Hostname: page-component-745bb68f8f-grxwn Total loading time: 0 Render date: 2025-02-10T10:30:10.428Z Has data issue: false hasContentIssue false
  • English
  • Français

Childhood maltreatment and eating disorder pathology: a systematic review and dose-response meta-analysis

Published online by Cambridge University Press:  19 January 2017

M. L. Molendijk ,
H. W. Hoek ,
T. D. Brewerton  and
B. M. Elzinga
M. L. Molendijk*
Affiliation:
Institute of Psychology, Clinical Psychology Unit, Leiden University, Leiden, The Netherlands Leiden Institute for Brain and Cognition, Leiden University Medical Center, Leiden, The Netherlands
H. W. Hoek
Affiliation:
Parnassia Psychiatric Institute, The Hague, The Netherlands Department of Psychiatry, University Medical Center Groningen, Groningen, The Netherlands Department of Epidemiology, Columbia University, New York, USA
T. D. Brewerton
Affiliation:
Department of Psychiatry and Behavioral Sciences, Medical University of South Carolina, Charleston, South Carolina, USA
B. M. Elzinga
Affiliation:
Institute of Psychology, Clinical Psychology Unit, Leiden University, Leiden, The Netherlands Leiden Institute for Brain and Cognition, Leiden University Medical Center, Leiden, The Netherlands
*
*Address for correspondence: M. L. Molendijk, Ph.D., Institute of Psychology, Clinical Psychology Unit, Leiden University, Leiden, The Netherlands and Leiden Institute for Brain and Cognition, Leiden University Medical Center, Leiden, The Netherlands. (Email: m.l.molendijk@fsw.leidenuniv.nl)
Rights & Permissions [Opens in a new window]

Abstract

Background

Meta-analyses have established a high prevalence of childhood maltreatment (CM) in patients with eating disorders (EDs) relative to the general population. Whether the prevalence of CM in EDs is also high relative to that in other mental disorders has not yet been established through meta-analyses nor to what extent CM affects defining features of EDs, such as number of binge/purge episodes or age at onset. Our aim is to provide meta-analyses on the associations between exposure to CM (i.e. emotional, physical and sexual abuse) on the occurrence of all types of EDs and its defining features.

Method

Systematic review and meta-analyses. Databases were searched until 4 June 2016.

Results

CM prevalence was high in each type of ED (total N = 13 059, prevalence rates 21–59%) relative to healthy (N = 15 092, prevalence rates 1–35%) and psychiatric (N = 7736, prevalence rates 5–46%) control groups. ED patients reporting CM were more likely to be diagnosed with a co-morbid psychiatric disorder [odds ratios (ORs) range 1.41–2.46, p < 0.05] and to be suicidal (OR 2.07, p < 0.001) relative to ED subjects who were not exposed to CM. ED subjects exposed to CM also reported an earlier age at ED onset [effect size (Hedges’ g) = −0.32, p < 0.05], to suffer a more severe form of the illness (g = 0.29, p < 0.05), and to binge-purge (g = 0.31, p < 0.001) more often compared to ED patients who did not report any CM.

Conclusion

CM, regardless of type, is associated with the presence of all types of ED and with severity parameters that characterize these illnesses in a dose dependent manner.

Keywords

Anorexiabulimiachildhood abuseeating disordersmeta-analysis
Type
Original Articles
Information
Psychological Medicine , Volume 47 , Issue 8 , June 2017 , pp. 1402 - 1416
Copyright
Copyright © Cambridge University Press 2017 

Introduction

The Diagnostic and Statistical Manual of Mental Disorders 4th edition (DSM-IV-TR; APA, 2000) specified only two types of eating disorders (EDs): anorexia nervosa (AN) and bulimia nervosa (BN), and a residual category ED not otherwise specified (EDNOS), which included binge-eating disorder (BED) for which in DSM-IV a criteria set has been proposed for further study. In the DSM-5 (APA, 2013), BED has been added as an official new diagnostic category and besides for BED the EDNOS category also made way for two new residual categories: (1) other specified feeding or ED, including for instance sub-threshold BN and (2) unspecified feeding or ED. The lifetime prevalence of eating disorders according to DSM-5 criteria is estimated to be up to 4% for AN, 2% for BN, and 2% for BED (Smink et al. Reference Smink, Van Hoeken and Hoek2013).

