectic behavior (Haynos and Fruzzetti,2011) and previous studies (Engel et al., 2005; Wildes et al., 2012),these results support the notion that caloric restriction in AN is

Table 1Socio-demographics and overall psychopathology presented separately for participants with anorexia nervosa (AN), bulimia nervosa (BN) and non-eating disorderedcontrols (CG).

AN n¼39 BN n¼34 CG n¼34 Statistics Scheffé post-hoc tests

M (S.D.)/Frequency M (S.D.)/Frequency M (S.D.)/Frequency F or χ2 p p (AN vs. BN) p (AN vs. CG) p (BN vs. CG)

Age (years) 25.54 (11.06) 25.94 (8.34) 25.88 (10.73) 0.02 0.98 – – –

Marital status χ2 (6)¼11.07 0.09 – – –

Single 23 17 10Partnership 12 15 17Married 3 2 3Divorced 1 0 4

Education level χ2 (4)¼1.26 0.87 – – –

Low education 2 2 1Middle education 16 11 11High education 21 21 22

BMI 15.05 (1.89) 22.44 (4.20) 21.23 (2.14) 69.60 o0.001 o0.001 o0.001 0.23EDE-QRE 5.27 (1.69) 4.71 (1.53) 1.33 (0.61) 83.74 o0.001 0.23 o0.001 o0.001EDE-QEC 4.47 (1.72) 4.71 (1.32) 1.11 (0.15) 84.74 o0.001 0.73 o0.001 o0.001EDE-QWC 4.87 (1.60) 5.25 (1.20) 1.44 (0.63) 99.76 o0.001 0.43 o0.001 o0.001EDE-QSC 5.38 (1.41) 5.82 (0.99) 1.78 (0.65) 144.14 o0.001 0.23 o0.001 o0.001EDE-QGS 5.07 (1.44) 5.24 (1.02) 1.48 (0.51) 132.00 o0.001 0.81 o0.001 o0.001BDI 28.23 (10.53) 25.94 (10.11) 3.20 (2.61) 89.24 o0.001 0.53 o0.001 o0.001

Note: BMI¼body mass index (weight/height2); EDE-Q¼eating disorder examination questionnaire; RE¼restraint subscale; EC¼eating concerns subscale; WC¼weightconcerns subscale; SC¼shape concerns subscale; GS¼global score; and BDI¼beck depression inventory.

0

20

40

60

80

100

baseline post-film

Mea

n D

TR

AN

BN

CG

0

20

40

60

80

baseline post-film

Mea

n D

TE

AN

BN

CG

Fig. 1. Means on the desire to engage in dietary restriction (DTR; (a)) and desire to exercise (DTE; (b)) before and after a sadness-inducing film clip. Results are presentedseparately for women with anorexia nervosa (AN), bulimia nervosa (BN) and non-eating disordered controls (CG). Bars represent standard deviations.

E. Naumann et al. / Psychiatry Research 215 (2014) 711–717714

triggered, and therefore possibly maintained by negative emo-tional states. Thus, similarly to BN and BED (Alpers and Tuschen-Caffier, 2001; Hilbert and Tuschen-Caffier, 2007; Smyth et al.,2007) results suggest that individuals with AN might engage indietary restriction in order to escape from or control negativeaffect.

Noteworthy, the present study found no changes in DTR aftersadness induction in non-eating disordered controls and womenwith BN. This is an interesting finding, given the fact that caloricrestriction is also highly prevalent in BN (Rossiter et al., 1989). It ispossible that the emotion regulation functioning of dietary restric-tion is rather specific for patients with AN. Interestingly, whenanalyzing the two AN subtypes separately, we found that ANpatients of bulimic and restrictive subtype both reacted to the filmclip with an increase in their DTR. Though AN females of thebulimic subtype also suffer from binge eating and purging, ourresults indicate that within the emotion regulation domain theymight have more in common with their restrictive counterparts.Thus, these results provide further support for the validity andrelevance of the diagnostic system regarding the classification ofAN subtypes.

