Prevalence of Eating Disorders amongst Dancers: A Systemic Review and Meta-Analysis

Abstract

Eating disorders in dancers are thought to be common, but the exact rates remain to be clarified. The aim of this study is to systematically compile and analyse the rates of eating disorders in dancers. A literature search, appraisal and meta-analysis were conducted. Thirty-three relevant studies were published between 1966 and 2013 with sufficient data for extraction. Primary data were extracted as raw numbers or confidence intervals. Risk ratios and 95% confidence intervals were calculated for controlled studies. The overall prevalence of eating disorders was 12.0% (16.4% for ballet dancers), 2.0% (4% for ballet dancers) for anorexia, 4.4% (2% for ballet dancers) for bulimia and 9.5% (14.9% for ballet dancers) for eating disorders not otherwise specified (EDNOS). The dancer group had higher mean scores on the EAT-26 and the Eating Disorder Inventory subscales. Dancers, in general, had a higher risk of suffering from eating disorders in general, anorexia nervosa and EDNOS, but no higher risk of suffering from bulimia nervosa. The study concluded that as dancers had a three times higher risk of suffering from eating disorders, particularly anorexia nervosa and EDNOS, specifically designed services for this population should be considered. Copyright © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

Full text

REVIEW
Prevalence of Eating Disorders amongst Dancers: A Systemic
Review and Meta-Analysis
Jon Arcelus
1,2
*, Gemma L. Witcomb
1,2
& Alex Mitchell
3,4
1
Loughborough University Centre for Research into Eating Disorders (LUCRED), Loughborough University, Loughborough, Leicestershire, UK
2
Eating Disorders Service, Leicester Partnership Trust, Bennion Centre, Leicester Glenfield Hospital, Leicester, UK
3
Department of Psycho-oncology, Leicester Partnership Trust, Leicester, UK
4
Honorary Senior Lecturer in Psycho-oncology, Department of Cancer and Molecular Medicine, Leicester Royal Infirmary, University of Leicester, Leicester, UK
Abstract
Eating disorders in dancers are thought to be common, but the exact rates remain to be clarified. The aim of this study is to systematically
compile and analyse the rates of eating disorders in dancers. A literature search, appraisal and meta-analysis were conducted. Thirty-three
relevant studies were published between 1966 and 2013 with sufficient data for extraction. Primary data were extracted as raw numbers
or confidence intervals. Risk ratios and 95% confidence intervals were calculated for controlled studies. The overall prevalence of eating
disorders was 12.0% (16.4% for ballet dancers), 2.0% (4% for ballet dancers) for anorexia, 4.4% (2% for ballet dancers) for bulimia and
9.5% (14.9% for ballet dancers) for eating disorders not otherwise specified (EDNOS). The dancer group had higher mean scores on the
EAT-26 and the Eating Disorder Inventory subscales. Dancers, in general, had a higher risk of suffering from eating disorders in general,
anorexia nervosa and EDNOS, but no higher risk of suffering from bulimia nervosa. The study concluded that as dancers had a three
times higher risk of suffering from eating disorders, particularly anorexia nervosa and EDNOS, specifically designed services for this
population should be considered. Copyright © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.
Keywords
eating disorders; anorexia; bulimia; EDNOS; review; meta-analysis; dance; ballet; epidemiology
*Correspondence
Professor Jon Arcelus, Eating Disorders Service, Leicester Partnership Trust, Bennion Centre, Leicester Glenfield Hospital, Leicester, UK. Tel: +44 (0)1509 223032.
Email: J.Arcelus@lboro.ac.uk
Published online 26 November 2013 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/erv.2271
Introduction
The lifetime prevalence for anorexia nervosa (AN) in women has
been estimated to be between 0.3 and 1 per cent, with a greater
number suffering from bulimia nervosa (BN) (Hoek & Van
Hoeken 2003; Preti et al., 2009). However, this prevalence appears
to be much higher for specific groups of society, such as models
(Smethurst, Wales, & Arcelus, 2010), athletes (Sundgot-Borgen
& Torstveit, 2004) and dancers (Herbrich, Pfeiffer, Lehmkuhl, &
Schneider, 2011; Ringham et al., 2006). Research investigating
the aetiology of eating disorders has identified a strong correlation
between these disorders and specific personality factors, such a
low self-esteem (Arcelus, Haslam, Farrow, & Meyer, 2013), per-
fectionism (Penniment & Egan, 2012; Ringham et al., 2006) and
high self-standards (Gunnard et al., 2012). The high levels of per-
fectionism (Nordin-Bates, Cumming, Sharp, & Aways, 2011;
Zoletić& Duraković-Belko, 2009) and low self-esteem (Benn &
Walters, 2001) found among dancers may explain why they ap-
pear to be more prevalent among this group of people. However,
the findings in this area are somewhat mixed and inconsistent.
Whereas some authors have found that up to 83% of dancers
report some form of lifetime eating pathology (Ringham et al.,
2006), others have not found a higher number of eating disorders
among dancers when compared to healthy controls (Toro,
Guerrero, Sentis, Castro, & Puertolas, 2009). The disparity of
the results may be explained by the differences in the populations
involved (e.g. students vs professionals), the different type of
dancers studied (e.g. ballet vs general) or the different tools used
to make a diagnosis.
In order to investigate the rates of eating disorders among
dancers, the primary aim of this study was to systematically com-
pile and analyse the current literature examining the rates of eat-
ing disorders in the dance population, taking into consideration
the great methodological variation, the type of dancers studied
and the change in diagnosis over the years.
Methods
Search strategy
Systematic literature search, appraisal and meta-analysis were
conducted. The following abstract databases were searched to July
2013: Medline/Pubmed, PsycINFO and Embase. Four full-text
collections were searched; Science Direct, Ingenta Select, Ovid
Full text and Wiley-Blackwell Interscience. A broad range of sub-
ject headings were used to identify the relevant disorders and
diagnoses. For each database search, combinations of the follow-
ing eight search components were used: eating disorders, eating
problems, AN, BN, eating disorders not otherwise specified
92 Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

(EDNOS), dance, dancers and ballet. Two investigators selected
the studies, extracted the data independently, cross-checked them
and resolved disagreements. Studies that included 10 or more par-
ticipants at the time of analysis were included. Those with less
than 10 were considered selected case-series not suitable for ascer-
tainment of prevalence. Reference lists of relevant articles were
screened for further potential studies and citation searches were
conducted. For all the included studies the following information
was recorded: author, year of publication, country, response rates,
number of individuals involved in the study, gender, type of
dance, setting (whether the population studied were students
(from elite or non-elite schools), amateurs or professionals), tools
used, diagnosed life time prevalence, point prevalence rate of eat-
ing disorders, AN, BN and EDNOS. Information from the results
of questionnaires used to measure eating psychopathology was
also recorded, which allowed for determination of caseness as
per the Eating Attitude Test (EAT; Garner, Olmsted, Bohr, &
Garfinkel, 1982). Mean scores for the EAT (EAT-40 and
EAT-26), Eating Disorders Inventory (EDI; Garner, Olmstead, &
Polivy, 1983), subscales of the EDI (drive for thinness [DT],
bulimia [B] and body dissatisfaction [BD]), and Eating Disorders
Questionnaire (EDE-Q; Fairburn & Beglin, 1994) were selected.