EDs are associated with co-morbidity of other mental disorders, high mortality rates and high costs of treatment (Fairburn & Harrison, Reference Fairburn and Harrison2003; Arcelus et al. Reference Arcelus, Mitchell, Wales and Nielsen2011; Mischoulon et al. Reference Mischoulon, Eddy, Keshaviah, Dinescu, Ross, Kass, Franko and Herzog2011; Mitchell et al. Reference Mitchell, Mazzeo, Schlesinger, Brewerton and Smith2012; Smink et al. Reference Smink, Van Hoeken and Hoek2013; Yao et al. Reference Yao, Kuja-Halkola, Thorton, Runfola, D'Onofrio, Almqvist, Lichtenstein, Sjölander, Larsson and Bulik2016). Their etiology is complex. Heritability estimates for EDs vary between 40% and 60% (Trace et al. Reference Trace, Baker, Peñas-Lledó and Bulik2013). Non-shared environmental risk factors such as childhood maltreatment (CM) also contribute to the risk of developing an ED, according to some, an estimated 17–46% (Klump et al. Reference Klump, Wonderlich, Lehoux, Lilenfeld and Bulik2002).

Dozens of studies have been published on the association between CM, usually defined as self-reported exposure to emotional abuse or neglect and physical and/or sexual abuse before the age of 18 years, and ED pathology. Among these studies are four meta-analyses (Rind et al. Reference Rind, Tromovitch and Bauserman1998; Smolak & Murnen, Reference Smolak and Murnen2002; Chen et al. Reference Chen, Murad, Paras, Colbenson, Sattler, Goranson, Elamin, Seime, Shinozaki, Prokop and Zirakzadeh2010; Caslini et al. Reference Caslini, Bartoli, Crocamo, Dakanalis, Clerici and Carrà2016) and some narrative reviews (Kent & Waller, Reference Kent and Waller2000; Brewerton Reference Brewerton2007; Röhr et al. Reference Röhr, Dölemeyer, Klinitzke, Steinig, Wagner and Kersting2015). The data pooled in the meta-analyses, the highest level of evidence available (Haidich, Reference Haidich2010), clearly shows a relationship between CM and the presence of an ED diagnosis, although heterogeneity in the magnitude of effect-size estimates is apparent (Smolak & Murnen, Reference Smolak and Murnen2002).

However, the meta-analytical results published so far are rather specific, as they pertain to college samples only (Chen et al. Reference Chen, Murad, Paras, Colbenson, Sattler, Goranson, Elamin, Seime, Shinozaki, Prokop and Zirakzadeh2010) or sexual abuse only (Smolak & Murnen, Reference Smolak and Murnen2002; Chen et al. Reference Chen, Murad, Paras, Colbenson, Sattler, Goranson, Elamin, Seime, Shinozaki, Prokop and Zirakzadeh2010). The most recent meta-analysis summarized the associations between all CM types and all ED types up to January 2014, but did not include a psychiatric control group nor assessed the effect of CM on psychiatric co-morbidity or other severity parameters (Caslini et al. Reference Caslini, Bartoli, Crocamo, Dakanalis, Clerici and Carrà2016). What is lacking is an in-depth overview that summarizes all CM types and all ED types and compares the prevalence not only with a healthy control group, but also with a control group with other mental disorders than EDs. Moreover, from a clinical perspective it is highly relevant to elucidate to what extent CM affects the defining features of EDs, such as its severity or age at onset since these define disease to a large extent (Fairburn & Harrison, Reference Fairburn and Harrison2003; Smink et al. Reference Smink, Van Hoeken and Hoek2013), beyond only the presence of a diagnostic entity. To date, this has not been subjected to systematic review.

The aim of this study hence is to provide a systematic overview on the associations between exposure to CM (i.e. emotional, physical and sexual abuse) on the occurrence of all types of EDs and some of its defining features such as severity and age at onset.

Method

The methodology that we used adhered to the guidelines that are recommended by the preferred reporting items for systematic reviews and meta-analyses statement (Moher et al. Reference Moher, Liberati, Tetzlaff and Altman2009).