Contrary to our hypothesis, DTE significantly decreased afterthe sadness-eliciting film clip in all participants. These findingsstand in contrast to the notion that physical activity serves anaffect regulation functioning in AN (Haynos and Fruzzetti, 2011) aswell as studies showing that excessive exercise in AN is positivelyconnected to chronic negative emotions (Peñas-Lledó et al., 2002;Vansteelandt et al., 2007). However, it is possible that our resultson the DTE are highly influenced by the side effects of the sadnessinduction. That is, previous research has shown that sadness isassociated with mental and physical passivity (Schwartz et al.,1981; Rucken and Petty, 2004), which in short term might haveinhibited the DTE in all participants. However, participants of thecurrent study rated their DTE directly after they had watched thefilm clip, i.e., when a strong feeling of sadness was still present.It therefore remains unclear whether individuals with AN usephysical activity to cope with their depressive mood at a later levelof the sadness generation process. It would be interesting in futureto also examine the long term effects of sadness on exercise ineating disorders. Furthermore, it would be interesting for futureresearch to induce other negative emotions than sadness, for it ispossible that different emotions lead to different eating disordersymptoms. For example, anger as a further highly prevalentemotion in eating disorders (Waller et al., 2003) is suggestedto be associated with a high action readiness and heightenedphysiological arousal (Schwartz et al., 1981; Frijda, 1987), andtherefore possibly is closer related to exercising in eating disordersthan sadness. Another possible explanation for our results on theDTE is that the conceptualization of the corresponding item wasvery broad, given the diversity of the concept (e.g., there is adifference between an urge to move, a desire to engage in sports, afeeling of restlessness). Therefore, future studies should considerusing more specific items to operationalize DTE. Nevertheless,there also is literature emphasizing that the drive for activity in ANrather results from biological consequences of the malnutritionstatus of the patients, and thus is relative unrelated to emotionalstates (Casper, 2006). In fact, animal research displays thatstarvation in rats leads to hyperactivity and increased exercisingbehavior (Pirke et al., 1993); a connection that is supposed to bemediated by the low plasma leptin levels found in food-restrictedrats and patients with AN (Balligand et al., 1998; Hebebrand et al.,2003; Holtkamp et al., 2003). However, correlational analyseson the current samples showed that there were no significantassociations of BMIs and the DTE ratings (all rso�0.326, allps40.060). This, along with research on the mood stabilizingeffect of physical activity (Blumenthal et al., 2007) as well as our

results on the significant change in the DTE after the sadnessinduction stand against the assumption that activity levels in ANare solely determined by the hypoleptinemia associated with foodrestriction. Hence, future studies are needed to further explore therole of emotions in the occurrence of exercise in eating disorders(e.g., by using EMA).

With regard to the limitations of the present study, it isimportant to critically note that only self-report measures wereincluded. In the context of bulimic spectrum eating disordersresearch has shown that binge eating is preceded by a strong desireto binge (Steiger et al., 1999; Engelberg et al., 2005). On the otherhand results of a study by Waters et al., 2001 illustrate that notevery urge to binge automatically leads to an actual binge attack.Similarly, in individuals diagnosed with AN the DTR or DTEpresumably is not always equivalent to real behavior, since externalfactors like social circumstances also need to be taken into account.Future studies that integrate more objective measures of restrictiveeating (e.g., real food intake in a laboratory task) and exercising(e.g., monitoring movement via a kinesiometer) are clearly needed.In this regard, it must be critically mentioned that our items also donot directly assess the motivation for the eating disordered beha-vior. As such, a self-reported increase in the item “At the momentI do not want to eat anything” may not necessarily be associatedwith a desire to regulate weight. More specifically, it is well possiblethat increases in DTR are merely a consequence of decreases inappetite following a negative emotional state rather than reflectingan increased desire to control or regulate weight. Therefore, futurestudies should more thoroughly focus on the motivational aspectsunderlying the desire to restrict food intake following a sadnessinduction. Additionally, these studies should definitely includemultiple and validated items to assess DTR and DTE to ensure andstrengthen the reliability of the results found. Ultimately, bogustaste tests implemented after sadness induction would improve theecological validity of the reported results. Likewise, EMA could beused to test the impact of sadness on exercise in AN.