Information was also recorded from controls, plus a description
of the type of controls used (e.g. non-dancers, athletes). The
results are reported for the following: (1) all dancers; and (2) bal-
let dancers, specifically (as a big proportion of studies exclusively
selected this type of dancer). See Table 1 for further details.
Procedure and statistical analysis
Studies meeting the inclusion criteria were examined, and study
demographics (age, gender, type of dancers and setting) and in-
formation about eating disorders (prevalence data and mean
scores of questionnaires) were recorded. The articles were divided
into subgroups of controlled and uncontrolled studies. A study
was considered controlled if the prevalence of eating disorders
or the results of the questionnaires in dancers was compared to
non-dancers. Eating disorder was assessed using structured and
semi-structured diagnostic interviews and diagnosed according
to the Diagnostic and Statistical Manual of Mental Disorders
(DSM; APA, 1952, 1968, 1980, 1987, 2000) that was current at
the time of the study. In the majority of the studies, the severity
of the eating disorders was assessed by self-report symptoms
scales (e.g. EAT, EDI or EDE-Q). In some of these studies,
caseness was determined by scores above a specific threshold of
the EAT. Here, these individuals will be described as possibly suf-
fering from an eating disorder.
There were two outcomes of interest: (1) the number of
dancers diagnosed with an eating disorder at the time of study
(point prevalence); and (2) the mean results of the different
questionnaires describing eating disorder psychopathology (mean
severity). The primary data of the epidemiological studies describ-
ing eating disorders in dancers was extracted as raw numbers. For
the purpose of this analysis, study design was explored as a dichot-
omous variable. Risk ratios (RRs) were calculated only for
controlled studies, because these calculations derived from com-
paring the risk of eating disorders in dancers with the risk in the
non-dancers group. For those studies describing only mean re-
sults of their questionnaires, the effect size for the continuous var-
iables was calculated. An aggregate effect size weighted by sample
size was computed to provide an overall effect size across the stud-
ies. Homogeneity among studies was computed using the Q statis-
tic and the I
2
statistic. A significant Q statistic suggests that the
distribution of effect size around the mean is greater than would
be predicted from sampling error alone, whereas I
2
provides an
estimate of the proportion of the variance in the aggregate effect
size that is attributable to between-studies heterogeneity (Higgins
& Thompson, 2002). Random effects models were fitted if there
was heterogeneity. RRs and 95% confidence intervals (CIs) were
calculated. The relative weighted contribution of each study, as
well as the RR estimate pooled across all studies, were also
assessed. Meta-analysis was performed using the Comprehensive
Meta-Analysis software programme-2 for Windows (CMA;
Borenstein, Hedges, Higgins, & Rothstein, 2005) according to the
Cochrane reviewers’handbook (Higgins & Green, 2011). Signif-
icance was set up at p<0.05. As the outcomes of interest were
prevalence and severity, not the effect of certain factors such as
age or country where the study was conducted, meta-regression
analysis and grouping were not used.
Included and excluded studies
Two independent assessors (J. A. and G. L. W.) identified
91 relevant articles, which were screened in detail. Following
this, 50 papers were excluded. The reasons for the exclusions
were the following: non-English publication (with no English
abstract; n= 7); conference only paper (with paper not printed
Table 1 Criteria for searches on eating disorders and dance
Category Criteria
Study population Humans
All races, ethnicities and cultural groups
All ages
Study settings and geography All nations
Time period Published from 1966 through July 2013
Publication criteria Included the following:
-English
-Articles in print
Excluded the following:
-Articles in grey literature or non-peer-reviewed
journals or unobtainable during the review period
Admissible evidence
(study design and other
criteria)
-Anorexia nervosa must be diagnosed according
to DSM I, II, III, DSM-III-R, DSM-IV, ICD-10,
Feighner or Russell criteria.
-Bulimia nervosa must be diagnosed according to
DSM I, II, III, DSM-III-R, DSM-IV or ICD-10
criteria.
-Eating disorders not otherwise specified or
atypical eating disorders must be diagnosed
according to DSM-III and DSM-IV criteria.
-Eligible study designs include outcomes studies;
observational studies including prospective and
retrospective cohort studies.
-Populations must include 10 or more
participants (in total) at the time of analysis.
DSM, Diagnostic and Statistical Manual of Mental Disorders.
J. Arcelus et al. Eating Disorders and Dance
93Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

in a peer reviewed journal n= 8); inadequate sample size (less
than 10 patients; n= 3); lack of tools used to measure eating
psychopathology (n= 28), and not involving dancers but
involving aerobic teaches or skaters only (n=3). Out of the
42 papers retrieved for more detailed evaluation, eight were
excluded for having no description of eating disorders and/or
not using tools to measure eating disorder psychopathology
(not identifiable through the abstract). In total, 33 studies
were reviewed systematically.
Summary of relevant studies
Thirty-three relevant empirical quantitative studies in eating dis-
orders or eating problems in dancers were published between
1966 and 2013. The oldest study was from 1985 (Szmukler, Eisler,
Gillies, & Hayward, 1985) and the most recent one from 2012
(Nascimento & Fontenelle, 2012). Twenty-three (69.7%) papers
included exclusively ballet dancers, nine (27.3%) included general
dancers and one (3%) Turkish dancers (Batur, Kabakci, & Gulol,
2003). Only five papers (16.1%) included men. Those studies
were not used to calculate meta-analytical means. Thirteen
(39.4%) studies were European: three Italian (Dotti et al., 2002;
Ravaldi et al., 2003, 2006), two from the UK (Szmukler et al.,
1985), two from Germany (Bettle, Bettle, Neumaker, &
Neumaker, 1998; Herbrich et al., 2011) and one study from the
Netherlands (Fogelholm, Van Marken Lichtenbelt, Ottenheijm,
& Westerterp, 1996), France (Martin & Bellisle, 1989), Spain
(Toro et al., 2009), Switzerland (Burckhardt, Wynn, Kreig,
Bagutti, & Faouzi, 2011), Portugal (Francisco, Narciso, & Alarcao,
2012) and Belgium (Van Durme, Goossens, & Braet, 2012).