Search strategy

The electronic databases PUBMED, PsycINFO, and EMBASE were searched, up to 4 June 2016, to identify relevant articles. The broad set of search terms can be found in section A of the Supplementary material. Additionally, we checked the references that were made to the two seminal papers on the association of interest (Oppenheimer et al. Reference Oppenheimer, Howells, Palmer and Chaloner1985; Root & Fallon, Reference Root and Fallon1988) and reviewed the reference lists of the identified articles. All searches, from identification to decision on inclusion, were performed by at least two independent reviewers (M.L.M., M.vS., M.O., V.K.). Inconsistencies between reviewers were resolved through consensus.

Inclusion and exclusion criteri

This review included human studies (all age groups) that assessed the presence of a current or lifetime ED as defined by the DSM (APA, 1980, 2000, 2013) or the International Classification of Diseases (ICD; WHO, 2016a ) and CM (before the age of 18 years). We largely followed the definition and categorization of the WHO in what we considered as exposure to CM. The WHO (2016b ) distinguishes five subtypes of CM: physical CM, sexual CM, neglect, emotional CM and exploitation. However, given that neglect and emotional abuse are often reported as one entity we decided to pool these categories as ‘emotional CM’. Furthermore, since exploitation is hardly a topic in the ED literature, we decided to not include this category in our meta-analysis. A general exclusion criterion was the use of non-diagnostic, quantitative assessment of eating behavior (e.g. disordered eating assessed through a questionnaire). Studies had to be published in English, French, German, Spanish or Dutch in order to be included.

Data extraction

Data that were extracted included: demographic and methodological characteristics (e.g. mean age of the samples that were used and study design), CM prevalence by diagnostic group, i.e. AN, BN, BED, EDNOS, healthy controls (HCs) and psychiatric controls (PCs) and type of CM (i.e. reported emotional, physical, and sexual abuse), CM characteristics (i.e. type, mean age, assessment method), and ED characteristics (i.e. type, mean age at onset, the number of self-reported binge and purge episodes per week/month, the severity of the ED, assessed for instance by means of the eating attitudes test (Mintz & O'Halloran, Reference Mintz and O'halloran2000), the eating disorder examination (Luce et al. Reference Luce, Crowther and Pole2008), or the eating disorder inventory (Wear & Pratz, Reference Wear and Pratz1987), the self-reported mean severity of depressive/anxiety symptoms, psychiatric co-morbidity, suicidal behavior or self-harm, and inpatient v. outpatient status) when provided. We contacted by email the corresponding authors of papers who we expected to have data that would fit our purposes, but who did not provide the data in their paper, and requested them to provide us with the necessary data. Data extraction was performed independently by at least two of us (M.L.M., M.vS., M.O., V.K.). Inconsistencies were resolved in consensus meetings.

Quality assessment

Based on the Newcastle-Ottawa scale (Wells et al. Reference Wells, Shea and O'Connell2016) the methodological quality of the included studies was assessed. This scale is recommended by the Cochrane collaboration (Cochrane Community Handbook, 2016). Detailed information on quality assessment is provided in Supplementary Tables S1 and S2.

Data analyses

Data analyses were performed in Stata version 13 (StataCorp, 2013). All associations were tested for statistical significance at a confidence interval (CI).

First, the prevalence rates and 95% CI of CM (by type) for all EDs and for HCs and PCs were summarized over studies, regardless whether they included a HC and/or PC reference group. Next, pooled effect-size estimates were calculated on the association between CM (coded as yes v. no exposure) and the presence of an ED using data from studies that contained an ED sample and a HC and/or PC reference group. Within ED samples pooled effect-size estimates were calculated on the association between reported CM exposure and age at onset and severity of the ED, self-reported number of binge or purge episodes per week/month, severity of depressive and anxiety symptoms, psychiatric co-morbidity, suicidal behavior, and inpatient v. outpatient status.

The random-effects model, a model that includes both sampling- and study-level error (Borenstein et al. Reference Borenstein, Hedges and Higgins2009) was adopted in all instances. As effect-size measures we chose to use odds ratios (ORs) for dichotomous outcome variables and Hedges’ g for continuous outcome variables (Hedges & Olkin, Reference Hedges and Olkin1985).