A further limitation of the current study is that we have notcollected information on the time of the last meal, the last bingeeating episode, or the last time the participants engaged inexercise. Obviously, the engagement in such behavior could haveinfluenced our results. On the other hand, all participants wereinstructed to eat 1–2 h prior to the experiment and not to eatanything within 1 h prior to the experimental session. Moreimportant though, baseline levels of current hunger did not differwithin the two experimental conditions. Nevertheless, futurestudies should not solely rely on self-report data and rather servea standardized meal prior to the experimental session.

Another issue concerns the fact that no neutral film clip wasincluded in the experimental design. Thus, it is not clear whether theemotional content of the film stimulus was responsible for our results.That is, DTR in participants with AN might have increased for reasonsother than negative emotions. Such factors may include naturallyoccurring daily fluctuations or time of day. This seems unlikely,though, as participants were tested throughout the day. Also, againstthis assumption we point out that Wildes et al. (2012) showed eatingdisorder symptoms in AN not to change following a neutral film clip.

Last but not least, the low sample size of the group with ANbulimic subtype (n¼10) compared to the group with AN restrictivesubtype (n¼29) is a further limitation of the current study.Though this is in line with research showing that the restrictivesubtype is more prevalent than the bulimic subtype (Garfinkelet al., 1996), future studies with more balanced and higher samplesizes regarding the AN subtypes are needed to provide reliableinformation on whether both AN subtypes differ in their reactionsto negative emotions.

In sum, our results suggest that similarly to BN and BED, negativeemotions are an antecedent condition in the core pathology of AN as

E. Naumann et al. / Psychiatry Research 215 (2014) 711–717 715

well. However, further support from studies inducing differentemotions (e.g., anger, disgust, shame, anxiety) and using moredifferentiated assessments of dietary restriction and exercising (e.g., multiple specific items, including objective measures) is clearlyneeded. Even though our data suggest an increase of DTR bysadness, the underlying mechanisms remain yet unclear. In anempirically well validated model, Gross (1998) defined emotionregulation (ER) as “the process by which we influence what kinds ofemotions we have, when we have them, and how we perceive andexpress them” [Gross, 2002; p. 282]. Two frequently examined ERstrategies are expressive suppression and cognitive reappraisal.Along a temporal continuum, reappraisal occurs early, and shouldthus enable an individual to modify an emotional sequence beforeemotion response tendencies have been fully generated. Suppres-sion, on the other hand, occurs late in the emotion-generativeprocess and therefore requires an effortful management of emotionresponse tendencies as they continually arise. Evidence suggeststhat compared to individuals without AN, those with AN more oftensuppress and less often reappraise their emotions (Svaldi et al.,2012). As such, it can be assumed that patients with AN adopteddifferent emotion regulation strategies following the sadness induc-tion. To address this issue, studies that analyze the differentialimpact emotion regulation strategies have on DTR in women withAN are needed.

Taken together with our findings, this research could haveprominent implications for current clinical efforts to includetrainings of functional emotion regulation skills in the therapy ofAN. Future studies are needed to investigate whether the integra-tion of such emotion focused interventions improve the effective-ness of existing AN treatment programs.

Acknowledgments

We want to thank the Scientific Society of Freiburg for theirgenerous grant, which enabled us to complete this study. Wegratefully acknowledge the excellent support of the staff of theSchoen Clinic Roseneck in Germany. We thank all the participantsfor their commitment and dedication.

References

Agras, W.S., Walsh, T., Fairburn, C.G., Wilson, G.T., Kraemer, H.C., 2000. A multi-center comparison of cognitive-behavioral therapy and interpersonal psy-chotherapy for bulimia nervosa. Archives of General Psychiatry 57, 459–466.