Twelve (36.4%) studies were from North America: one from Can-
ada (Piran, 1999) and 11 from the USA. Of the remaining papers,
three (9.1%) were Australian (Abraham, 1996; Anshel, 2004;
Penniment & Egan, 2012), two (6.1%) were Brazilian
(Nascimento & Fontenelle, 2012; Ribeiro & da Veiga, 2010), one
was Taiwanese (Tseng et al., 2007), one was Turkish (Batur
et al., 2003) and one was South African (Le Grange, Tibbs, &
Noakes, 1994). Twelve (36.4%) studies involved professional
dancers only: 10 (30.3%) ballet dancers and two (6.1%) general
dancers. Three (9.1%) studies included professional and amateur
dancers together (two with ballet dancers and one with Turkish
dancers); however, the authors of these studies present the results
separately for professionals and amateurs. The rest of the studies
(57.6%, n= 19) included students, 14 of them from ballet schools.
Among the studies from students, 12 (36.4%) provided eating dis-
orders information from students attending elite dance schools
(10 of them from ballet schools) and the rest (21.2%, n=7) from
non-elite schools (four from ballet schools). In view of the small
number of studies describing males and the low sample size and
power in this subgroup, only studies including females were
selected for the overall analysis. In view of the large number of
studies describing ballet dancers, the results of this population will
be shown separately.
Results
General descriptive statistics
The total number of dancers that participated in the studies was
3337. The mean age of participants was 19.5 years (SD = 5.76;
minimum and maximum mean age was 12.6 and 38 years, respec-
tively). The number of participants included in the studies varied
greatly from 13 to 665.
There were considerable differences among the studies in de-
sign, group size and methods. Nearly a quarter of the studies
(n= 8, 24.2%) offered prevalence data on clinical eating disorders
using a validated structured interview. Most of the studies used
eating disorders questionnaires (EAT), which does not prove a di-
agnosis but gives information on eating disorder psychopathol-
ogy. Twenty studies (60.6%) used a control group to compare
their results, either healthy controls (n= 12, 36.4%), a different
group of dancers (n= 6, 18.2%) or other athletes (gymnasts,
body-builders or gym users; n= 2, 6.1%).
Point prevalence of eating disorders
All dancers versus non-dancers
Seven studies used a diagnostic interview in order to identify
point prevalence of eating disorders in general. Eight studies used
it to identify point prevalence for anorexia, six for B and six for
EDNOS. Except for one study, which described the prevalence
of eating disorders in general dancers from Taiwan (Tseng et al.,
2007), all of the remaining studies included ballet dancers only.
The overall prevalence of eating disorders using random effects
meta-analysis was 12.0% (95% CI, 10.0–14.2). For anorexia it
was 2.0% (95% CI, 0.9–4.3), for B it was 4.4% (95% CI,
3.2–6.2) and for EDNOS it was 9.5% (95% CI, 7.6–11.8). Eleven
studies described the prevalence of possible cases of eating disor-
ders among dancers using caseness from the results of one ques-
tionnaire. The questionnaire used was the EAT (EAT; 40 or 26),
which described caseness when individuals score above 30 for
the EAT-40 and above 20 for the EAT-26. These questionnaires
have been very popular in this field, as more than half of the stud-
ies (n= 18, 58%) used them. Some studies used the EAT with the
aim of identifying possible cases of eating disorders as this usually
needs to be followed by an interview in order to reach a diagnosis.
Most of the studies did not aim to measure the prevalence of eat-
ing disorders but to identify a correlation between disordered eat-
ing and other variables, such as menstruation or bone density.
Seven of those studies included ballet dancers only. The preva-
lence of possible cases of eating disorders in professional dancers
ranged from 0 to 13.6% for males (only 2 studies) and from 7.4 to
50% in females.
Ballet dancers versus non-dancers
Six studies used diagnostic interviews in order to identify
current prevalence rates of eating disorders among ballet dancers.
The overall number of individuals involved in those studies
were 484 ballet dancers, with an overall mean age of 16.7 years
(SD = 1.13), ranging from 15.6 to 18.9 years. The maximum
number of individuals in one single study was 113 (Ravaldi
et al., 2003). All of the individuals were students. The prevalence
of eating disorders in ballet dancers ranged from 1.90 to
26.60%. The overall prevalence of eating disorders using random
effects meta-analysis was 16.4 (95% CI, 13.1–20.5). Lower
prevalence was identified in studies using non-specifically
designed eating disorders diagnostic interviews (general DSM
interviews). There were also a higher number of cases suffering
Eating Disorders and Dance J. Arcelus et al.
94 Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

from eating disorders among students not attending elite ballet
schools, although this was not statistically assessed. The preva-
lence for AN ranged from 0.9 to 4.1% and for BN from 0 to
2.7%. The overall prevalence of AN using random effect meta-
analysis was 4.0 (95% CI, 2.3–6.7) and for B 2.0 (95% CI,
0.9–4.3). The lower prevalence of BN may be explained by the
fact that this disorder tends to develop at a later age. As expected,
there was a high number of EDNOS cases (between 8.3 and
22.10%). The overall prevalence of EDNOS among ballet
dancers was 14.9 (95% CI, 11.3–19.4). Some of those cases
may have been diagnosed as AN or BN under the current
DSM-V diagnosis (APA, 2013).
Three studies, which included professional ballet dancers,
measured lifetime prevalence of eating disorders (Holderness,
Brooks-Gunn, & Warren, 1994; Nascimento & Fontenelle,
2012; Ringham et al., 2006). The lifetime prevalence found in
professional ballet dancers ranged from 15.78 to 82.6%. This
large variation can be explained by the number of EDNOS cases
found in one of the studies (72%; Ringham et al., 2006). The
lifetime prevalence for AN ranged from 3.4 to 31% and for BN
from 6.90 to 12%.
Prevalence of possible cases of eating disorders based on
questionnaire results
Seven studies described the prevalence of possible cases of eat-
ing disorders among ballet dancers using caseness. Four of the
seven studies describing caseness included professional ballet
dancers and three involved students from elite ballet schools.
The prevalence of possible cases of eating disorders in professional
dancers was 13.6% for males (one study) and ranged from 10.3 to
73.10% in females (Tables 2 and 3, and Figures 1–4).