To assess the sensitivity of our results, all analyses were repeated with the data split-up as a function of type of childhood maltreatment (i.e. sexual, physical, and emotional maltreatment) and type of eating disorders [i.e. AN (subdivided by restricted [AN-R] and binge-and-purge subtype [AN-BP]), BN, EDNOS, and mixed samples].

The I 2 measure (Higgins & Thompson, Reference Higgins and Thompson2002) was used to quantify the amount of between-study heterogeneity in outcomes. Statistical significance of this measure was assessed using the χ2 statistic. Meta-regression analyses were performed in case of significant between-study heterogeneity in outcomes. Predictors in these analyses were: mean age in years and gender distribution of the sample, inpatient v. outpatient status, country in which the study was performed (coded as Western v. non-Western) and the methodological quality score of the study.

Publication-bias was assessed by means of Egger's test (Egger et al. Reference Egger, Davey Smith, Schneider and Minder1997). Where publication bias was evident, trim-and-fill procedures were applied to estimate pooled effect sizes while taking bias into account (Duval & Tweedie, Reference Duval and Tweedie2000).

Results

The electronic searches yielded 3938 publications after duplicates were removed and 22 additional records through backward searches and reference lists. Of these, 3878 were excluded leaving a number of 82 studies that reported on at least one effect size that met our inclusion criteria. Study selection, from search to inclusion, is presented in the flowchart in Supplementary Fig. S1. Table 1 and Supplementary Table S3 list the papers that were included for analyses and provide information on the general characteristics of them. Full references to the included papers also are presented in the Supplementary reference list (pp. 14–18).

Table 1. Basic characteristics of the included studies

AN, Anorexia nervosa; AN-BP, anorexia nervosa binge-purge subtype, AN-R, anorexia nervosa restrictive subtype; BED; binge-eating disorder, BN, bulimia nervosa; EA, emotional abuse (including emotional abuse and neglect and psychological abuse); HC, healthy controls; N.K. not known, PA, physical abuse; PC, psychiatric controls; SA, sexual abuse.

a This column indicates in which meta-analysis the study in the corresponding row is included:

[1] Childhood maltreatment (CM) exposure in patients with an ED.

[2] CM exposure in patients with AN v. that in HCs.

[3] CM exposure in patients with BN v. that in HCs.

[4] CM exposure in patients with BED/EDNOS v. that in HCs.

[5] CM exposure and characteristics of EDs such as age at onset or psychiatric co-morbidity (for more information we refer to the Method section of the main text).

Prevalence rates of CM in EDs and healthy and psychiatric reference samples

Table 2 provides the prevalence rates (in percentage and 95% CIs) of self-reported CM exposure (by type) for all the EDs [total number of estimates (k) = 214, N = 13 059 unique individuals], for healthy and psychiatric reference samples (k = 50, N = 15 092 and k = 26, N = 7736, respectively). In general, prevalence rates of all types of CM appeared to be two- to fourfold higher in ED samples as HC samples. However, in patients with AN of the restrictive subtype the prevalence rates of reported CM exposure were less pronounced (see Table 2) and for some ED subtypes there was overlap in 95% CIs when compared to HCs (e.g. in the case of sexual, physical, and emotional maltreatment in samples of patients with AN of the restrictive subtype). Prevalence rates of emotional CM and exposure to >1 type of CM was at least twofold higher in ED samples as compared to those reported in PC samples. ED samples did not differ significantly from PC samples regarding the prevalence of physical and sexual CM.

Table 2. Prevalence (in bold) of Child maltreatment (CM) exposure (by type) and estimated 95% CI and diagnostic status (by type)

AN, Anorexia nervosa; BN, bulimia nervosa; BED, binge-eating disorder; ED, Eeating disorder; EDNOS, eating disorder not otherwise specified.

a Double maltreatment counts entails exposure to >1 type of CM (regardless of the exact CM type).

b No confidence interval could be calculated because only one estimate was available.

The number of studies and individuals given here in this table are not ‘unique’. Individuals may appear in various cells (e.g. subjects on which data is presented in an ‘emotional CM cell’ may also appear in a ‘sexual CM cell’. N’s thus do not add up to the unique N, just as the number of effect sizes does not add up to the number of studies.