Alpers, G.W., Tuschen-Caffier, B., 2001. Negative feelings and the desire to eat inbulimia nervosa. Eating Behaviors 2, 339–352.

American Psychiatric Association (APA), 2000. Diagnostic and Statistical Manual ofMental Disorders, fourth ed. American Psychiatric Association, Washington, DC.(Text revision).

American Psychiatric Association (APA), 2013. Diagnostic and Statistical Manual ofMental Disorders, fifth ed. American Psychiatric Publishing, Arlington, VA.

Archer, T., 2012. Influence of physical exercise on traumatic brain injury deficits:scaffolding effect. Neurotoxicity Research 21, 418–434.

Baker, J.D., Williamson, D.A., Sylve, C., 1995. Body image disturbance, memory bias, andbody dysphoria: effects of negative mood induction. Behavior Therapy 26, 747–759.

Balligand, J.L., Brichard, S.M., Brichard, V., Desager, J.P., Lambert, M., 1998.Hypoleptinemia in patients with anorexia nervosa: loss of circadian rhythmand unresponsiveness to short-term refeeding. European Journal of Endocri-nology/European Federation of Endocrine Societies 138, 415–420.

Beck, A.T., Steer, R.A., Brown, G.K., 1996. Manual for the beck depression inventory-II. Psychological Corporation, San Antonio.

Berkman, N.D., Lohr, K.N., Bulik, C.M., 2007. Outcomes of eating disorders: a systematicreview of the literature. International Journal of Eating Disorders 40, 293–309.

Bewell-Weiss, C.V., Carter, J.C., 2010. Predictors of excessive exercise in anorexianervosa. Comprehensive Psychiatry 51, 566–571.

Blumenthal, J.A., Babyak, M.A., Doraiswamy, P.M., Watkins, L., Hoffman, B.M.,Barbour, K.A, Herman, S., Craighead, W.E., Brosse, A.L., Waugh, R., Hinderliter,A., Sherwood, A., 2007. Exercise and pharmacotherapy in the treatment ofmajor depressive disorder. Psychosomatic Medicine 69, 587–596.

Carter, J.C., Blackmore, E., Sutandar-Pinnock, K., Woodside, D.B., 2004. Relapse inanorexia nervosa: a survival analysis. Psychological Medicine 34, 671–679.

Casper, R.C., 2006. The ‘drive for activity’ and “restlessness” in anorexia nervosa:potential pathways. Journal of Affective Disorders 92, 99–107.

Chua, J.L., Touyz, S., Hill, A.J., 2004. Negative mood-induced overeating in obesebinge eaters: an experimental study. International Journal of Obesity andRelated Metabolic Disorders: Journal of the International Association for theStudy of Obesity 28, 606–610.

Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences. vol. 2.Hillsdale, Lawrence Erlbaum, New Jersey.

Cooper, P.J., Bowskill, R., 1986. Dysphoric mood and overeating. British Journal ofClinical Psychology/The British Psychological Society 25, 155–156.

Davis, C., Katzman, D.K., Kaptein, S., Kirsh, C., Brewer, H., Kalmbach, K., Olmsted, M.P.,Woodside, D.B., Kaplan, A.S., 1997. The prevalence of high-level exercise in theeating disorders: etiological implications. Comprehensive Psychiatry 38, 321–326.

Deslandes, A., Moraes, H., Ferreira, C., Veiga, H., Silveira, H., Mouta, R., Pompeu, F.A.,Coutinho, E.S., Laks, J., 2009. Exercise and mental health: many reasons tomove. Neuropsychobiology 59, 191–198.

Dignon, A., Beardsmore, A., Spain, S., Kuan, A., 2006. 0Why I won0t eat’: patienttestimony from 15 anorexics concerning the causes of their disorder. Journal ofHealth Psychology 11, 942–956.