Severity of eating psychopathology—mean results
of questionnaires
All dancers versus non-dancers
Seven studies used the EAT-40. The mean overall meta-analytical
score was 22.73 (SED =0.46). However, there was large variation in
the mean scores among the dancers in these studies ranging from
17.9 to 51.2. This variation could be explained by the differences
in the populations studied. There was a similar variation of means
of the EAT-26 (meta-analytical mean score = 3.54, SED = 0.07)
ranging from 2.37 to 22.90. However, this variation cannot be
explained by mean age, whether the dancers were students or
Table 2 Eating disorders in dancers
Year Author Country NM/F
Mean
Age
Type
Dance Setting Tools
Life
time
ED
(%)
AN
(%)
BN
(%)
EDNOS
(%)
EAT-26
1985 Szmukler et al. UK 100 F 15.6 Ballet Elite students Clinical interview N/a 7 7 N/a N/a
EAT-40
1994 Le Grange et al. S Africa 49 F 18.9 Ballet Elite students Clinical interview (DSM-III) N/a 12.3 4.1 0 N/a
EAT-26
1994 Holderness et al. USA 50 F 20.9 Ballet Professionals Clinical interview (DSM-III) 43 N/a 31 12 N/a
1996 Abraham Australia 60 F 16.9 Ballet Elite students Clinical interview (DSM-III) N/a 11.5 1.6 1.6 8.3
2003 Ravaldi et al. Italy 113 F 16.2 Ballet Non elite students EDE N/a 26.6 1.8 2.7 22.1
2006 Ringham et al. USA 29 F 19.7 Ballet Professionals EDI I& SCID interview 83 N/a 3.4 6.9 72.3
2006 Ravaldi et al. Italy 110 F 16.5 Ballet Non-elite students EDE N/a 13.6 0.9 1.8 10.9
EAT-26
2007 Tseng et al. Taiwan 613 F 15.8 General Elite school students SCID interview N/a 8 0.7 2.5 4.8
2009 Toro et al. Spain 76 F 14.4 General Non-elite students EAT-26 & CETCA N/a N/a 1.3* 15.8* N/a
EDI I
2011 Herbrich et al. Germany 52 F 16.4 Ballet Elite students SCID interview N/a 1.9 0 N/a 73.1
Interview (DSM-IV)
2012 Nascimento & Fontenelle Brazil 19 F 34.4 Ballet Professional BITE 15.7 N/a N/a N/a N/a
F, female; n/a, not available; SCID, Structured Clinical Interview for DSM Axis I Disorders; EDE, Eating Disorders Examination; CETCA, eating disorders assessment ques-
tionnaire; BITE, Bulimia investigation test Edinburgh.
*Probable anorexia or bulimia nervosa.
Table 3 Meta-analytical prevalence of eating disorders and meta-analytical
means of questionnaires
Ballet dancers n= 1729 All dancers n= 3337
Mean age 20.34 years 18.78 years
Eating disorders 16.4% 12.0%
Anorexia nervosa 4% 2.0%
Bulimia nervosa 2% 4.4%
EDNOS 14.9% 9.5%
Mean EAT-40 27.42 22.73
Mean EAT-26 3.03 3.54
Mean EDI 46.59 46.59
Drive for thinness (EDI) 15.78 6.21
Bulimia (EDI) 10.12 0.82
Body dissatisfaction (EDI) 17.24 13.06
EDE-Q 1.07 1.58
EDNOS, eating disorders not otherwise specified; EAT, Eating Attitude Test; EDI,
Eating disorders Inventory; EDEQ, Eating Disorders Examination Questionnaire.
J. Arcelus et al. Eating Disorders and Dance
95Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

Figure 1. Meta-analytical analysis of the mean prevalence of eating disorders in general in all dancers
Figure 2. Meta-analytical analysis of the mean prevalence of anorexia nervosa in all dancers
Figure 3. Meta-analytical analysis of the mean prevalence of bulimia nervosa all dancers
Figure 4. Meta-analytical analysis of the mean prevalence of eating disorders not otherwise specified for all dancers
Eating Disorders and Dance J. Arcelus et al.
96 Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

professionals, or by the country where the study took place. Seven
studies used the EDI to identify eating disorder psychopathology.
The overall mean EDI was 46.59 (SED= 0.60). The mean EDI is
rarely used in the eating disorders literature and scores on the
different eating disorders risk composite scales (DT, B and BD)
are more commonly reported. The overall mean for DT was 6.21
(SED = 0.21), for B was 0.82 (SED = 0.05) and for BD was 13.06
(SED = 0.25). Only three studies used the EDE-Q. The mean score
of the EDE-Q was 1.58 (SED = 0.07).
Ballet dancers versus non-dancers
Mean EAT-40 scores were available from three studies and
seven studies provided mean EAT-26 scores. Total EDI scores
and scores on the subscales of DT and B were available from seven
studies. Scores on the EDI-BD subscale were available from four
studies and EDE-Q mean scores were also available from four
studies. Some of the studies report mean scores in different
subgroups (i.e. primary and secondary students, or dancers
attending national, regional and local competitions; Piran
[1999], and Thomas, Keel, & Heatherton [2005], respectively),
which were analysed separately. The meta-analytical mean score of
the EAT-40 was 27.42 (SED = 12.5) and for the EAT-26 it was
3.03 (SED= 0.07). The mean score for the total EDI was 46.59
(SED = 0.60). The DT subscale mean was 15.78 (SED = 1.90), the
B subscale mean was 10.12 (SED = 2.47) and the BD subscale mean
was 17.24 (SED= 0.31). The EDE-Q meta-analytic mean score was
1.07 (SED = 0.58) (Table 4).
Relative risk of eating disorders in dancers in
general versus non-dancers
Meta-analysis was performed using only studies that provided
control data. Studies that provided control data from other ath-
letes such as gym users or gymnasts were not included as high
levels of eating disorders have already been reported in athletes.
Only five studies had enough information for the meta-analysis
to be carried out. They included a total of 948 dancers and 1619
non-dancers. Heterogeneity was low (I
2
= 0%) and the analysis
showed that dancers had a higher risk of suffering from eating dis-
orders in general (RR = 2.56, z= 5.83, p= 0.0001); in AN (RR =
3.60, z= 2.28, p= 0.02); and in EDNOS (RR = 3.02, z= 5.25,
p= 0.0001). The risk of developing B was not statistically signifi-
cantly higher in dancers when comparing to non-dancers (RR =
1.29, z= 1.21, p= 0.26).
Relative risk of eating disorders in ballet dancers
versus non-dancers
Only four studies provided enough control data to examine the
risk of suffering from an eating disorder in ballet dancers com-
pared to non-dancers. They included a total of 335 ballet dancers
and 438 non-dancers. Heterogeneity was low (I
2
= 0%). A statisti-
cally significantly higher risk for eating disorders in general (RR =
1.78, z= 2.70, p= 0.007) and for EDNOS (RR = 2.12, z= 2.90,
p= 0.004) was found among ballet dancers. No differences were
found for anorexia (RR = 3.85, z= 1.68, p= 0.09) or for B (RR =
0.62, z=0.9, p= 0.36).