CM in EDs v. healthy reference samples: direct comparisons

The data presented in Table 2 are derived from studies that reported CM prevalence rates in ED samples only, excluding the possibility for between-group comparisons. In a next step, we excluded studies that lacked a reference group and ran meta-analyses on the difference in prevalence rates of CM in ED samples v. healthy reference groups. Analyses were run stratified by type of CM and type of ED. From the data it appeared that in ED samples (any type), a history of CM (any type) is much more often reported compared to HC reference groups (ORs in most cases >2). In some specific instances there were no significant differences in CM exposure between ED and HC samples (e.g. in sexual, and physical maltreatment in samples of patients with AN of the restrictive subtype and ED-NOS, respectively). However, in all these instances that were based on a limited number of studies, the effect-size pointed in the direction of higher reported CM in ED samples. Table 3 provides an overview of the results of these analyses together with data on between-study heterogeneity and publication bias. Forest plots, by type of ED, are provided in Supplementary Figs S2–S5. For most pooled effect-size estimates there was little evidence for between-study heterogeneity or publication bias (see Table 3).

Table 3. Child maltreatment (CM) exposure (overall and by type) and the odds (in bold) on ED (by type) v. healthy controls by type

AN, Anorexia nervosa; BN, bulimia nervosa; BED, binge-eating disorder; ED, eating disorder; EDNOS, eating disorder not otherwise specified; n.a., not available; OR, odds ratio; CI, confidence interval.

a Although BED was subsumed in the EDNOS category in DSM-IV we decided to report data separately for these illnesses where possible because EDNOS is not necessarily BED.

b This could not be calculated because only 1 or 2 estimates were available.

c Double maltreatment counts entails exposure to >1 type of CM (regardless of the exact CM type).

*Statistically significant at p < 0.05; **statistically significant at p < 0.01; ***statistically significant at p < 0.001.

CM in EDs v. psychiatric reference samples: direct comparisons

Direct comparisons of CM prevalence rates in ED samples (N = 1809) v. psychiatric reference samples (k = 36, N = 13 186; composed mostly of depressed and substance-dependent persons) showed that a history of CM was more often reported in the EDs as relative to other psychiatric illnesses (OR 1.31, 95% CI 1.08–1.58, p < 0.001). There were no marked differences for type of CM or for type of ED. A forest plot is provided (Supplementary Fig. S6) by type of ED. There was evidence for between-study heterogeneity in outcome (I 2 = 55.7, χ2 = 76.8, p < 0.001) and publication bias (Egger's t = 3.22, p < 0.01). When we accounted for publication bias by means of trim-and-fill estimates, the difference in CM exposure was no longer significant.

Association between psychiatric co-morbidity and CM severity within ED patients

Within ED patients we assessed whether CM exposure was associated with the presence of psychiatric co-morbidity and suicidal/self-harm behavior. We found CM to be associated with a statistically significant increase in the odds on psychiatric co-morbidity and suicidal/self-harm behavior. Results by type of co-morbidity are provided in Table 4a and in the Supplementary material as a forest plot by type of ED (Supplementary Fig. S7). These associations were not due to a particular CM or ED type although statistical power to show differences may have been too low due to a relatively limited number of studies (range k = 7–16, range N = 497–2109). We did not detect evidence for between-study heterogeneity (p = 0.38), but publication bias was observed (Egger's t = 2.4, p = 0.02). Effect-size estimates were not markedly different after we applied trim-and-fill estimates (data not shown).

Table 4. Child maltreatment exposure, co-morbidity (part a), and eating disorder (ED) characteristics (part b) in ED patients

OR, Odds ratio; CI, confidence interval.

a ED severity was measured eight times with the EAT (Mintz & O'Halloran, Reference Mintz and O'halloran2000), four times with the EDE (Luce & Crowther, 1999); three times with the EDI (Wear & Pratz, Reference Wear and Pratz1987) and one time with the BITE (Henderson & Freeman, Reference Henderson and Freeman1987).

b In case there was between-study heterogeneity in outcomes, we tested whether this could be explained by the pooling together of several related but distinct constructs (e.g. different Axis II disorders or different questionnaires to measure ED severity with). Heterogeneity in outcomes, however, were in all cases independent of pooling constructs.