Engel, S.G., Wonderlich, S.A., Crosby, R.D., Wright, T.L., Mitchell, J.E., Crow, S.J.,Venegoni, E.E., 2005. A study of patients with anorexia nervosa using ecologicmomentary assessment. International Journal of Eating Disorders 38, 335–339.

Engelberg, M.J., Gauvin, L., Steiger, H., 2005. A naturalistic evaluation of the relationbetween dietary restraint, the urge to binge, and actual binge eating: aclarification. International Journal of Eating Disorders 38, 355–360.

Fairburn, C.G., Beglin, S.J., 1994. Assessment of eating disorders: interview or self-report questionnaire? International Journal of Eating Disorders 16, 363–370.

Fairburn, C.G., Cooper, Z., 1993. The eating disorder examination. In: Fairburn, C.G.,Wilson, G.T. (Eds.), Binge Eating: Nature, Assessment and Treatment, 12th ed.Guilford Press, New York, pp. 317–360.

Fairburn, C.G., Cooper, Z., Shafran, R., 2003. Cognitive behavior therapy for eatingdisorders: a “transdiagnostic” theory and treatment. Behavior Research andTherapy 41, 509–528.

Federici, A., Kaplan, A.S., 2008. The patient0s account of relapse and recovery inanorexia nervosa: a qualitative study. European Eating Disorders Review: TheJournal of the Eating Disorders Association 16, 1–10.

Fichter, M.M., Quaflieg, N., Hedlund, S., 2006. Twelve-year course and outcomepredictors of anorexia nervosa. International Journal of Eating Disorders 39, 87–100.

First, M.B., Spitzer, R.L., Gibbon, M., Williams, T.W., 1997. Structured ClinicalInterview for DSM-IV Axis I and II Disorders. American Psychiatric Press,Washington, DC.

Frank, G.K., Kaye, W.H., Weltzin, T.E., Perel, J., Moos, H., McConaha, C., Pollice, C.,2001. Altered response to meta-chlorophenylpiperazine in anorexia nervosa:support for a persistent alteration of serotonin activity after short-term weightrestoration. International Journal of Eating Disorders 30, 57–68.

Frijda, N.H., 1987. Emotion, cognitive structure, and action tendency. Cognition andEmotion 1, 115–143.

Garfinkel, P.E., Lin, E., Goering, P., Spegg, C., Goldbloom, D., Kennedy, S., Kaplan, A.S.,Woodside, D.B., 1996. Should amenorrhoea be necessary for the diagnosis ofanorexia nervosa? Evidence from a Canadian community sample. BritishJournal of Psychiatry: The Journal of Mental Science 168, 500–506.

Gross, J.J., 1998. The emerging field of emotion regulation: an integrative review.Review of General Psychology 2, 271–299.

Gross, J.J., 2002. Emotion regulation: Affective, cognitive, and social consequences.Psychophysiology 39, 281–291.

Harrison, A., Sullivan, S., Tchanturia, K., Treasure, J., 2009. Emotion recognition andregulation in anorexia nervosa. Clinical Psychology and Psychotherapy 16,348–356.

Hautzinger, M., Keller, F., Kuehner, C., 2007. Deutsche Adaptation des BeckDepressions-Inventars BDI-II. Harcourt Test Services, Frankfurt.

Haynos, A.F., Fruzzetti, A.E., 2011. Anorexia nervosa as a disorder of emotiondysregulation: evidence and treatment implications. Clinical Psychology:Sciene and Practice 18, 183–202.

Hebebrand, J., Exner, C., Hebebrand, K., Holtkamp, C., Casper, R.C., Remschmidt, H.,Herpertz-Dahlmann, B., Klingenspor, M., 2003. Hyperactivity in patients withanorexia nervosa and in semistarved rats: evidence for a pivotal role ofhypoleptinemia. Physiology and Behavior 79, 25–37.

Hilbert, A., Tuschen-Caffier, B., 2007. Maintenance of binge eating through negativemood: a naturalistic comparison of binge eating disorder and bulimia nervosa.International Journal of Eating Disorders 40, 521–530.