There were only three studies using control data for the EAT-
26 and for the EDI. Meta-analysis was used to compare the mean
scores of the questionnaires between all dancers and non-dancers
in these studies. The scores in the dancer group were statistically
significantly higher for all the questionnaires: EAT-26 (z= 13.31,
p= 0.001), EDI-DT (z= 8.61, p= 0.001), EDI-B (z= 4.79) and
EDI-BD (z= 6.144, p= 0.001). There were not enough control
studies to use meta-analysis to compare questionnaires results of
just the ballet dancers (as opposed to all dancers) with non-dance
controls.
Discussion
The aim of the study was to systematically summarise the avail-
able literature in the field of eating disorders and dance. We were
able to identify 33 studies describing eating disorder diagnosis or
reporting mean values of questionnaires aimed at measuring eat-
ing disorder psychopathology. As per any meta-analysis, the study
is limited by the amount of studies available in the field. Although
the overall number of studies appears to be substantial (n= 33)
when examining the available data, there was a small number of
studies that were suitable for a proper meta-analysis. Meta-analy-
sis have been criticised by the fact that different populations and
variables have been compared (Eysenck, 1994). This study aimed
to compare a group of subjects as homogeneous as possible. As a
result, the study is strengthen by the homogeneity of the popula-
tion selected (female dancers or female ballet dancers) and the
variables collected (only studies that report eating disorders prev-
alence using strong methodology, such as semi-structured inter-
views). However, as a result, it has been limited by the small
number of studies available and the small populations studied.
Therefore, the results of the study need to be considered within
these limitations.
The overall prevalence of eating disorders in all dancers was
found to be 12%, which was slightly lower than the prevalence
found in ballet dancers (16.4%). The difference appeared to be
particularly related to the high number of EDNOS (14.9%) and
anorexia (4%) cases found among ballet dancers. A strength of
this analysis is the fact that this group is very homogeneous, all
of them are female and ballet students. In that respect, it appears
that nearly one fifth of ballet dance students suffer from an eating
disorder, predominantly from the EDNOS type. When comparing
the result of ballet students with non-dancers, we found that ballet
students had a significantly higher risk of suffering from eating
disorders, particularly EDNOS. In fact, they had twice the chance
of developing EDNOS when compared to non-dancers. Unfortu-
nately, there was not enough information about the type of
EDNOS. Research has identified that individuals suffering from
EDNOS of anorexia subtype present no significant differences in
eating related and general psychopathology when compared to
those suffering from AN (Le Grange et al., 2013). In fact, some
of the dancers suffering from EDNOS may have been diagnosed
with AN under the new DSM-V diagnostic criteria (APA, 2013).
Interestingly no significant difference was found in anorexia or
B rates. The former may be explained by the rarity of the disease
and the low number of studies in the field, whereas the latter may
be explained by the age of the population studied as this disorder
tends to appear at a later age.
The results were slightly different when comparing all dancers
with non-dancers. The population of dancers that has been in-
volved in these studies is too diverse to reach full conclusions as
J. Arcelus et al. Eating Disorders and Dance
97Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

Table 4 Mean scores of the eating disorders questionnaires in dancers
Year Author Country NM/F
Mean
age Dance Setting Tools
Caseness
for EAT EAT-40 EAT-26 EDI DT (EDI) B (EDI) BD (EDI) EDE-Q Control
1985 Szmukler et al. UK 100 F 15.6 Ballet Elite st EAT-26 16 18.9 N/a N/a N/a N/a N/a N/a No
1986 Schnitt, Schnitt, & Del A’Une USA 62 F 22.6 General Non elite
students
EAT-40 8 18 N/a N/a N/a N/a N/a N/a No
1986 Brooks-Gunn, Warren,
& Hamilton
USA 55 F 24.7 Ballet Prof. EAT-26 N/a N/a 3.57 N/a N/a N/a N/a N/a No
1987 Evers USA 21 F 21.4 General Non elite
students
EAT-40 33 21.8 N/a N/a N/a N/a N/a N/a Yes
1988 Hamilton, Brooks-Gunn,
Warren, & Hamilton (Chinese)
USA 17 F 24.6 Ballet Prof. EAT-26 N/a N/a 2.45 N/a N/a N/a N/a N/a No
1988 Hamilton et al. (American) USA 19 F 22.3 Ballet Prof. EAT-26 N/a N/a 2.37 N/a N/a N/a N/a N/a No
1989 Martin & Bellisle France 23 F 18.3 Ballet Elite st EAT-40 N/a 51.2 n/a N/a N/a N/a N/a N/a No
1994 Le Grange et al. S Africa 49 F 18.9 Ballet Elite st EAT-40 16 17.9 N/a N/a N/a N/a N/a N/a No
1994 Holderness et al. USA 50 F 20.9 Ballet Prof. EAT-26 16 N/a 13.55 N/a N/a N/a N/a N/a Yes
1996 Fogelholm et al. Netherlands 113 F 23.3 Ballet Prof. EDI I n/a N/a N/a 25.8 N/a N/a N/a N/a No
1998 Bettle et al. Germany 32 M 14.4 Ballet Elite st EAT-40 0 N/a N/a N/a N/a N/a N/a N/a No
1998 Bettle et al. Germany 58 F 14.4 Ballet Elite st EAT-40 14 N/a N/a N/a N/a N/a N/a N/a No
1999 Piran (primary school) Canada 45 F 13.3 Ballet Elite st EAT-26
and EDI
13 N/a 9.2 17.1 N/a N/a N/a N/a No
1999 Piran (secondary school) Canada 23 F 13.4 Ballet Elite st EAT-26 50 N/a 17.3 21.9 N/a N/a N/a N/a No
2002 Kaufman et al. USA 21 F 23.2 Ballet Prof. EAT-26 N/a N/a 22.9 N/a N/a N/a N/a N/a Yes
2002 Dotti et al. (children) Italy 75 F 12.6 General Non elite
students
EAT-26, EDI N/a N/a 5.13 N/a 1.8 0.16 3.56 N/a No
2002 Dotti et al. (adolescents) Italy 36 F 16.1 General Non elite
students
EAT-26, EDI N/a N/a 10.92 N/a 5.64 0.81 10.22 N/a No
2002 Dotti et al. (young adults) Italy 30 F 20.5 General Non elite
students
EAT-26, EDI N/a N/a 9.33 N/a 4.4 1.93 7.27 N/a No
2002 Dotti et al. (adults) Italy 19 F 24.5 General Non elite
students
EAT-26, EDI, N/a N/a 11.21 N/a 6.37 1.84 9.11 N/a No
2003 Ravaldi et al. Italy 113 F 16.2 Ballet Non elite
students
EDEQ N/a N/a N/a N/a N/a N/a N/a 1.6 No
2003 Batur et al. Turkey 12 M 23.0 Turkish
dance
Prof. EAT-40 N/a 18.2 N/a N/a N/a N/a N/a N/a Yes
2003 Batur et al. Turkey 12 M 23.0 Turkish
dance
Amat. EAT-40 N/a 12.8 N/a N/a N/a N/a N/a N/a Yes
2003 Batur et al. Turkey 32 F 22.1 Turkish
dance
Prof. EAT-40 N/a 34.9 N/a N/a N/a N/a N/a N/a Yes
2003 Batur et al. Turkey 22 F 22.1 Turkish
dance
Amat EAT-40 N/a 30.1 N/a N/a N/a N/a N/a N/a Yes
2004 Anshel Australia 57 F 17.3 Ballet Non elite
students
EDI-II N/a N/a N/a N/a 9.26 (3.75) 2.56 (1.51) 16.56 (2.57) N/a Yes
2005 Thomas et al. (National) USA 63 F 15.0 Ballet Elite st. EDI I N/a N/a N/a 81.9 18.8 13.7 N/a N/a No
2005 Thomas et al. (Regional) USA 64 F 15.0 Ballet Elite st. EDI I N/a N/a N/a 72.1 13.8 11.9 N/a N/a No
2005 Thomas et al. (Local) USA 107 F 15.0 Ballet Non elite
students
EDI I N/a N/a N/a 78.5 16.3 13.0 N/a N/a No
2006 Ringham et al. USA 29 F 19.7 Ballet Prof. EDI I N/a N/a N/a 43.4 11.27 2.32 12.14 N/a No
(Continues)
Eating Disorders and Dance J. Arcelus et al.