Association between ED psychopathology and CM exposure within ED patients

We aggregated the reported associations between CM exposure and (i) severity of ED pathology measured on continuous scales such as the EAT, (ii) frequency of binge-purge episodes per week/month, (iii) age at ED onset, (iv) continuous scores on depression/anxiety severity within patients with an ED, and (v) the use of diuretics and laxatives. In general CM exposure was associated with higher severity of ED pathology, a larger number of binge/purge episodes, an earlier age at ED onset, higher symptom level of depression/anxiety, and a higher frequency of use of diuretics and laxatives (see Table 4b and forest plots; Supplementary Figs S8 and S9). In line with the idea that a history of CM is related with a more severe ED pathology was that prevalence rates of CM were in general higher in inpatients as compared to outpatients (prevalence rates of 0.45 v. 0.29 respectively, p for the difference <0.01).

Association between psychiatric co-morbidity and severity measures of CM exposure within ED patients

Forty-one effect-size estimates (N = 4683) used a continuous CM exposure measure (e.g. the sum score on the childhood trauma questionnaire) to predict general psychopathology (e.g. depression severity) and ED severity features (e.g. EAT score) within ED patients. Pooling these estimates showed that the severity of CM was positively correlated with continuous indices for depression and anxiety (weighted Pearson's r = 0.27, 95% CI 0.10–0.53, k = 19, N = 2197) but not with ED severity features (r = 0.26, 95% CI −0.11 to 0.67, k = 13, N = 1067).

Given the limited amount of studies, we were not able to formally test whether the above-reported dose-response associations were particularly due to one type of CM or ED.

Predictors of CM prevalence rates and between-study heterogeneity in outcomes

By means of meta-regression analyses we identified sources of between-study heterogeneity in prevalence rates of CM. We found that a small part of the heterogeneity could be explained by (1) the methodological quality of a study; studies of higher quality reported somewhat lower prevalence rates (R 2 = 0.05, p < 0.01); (2) sample size; studies that used a larger sample size reported somewhat lower prevalence rates (R 2 = 0.07, p < 0.01), (3) a lower age threshold for maltreatment to count as CM; studies that applied a lower age threshold (e.g. 16 v. 18 years age) reported somewhat lower prevalence rates (R 2 = 0.03, p < 0.01), and (4) year of publication; studies that were published more recently also yielded somewhat lower prevalence rates (R 2 = 0.04, p < 0.01). Besides, prevalence rates of CM were in general somewhat lower when they were acquired by means of an interview v. self-report questionnaire (prevalence rates of CM of 0.28 v. 0.34 respectively, p < 0.05). Mean age of the sample at assessment, gender distribution of a particular study, and whether the data was gathered in a Western v. a non-Western culture were not related to between-study heterogeneity in prevalence rates.

There were no statistically significant associations between the pre-specified moderators and between-study variation in OR estimates on the association between CM in ED samples v. HC and PC reference samples.