Hilbert, A., Tuschen-Caffier, B., Karwautz, A., Niederhofer, H, Munsch, S., 2007.Eating disorder examination-questionnaire. Evaluation der deutschsprachigenUebersetzung. Diagnostica 53, 144–154.

Hilbert, A., Tuschen-Caffier, B., Ohm, M., 2004. Eating disorder examination: EineDeutschsprachige version des strukturierten Essstoerungsinterviews. Diagnos-tica 50, 98–106.

Holtkamp, K., Herpertz-Dahlmann, B., Mika, C., Heer, M., Heussen, N., Fichter, M.,Herpertz, S., Senf, W., Blum,W.F., Schweiger, U., Warnke, A., Ballauff, A., Remschmidt,H., Hebebrand, J., 2003. Elevated physical activity and low leptin levels co-occur inpatients with anorexia nervosa. Journal of Clinical Endocrinology and Metabolism88, 5169–5174.

Hudson, J.I., Hiripi, E., Harrison, G.P., Kessler, R.C., 2007. The prevalence andcorrelates of eating disorders in the national comorbidity survey replication.Biological Psychiatry 61, 348–358.

Kaye, W., 2008. Neurobiology of anorexia and bulimia nervosa. Physiology andBehavior 94, 121–135.

E. Naumann et al. / Psychiatry Research 215 (2014) 711–717716

Kuehner, C., Buerger, C., Keller, F., Hautzinger, M., 2007. Reliabilitaet und Validitaetdes revidierten Beck-Depressionsinventars (BDI-II). Der Nervenarzt 78,651–656.

Le Grange, D., Swanson, S.A., Crow, S.J., Merikangas, K.R., 2012. Eating disorder nototherwise specified presentation in the US population. International Journal ofEating Disorders 45, 711–718.

Merwin, R., Zucker, N., Lacy, J., Elliott, C., 2010. Interoceptive awareness in eatingdisorders: distinguishing lack of clarity from nonacceptance of internal experi-ence. Cognition and Emotion 24, 892–902.

Mitchell, J.E., Crow, S., 2006. Medical complications of anorexia nervosa andbulimia nervosa. Current Opinion in Psychiatry 19, 438–443.

Morris, A., Do, D., Gottlieb-Smith, R., Ng, J., Jain, A., Wright, S., Shochet, R., 2012.Impact of a fitness intervention on medical students. Southern Medical Journal105, 630–634.

Nordbø, R.H.S., Espeset, E.M.S., Gulliksen, K.S., Skårderud, F., Holte, A., 2006. Themeaning of self-starvation: qualitative study of patients0 perception of anorexianervosa. International Journal of Eating Disorders 39, 556–564.

Norman, D.K., Herzog, D.B., 1983. Bulimia, anorexia nervosa, and anorexia nervosawith bulimia: a comparative analysis of MMPI profiles. International Journal ofEating Disorders 2, 43–52.

Peñas-Lledó, E., Vaz Leal, F.J., Waller, G., 2002. Excessive exercise in anorexianervosa and bulimia nervosa: relation to eating characteristics and generalpsychopathology. International Journal of Eating Disorders 31, 370–375.

Pirke, K.M., Broocks, A., Wilckens, T., Marquard, R., Schweiger, U., 1993. Starvation-induced hyperactivity in the rat: the role of endocrine and neurotransmitterchanges. Neuroscience and Biobehavioral Reviews 17, 287–294.

Plies, A., Florin, I., 1992. Effects of negative mood induction effects of negative moodinduction on the body image of restrained eaters. Psychology and Health 7,235–242.

Rossiter, E.M., Wilson, G.T., Goldstein, L., 1989. Bulimia nervosa and dietaryrestraint. Behavior Research and Therapy 27, 465–468.