98 Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

they incorporate different types of dance (e.g. ballet, jazz, na-
tional) and spanning levels from students to professionals. There-
fore, these results need to be considered with caution and
generalisability may be difficult. Nevertheless, our results showed
that general dancers had more than twice the risk of developing
an eating disorder and more than three times the risk of develop-
ing AN and EDNOS than non-dancers. In spite of the limited
number of questionnaire studies that used control data when
assessing eating disorder psychopathology, our analysis found a
significant difference on the EAT (EAT-26) and the D, B and
BD subscales of the EDI. BD and general body shame has been
found to predict an increase in anorexic symptoms but interest-
ingly not bulimic symptoms (Troop & Redshaw, 2012).
Although this study did not compare ballet dancers with all
dancers (which would include some dancers who do ballet as well
as other dance) because of the limited number of control studies
comparing ballet with general dancers, it is interesting that there
appears to be a difference in the subscales of the EDI between
the two groups, indicating that ballet dancers present with higher
levels of DT, B and BD. Further research in this area will be able to
explore the reason for this. For example, do ballet students have a
preexisting elevated risk of developing an eating disorder before
they become involved in ballet, or is it a result of the environ-
ment? Research into physical injuries in dancers has found that
both professional and amateur dancers have a tendency to
normalise pain (Markula, 2013). The belief that if no pain is felt,
then the dancers are not pushing themselves hard enough appears
to be commonplace (Turner & Wainright, 2003). A similar drive
to push oneself to achieve the desired ‘dancer’s body’, irrespective
of how hard this may be, may possibly play a role in the higher
levels of eating disorders seen in ballet dancers. Recent studies
have also found severity of appearance-based critical comments,
and lower self-esteem was associated with heightened eating
psychopathology (Goodwin, Arcelus, Marshall, Wick, & Meyer,
in press).
Limitations and future studies
This analysis has a number of limitations. Notably, we were reli-
ant on the quality and non-repetition of primary publications.
We sought to ensure data integrity by excluding small studies of
less than 10 patients. As a result, the resulting study is limited
by the number of available published data for this population,
and although the data were reasonably robust for ballet dancers,
the population of other dancers was non-homogenous and mixed.
Throughout the years, investigators have used a large number of
different diagnostic and outcome assessment measures, which
have reflected numerous definitions of diagnosis. The lack of con-
sistency in the measures used makes comparisons across studies
very difficult. Consolidation of measures, standardised definitions
and reporting guidelines are critical to the further advancement of
the field. The weakness of most of the studies used in this meta-
analysis is that they aim to measure eating disorder psychopathol-
ogy by studying its relationship with other factors, such as
menstruation or injuries. The number of epidemiological studies
in the field of dance are few, possibly because of the complexity
of undertaking such studies and the secrecy of the disorder.
However, dance organisations such as Dance UK have started to
Year Author Country NM/F
Mean
age Dance Setting Tools
Caseness
for EAT EAT-40 EAT-26 EDI DT (EDI) B (EDI) BD (EDI) EDE-Q Control
2007 Tseng et al. Taiwan 613 F 15.8 General Elite st EAT-26 N/a N/a 17.2 N/a N/a N/a N/a N/a Yes
2009 Toro et al. Spain 76 F 14.4 General
dance
Non elite
students
EAT-26 11.8 N/a N/a N/a N/a N/a N/a N/a No
2010 Ribeiro & Veiga Brazil 29 M 38.0 Ballet Prof. EAT-26 + BITE 14 N/a N/a N/a N/a N/a N/a N/a No
2010 Ribeiro & Veiga Brazil 50 F 36.8 Ballet Prof. EAT-26 + BITE 10 N/a N/a N/a N/a N/a N/a N/a No
2011 Herbrich et al. Germany 52 F 16.4 Ballet Elite
students
EDI I N/a N/a N/a N/a 20.58 14.06 29.63 N/a Yes
2011 Nordin-Bates, Cumming, et al. UK 85 M General Elite st EAT-26 8 N/a N/a N/a N/a N/a N/a N/a No
2011 Nordin-Bates, Walker,
& Redding
UK 261 F 14.4 General Elite st EAT-26 7 N/a N/a N/a N/a N/a N/a N/a No
2011 Burckhardt et al. Switz. 127 F 16.7 General Prof. EAT-40 N/a 12.6 N/a N/a N/a N/a N/a N/a No
2012 Penniment & Egan Australia 142 F 22.3 General Non-
prof
EDEQ N/a N/a N/a N/a N/a N/a N/a 1.85 No
2012 Francisco et al. Portugal 13 M 15.0 Ballet Elite st. EDE-Q N/a N/a N/a N/a N/a N/a N/a 0.46 Yes
2012 Francisco et al. Portugal 53 F 14.6 Ballet Elite st. EDE-Q N/a N/a N/a N/a N/a N/a N/a 1.54 Yes
2012 Francisco et al. Portugal 47 F 14.6 Ballet Non elite
students
EDE-Q N/a N/a N/a N/a N/a N/a N/a 0.69 Yes
2012 Van Durme et al. Belgium 44 F 14.5 Ballet Elite st. EDI-II N/a N/a N/a N/a 20.8 13.46 29.54 N/a Yes
F, female; M, males; EAT, Eating Attitude Test; EDI, Eating Disorders Inventory; EDEQ, Eating Disorders Examination Questionnaire; n/a, not available; Elite St, elite students; Prof., professionals; Amat, Amateurs;
Switz, Switzerland.