Rottenberg, J., Ray, R.D., Gross, J.J., 2007. Emotion elicitation using films. In: Coan, J.A., Allen, J.J.B. (Eds.), The Handbook of Emotion Elicitation and Assessment.Oxford University Press, New York, pp. 9–24.

Rucken, D.D., Petty, R.E., 2004. Emotion specificity and consumer behavior: anger,sadness, and preference for activity. Motivation and Emotion 28, 3–21.

Schwartz, G.E., Weinberger, D.A., Singer, J.A., 1981. Cardiovascular differentiation ofhappiness, sadness, anger, and fear following imagery and exercise. Psychoso-matic Medicine 43, 343–364.

Serpell, L., Treasure, J., Teasdale, J., Sullivan, V., 1999. Anorexia nervosa: friend orfoe?. International Journal of Eating Disorders, 25; , pp. 177–186.

Smyth, J.M., Wonderlich, S.A., Heron, K.E., Sliwinski, M.J., Crosby, R.D., Mitchell, J.E.,Engel, S.G., 2007. Daily and momentary mood and stress are associated withbinge eating and vomiting in bulimia nervosa patients in the natural environ-ment. Journal of Consulting and Clinical Psychology 75, 629–638.

Solenberger, S.E., 2001. Exercise and eating disorders: a 3-year inpatient hospitalrecord analysis. Eating Behaviors 2, 151–168.

Steiger, H., Lehoux, P.M., Gauvin, L., 1999. Impulsivity, dietary control and the urgeto binge in bulimic syndromes. International Journal of Eating Disorders 26,261–274.

Stice, E., 2001. A prospective test of the dual-pathway model of bulimic pathology:mediating effects of dieting and negative affect. Journal of Abnormal Psychol-ogy 110, 124–135.

Strober, M., Freeman, R., Morrell, W., 1997. The long-term course of severe anorexianervosa in adolescents: survival analysis of recovery, relapse, and outcomepredictors over 10-15 years in a prospective study. International Journal ofEating Disorders 22, 339–360.

Svaldi, J., Griepenstroh, J., Tuschen-Caffier, B., Ehring, T., 2012. Emotion regulationdeficits in eating disorders: a marker of eating pathology or general psycho-pathology? Psychiatry Research 197, 103–111.

Tabachnick, B.G., Fidell, L.S., 2007. Using Multivariate Statistics, 5th ed. PearsonEducation Inc, Boston.

Tuschen-Caffier, B., Voegele, C., 1999. Psychological and physiological reactivity tostress: an experimental study on bulimic patients, restrained eaters andcontrols. Psychotherapy and Psychosomatics 68, 333–340.

Vansteelandt, K., Rijmen, F., Pieters, G., Probst, M., Vanderlinden, J., 2007. Drive forthinness, affect regulation and physical activity in eating disorders: a daily lifestudy. Behavior Research and Therapy 45, 1717–1734.

Waller, G., Babbs, M., Milligan, R., Meyer, C., Ohanian, V., Leung, N., 2003. Anger andcore beliefs in the eating disorders. International Journal of Eating Disorders 34,118–124.

Waters, A., Hill, A., Waller, G., 2001. Internal and external antecedents of bingeeating episodes in a group of women with bulimia nervosa. InternationalJournal of Eating Disorders 29, 17–22.

Wildes, J.E., Marcus, M.D., Bright, A.C., Dapelo, M.M., Psychol, M.C., 2012. Emotionand eating disorder symptoms in patients with anorexia nervosa: An experi-mental study. International Journal of Eating Disorders 45, 876–882.

Wildes, J.E., Ringham, R.M., Marcus, M.D., 2010. Emotion avoidance in patients withanorexia nervosa: Initial test of a functional model. International Journal ofEating Disorders 43, 398–404.

Wittchen, H., Zaudig, M., Fydrich, T., 1997. Strukturiertes Klinisches Interview fuerDSM-IV, Achse I und II. Hogrefe, Goettingen.

E. Naumann et al. / Psychiatry Research 215 (2014) 711–717 717