Table 4 (Continued)
J. Arcelus et al. Eating Disorders and Dance
99Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

highlight the need for such studies in order to identify the need
for clinical services linked to dancers. Dancers suffering from
physical injuries are treated within specialist clinical services. It
can be argued then that as eating disorders are found to be consid-
erably more prevalent in this population, dancers should also have
access to dance-specific eating disorders services. Advances in
technology provide opportunities for the ways in which self-help
can be delivered (e.g. Cardi, Clarke, & Treasure, 2013), which
could be tailored to dancers. Alternatively, specifically designed
preventative programmes, as described in other populations,
(Beintner, Jacobi, & Taylor, 2012) may reduce levels of eating dis-
orders in dancers. A well-performed epidemiological study in the
dance world using diagnostic interviews and controlling for age
and type of dance may be able to address the question of whether
and how the presentation of eating disorders in dancers differs
from non-dancers.
REFERENCES
Abraham, S. (1996). Characteristics of eating disorders among young
ballet dancers. Psychopathology,29, 223–229.
American Psychiatric Association. (1952). Diagnostic and Statistical
Manual of Mental Disorders (1st edn). Washington, DC: America
Psychiatric Association.
American Psychiatric Association. (1968). Diagnostic and Statistical
Manual of Mental Disorders (2nd edn). Washington, DC:
America Psychiatric Association.
American Psychiatric Association. (1980). Diagnostic and Statistical
Manual of Mental Disorders (3rd rev edn). Washington, DC:
America Psychiatric Association.
American Psychiatric Association. (1987). Diagnostic and Statistical
Manual of Mental Disorders (4th edn). Washington, DC: America
Psychiatric Association.
American Psychiatric Association. (2000). Diagnostic and Statistical
Manual of Mental Disorders (4th rev edn). Washington, DC:
America Psychiatric Association.
American Psychiatric Association. (2013). Diagnostic and Statistical
Manual of Mental Disorders (5th). Washington, DC: America
Psychiatric Association.
Anshel, M. H. (2004). Sources of disordered eating patterns between
ballet dancers and non-dancers. Journal of Sport Behaviour,27,
115–133.
Arcelus, J., Haslam, M., Farrow, C., & Meyer, C. (2013). The role of
interpersonal functioning in the maintenance of eating psycho-
pathology: A systematic review and testable model. Clinical
Psychology Review,33, 156–167.
Batur, S., Kabakci, E., & Gulol, C. (2003). Maladaptive eating
attitudes of elite and amateur Turkish dancers: are they at risk?
Eating and Weight Disorders,8, 263–267.
Beintner, I., Jacobi, C., & Taylor, C. B. (2012). Effects of an internet-
based prevention programme for eating disorders in the USA and
Germany: a meta-analytic review. European Eating Disorders
Review,20,1–9.
Benn, T., & Walters, D. (2001). Between Scylla and charybdis
nutritional education versus body culture and the ballet aesthetic:
the effects on the lives of female dancers. Research in Dance
Education,2, 139–155.
Bettle, N., Bettle, O., Neumaker, U., & Neumaker, K. J. (1998).
Adolescent ballet school students: their quest for body weight
change. Psychopathology,31, 153–159.
Borenstein, M., Hedges, L., Higgins, J., & Rothstein, H. (2005).
Comprehensive Meta-Analysis. Version 2. Englewood NJ: Biostat.
Brooks-Gunn, J., Warren, P., & Hamilton, L. H. (1986). The relation
of eating problems and amenorrhea in ballet dancers. Medicine
and Science in Sports and Exercise,19,41–44.
Burckhardt, P., Wynn, E., Kreig, M.-A., Bagutti, C., & Faouzi, M.
(2011). The effects of nutrition, puberty and dancing on bone
density in adolescent ballet dancers. Journal of Dance Medicine
and Science,15,51–60.
Cardi, V., Clarke, A., & Treasure, J. (2013). The use of guided self-
help incorporating a mobile component in people with eating
disorders: a pilot study. European Eating Disorders Review,21,
315–322.
Dotti, A., Fioravanti, A., Balotta, M., Tozzi, F., Cannella, C., &
Lazzari, R. (2002). Eating behavior of ballet dancers. Eating and
Weight Disorders,7,60–67.
Evers, C. L. (1987). Dietary intake and symptoms of anorexia
nervosa in female university dancers. Perspectives in Practice,87,
66–68.
Eysenck, H. J. (1994). Systematic reviews: meta-analysis and its
problems. British Medical Journal,309, 789–790.
Fairburn, C. G., & Beglin, S. J. (1994). Assessment of eating disorder
psychopathology: interview or self-report questionnaire? Interna-
tional Journal of Eating Disorders,16, 363–370.
Fogelholm, M., Van Marken Lichtenbelt, W., Ottenheijm, R., &
Westerterp, K. (1996). Amenorrhea in ballet dancers in the
Netherlands. Medicine and Science in Sports and Exercise,28,
545–550.
Francisco, R., Narciso, I., & Alarcao, M. (2012). Specific predictors of
disordered eating among elite and non-elite gymnasts and ballet
dancers. International Journal of Sport Psychology,43, 479–502.
Garner, D. M., Olmstead, M. P., & Polivy, J. (1983). Development
and validation of a multidimensional eating disorder inventory
for anorexia nervosa and bulimia. International Journal of Eating
Disorders,2,15–34.
Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982).
The eating attitude test: psychometric features and clinical corre-
lates. Psychological Medicine,12, 871–878.
Goodwin, H., Arcelus, J., Marshall, S., Wick, S., & Meyer, C. (in
press). Critical comments concerning shape and weight: associa-
tions with eating psychopathology among full-time dance
students. Eating and Weight Disorders - Studies on Anorexia,
Bulimia and Obesity (pubmed ID: 24097346).
Gunnard, K., Krug, I., Jimenez-Murcia, S., Penelo, E., Granero, E.,
Treasure, J., et al. (2012). Relevance of social and self-standards in
eating disorders. European Eating Disorders Review,20,271–279.
Hamilton, L. H., Brooks-Gunn, J., Warren, M. P., & Hamilton,
W. G. (1988). The role of selectivity in the pathogenesis of eating
problems in ballet dancers. Medicine and Science in Sports and
Exercise,20, 560–565.
Herbrich, L., Pfeiffer, E., Lehmkuhl, U., & Schneider, N. (2011).
Anorexia athletica in pre-professional ballet dancers. Journal of
Sports Sciences,29, 1115–1123.
Higgins, J. P. T., & Green, S. (2011). Cochrane handbook for sys-
tematic reviews of interventions. The Cochrane collaboration.
www.cochrane-handbook.org (accessed March 2011)
Higgins, J. P. T., & Thompson, S. G. (2002). Quantifying heterogene-
ity in a meta-analysis. Statistics in Medicine,21, 1539–15558.
Hoek, H. W., & Van Hoeken, D. (2003). Review of the prevalence
and incidence of eating disorders. International Journal of Eating
Disorders,34, 383–96.
Holderness, C. C., Brooks-Gunn, J., & Warren, M. P. (1994).
Eating disorders and substance use: a dancing vs a nondancing
population . Medicine and Science in Sports and Exercise,26,
297–302.
Kaufman, B. A., Warren, M. P., Domingues, J. E., Wang, J.,
Heymsfield, S. T., & Pierson, R. N. (2002). The Journal of Clinical
Endocrinology and Metabolism,87, 2777–2783.
Le Grange, D., Crosby, R. D., Engel, S. G., Cao, L., Ndungu, A.,
Crow, S. J., et al. (2013). DSM-IV-Defined anorexia nervosa
versus sub threshold anorexia nervosa (EDNOS-AN). European
Eating Disorders Review,21,1–7.
Le Grange, D., Tibbs, J., & Noakes, T. D. (1994). Implications of a
diagnosis of anorexia nervosa in a ballet school. International
Journal of Eating Disorders,15, 369–376.
Markula, P. (2013). (Im)Mobile bodies: contemporary semi-
professional dancers’experiences with injuries. International
Review for the Sociology of Sport. Published online before print.
DOI: 10.1177/1012690213495745
Martin, C., & Bellisle, F. (1989). Eating attitudes and taste responses
in young ballerinas. Physiology and Behavior,46, 223–227.
Nascimento, A. L., & Fontenelle, L. F. (2012). Body dysmorphic
disorder and eating disorders in elite professional female ballet
dancers. Annals of Clinical Psychiatry,24, 191–194.
Nordin-Bates, S. M., Cumming, J., Sharp, L., & Aways, D. (2011).
Imagining yourself dancing to perfection? Cognitive correlates
of perfectionism among ballet and contemporary dancers.
Journal of Clinical Sport Psychology,5,58–76.
Nordin-Bates, S. M., Walker, I. J., & Redding, E. (2011). Correlates
of disordered eating attitudes among male and female young
talented dancers: findings from the UK centres for advanced
training. Eating Disorders,19, 211–233.
Penniment, K. I., & Egan, S. J. (2012). Perfectionism and
learning experiences in dance class as risk factors for eating
disorders in dancers. European Eating Disorders Review,20,
13–23.
Piran, N. (1999). Eating disorders: a trial of prevention in a high risk
school setting. The Journal of Primary Prevention,20,75–90.
Preti, A., De Girolamo, G., Vilagut, G., Alonso, J., Graaf, R. D.,
Bruffaerts, R., et al. (2009). The epidemiology of eating disorders
in six European countries: results of the ESEMeD-WMH project.
Journal of Psychiatric Research,43, 1125–1132.
Ravaldi, C., Vannacci, A., Bolognesi, E., Mancini, S., Faravelli, C., &
Ricca, V. (2006). Gender role, eating disorder symptoms, and
body image concern in ballet dancers. Journal of Psychosomatic
Research,61, 529–535.
Ravaldi, C., Vannacci, A., Zucchi, T., Mannucci, E., Cabras, P. L.,
Boldrin, M., et al. (2003) Eating disorders and body image
disturbances among ballet dancers, gymnasium users and body
builders. Psychopathology,36, 247–254.
Ribeiro, L. G., & da Veiga, G. V. (2010). Risk behaviors for eating
disorders in Brazilian dancers. International Journal of Sports
Medicine,31, 283–288.
Eating Disorders and Dance J. Arcelus et al.
100 Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.

Ringham,R.,Klump,K.,Kaye,W.,Stone,D.,Libman,S.,Stowe,
S., et al. (2006). Eating disorder symptomatology among
ballet dancers. International Journal of Eating Disorders,39,
503–508.
Schnitt, J. M., Schnitt, D., & Del A’Une, W. (1986). Anorexia nervosa
or thinness in modern dance students: comparison with balle-
rinas. Annuals of Sports Medicine,31,9–13.
Smethurst, W., Wales, J., & Arcelus, J. (2010). Puff the magic slim-
mer. European Eating Disorders Review,18, 431–433.
Sundgot-Borgen, J., & Torstveit, M. K. (2004). Prevalence of
eating disorders in elite athletes is higher than in the
general population. Clinical Journal of Sport Medicine,14,
25–32.
Szmukler, G. I., Eisler, I., Gillies, C., & Hayward, M. E. (1985). The
implication of anorexia nervosa on a ballet school. Journal of
Psychiatric Research,17, 177–181.
Thomas, J. J., Keel, P. K., & Heatherton, T. F. (2005). Disordered
eating attitudes and behaviors in ballet students: examination of
environmental and individual risk factors. International Journal
of Eating Disorders,38, 263–268.
Toro, J., Guerrero, M., Sentis, J., Castro, J., & Puertolas, C. (2009).
Eating disorders in ballet dancing students: problems and risk
factors. European Eating Disorders Review,17,40–49.
Troop, N. A., & Redshaw, C. (2012). General shame and bodily
shame in eating disorders: a 2.5 years longitudinal study.
European Eating Disorders Review,20, 373–378.
Tseng, M. M.-C., Fang, D., Lee, M.-B., Chie, W.-C., Liu, J. P., &
Chen, W. J. (2007). Two-phase survey of eating disorders in
gifted dance and non-dance high-school students in Taiwan.
Psychological Medicine,37, 1085–1096.
Turner, B.S., & Wainright, S. P. (2003). Corps de ballet: the case of the
injured ballet dancer. Sociology of Health & Illness,25,269–288.
Van Durme, K., Goossens, L., & Braet, C. (2012). Adolescent
aesthetic athletes: a group at risk for eating pathology? Eating
Behaviors,13, 119–122.
Zoletić, E., & Duraković-Belko, E. (2009). Body image distortion,
perfectionism and eating disorder symptoms in risk group of
female ballet dancers and models and in control group of female
students. Psychiatria Danubina,21, 302–309.
J. Arcelus et al. Eating Disorders and Dance
101Eur. Eat. Disorders Rev. 22 (2014) 92–101 © 2013 John Wiley & Sons, Ltd and Eating Disorders Association.