Discussion

The purpose of this study was to provide a quantitative overview of studies that report on the association between CM exposure (including type and severity) and a lifetime diagnosis of an ED (including type and severity parameters). Eighty-two studies, reporting on 13 059 individuals with an ED, 15 092 HC and 7736 PC subjects were pooled. This pool of data, the largest of its kind to date, indicated that reported CM, regardless of type (emotional, physical, or sexual), was strongly associated with the presence of all types of EDs. Overall, the lifetime prevalence rates of CM ranged from 1–35% in HCs, 5–46% in PCs, and 21–59% in individuals with any ED. The difference in prevalence rates of CM between patients with an ED and HCs is largely in line with the findings of previous meta-analyses (Rind et al. Reference Rind, Tromovitch and Bauserman1998; Smolak & Murnen, Reference Smolak and Murnen2002; Chen et al. Reference Chen, Murad, Paras, Colbenson, Sattler, Goranson, Elamin, Seime, Shinozaki, Prokop and Zirakzadeh2010; Caslini et al. Reference Caslini, Bartoli, Crocamo, Dakanalis, Clerici and Carrà2016). Some discrepancies were found when considering the specific types of CM. Caslini et al. (Reference Caslini, Bartoli, Crocamo, Dakanalis, Clerici and Carrà2016) for instance did not find a significant difference with respect to the prevalence of emotional CM among ED patients v. HCs whereas we do. The reason for this discrepancy probably is statistical power; we report on 10 studies that assessed this effect and Caslini et al. (Reference Caslini, Bartoli, Crocamo, Dakanalis, Clerici and Carrà2016) reported on six. Besides, their estimated point estimate on the association between emotional CM and EDs (OR range 2.13–4.15) lies within the CI that we estimate on this association (i.e. 2.05–7.08). In addition, we found that the associations between CM and ED were particularly strong for BN, BED and for AN of the binge-purge subtype, whereas this seemed to be less the case for AN of the restrictive subtype. This is in line with the finding that ED patients with CM reported more binging and purging. Our findings that ED patients reporting CM had an earlier age at ED onset, suffered a more severe form of the illness, and binge-purged more often as compared to their non-maltreated counterparts is new. Also new is the finding that individuals with an ED reporting CM were more likely to be diagnosed with a co-morbid psychiatric disorder and to be suicidal relative to ED patients who did not report exposure to CM is new. The finding of higher psychiatric co-morbidity in CM exposed patients with EDs is remarkable because this kind of co-morbidity tends to be the rule rather than the exception in individuals with EDs (Ulfvebrand et al. Reference Ulfvebrand, Birgegård, Norring, Högdahl and Hausswolff-Juhlin2015; Keski-Rahkonen & Mustelin, Reference Keski-Rahkonen and Mustelin2016). The fact that both CM exposed and non-exposed groups allegedly have similar co-morbidities makes a potential difference between the two less likely to be detected.

We found some evidence suggesting higher prevalence rates of CM in patients with an ED compared to PCs. This finding, however, did not remain statistically significant when we corrected for the likely presence of publication bias. So, the difference in CM exposure between patients with an ED as compared to PCs may reflect the effect of some small-scale studies reporting overly large effect-size estimates (and hence are more likely to get published). In general, these new findings should be considered with some caution because with each further breakdown of variables (i.e. subtypes of eating disorder, trauma type, or severity measure) analyses were run on smaller number of studies, and results become less reliable (Cochrane Community Handbook, 2016).

The consistency of our findings across most types of EDs and forms of maltreatment (see the forest plots), the strength and dose-response nature of the association, and the likely temporal precedence suggests a causal link of CM exposure on ED pathology. Nevertheless, in the absence of longitudinal and experimental data we cannot proof this here (Hill, Reference Hill1965). Furthermore, a large part of between-study heterogeneity in outcome could not be explained.

The mechanisms that may underlie the observed associations remain elusive. Moreover, the current meta-analytical approach cannot shed much light on the underlying processes. Yet, the outcomes do underline the importance of further elucidating the processes linking CM and EDs, with a particular focus on the associations with binging and purging behavior in the context of the enhanced negative affect (i.e. depression, suicidality and anxiety) and co-morbidity in EDs patients with reported CM. Clinically, it is important to elucidate whether similar or specific processes are at stake in patients with a history of CM compared to patients who do not report this (i.e. maladaptive emotion regulation styles (i.e. distorted cognitive schemas of self and body dissatisfaction, and dissociation), and whether this is further dependent on the type of CM a patient has been exposed to. Body dissatisfaction and dissociation may, for example, be specifically at stake in patients with histories of sexual CM (Dunkley et al. Reference Dunkley, Masheb and Grilo2010; Muehlenkamp et al. Reference Muehlenkamp, Claes, Smits, Peat and Vandereycken2011; Duarte et al. Reference Duarte, Ferreira and Pint-Gouveia2016; Preti et al. Reference Preti, Incani, Camboni, Petretto and Masala2016), while emotion dysregulation in general may be a process that is particularly prominent in individuals with a history of emotional CM (Michopoulos et al. Reference Michopoulos, Powers, Moore, Villarreal, Ressler and Bradley2015; Moulton et al. Reference Moulton, Newman, Power, Swanson and Day2015; Racine & Wildes, Reference Racine and Wildes2015). In addition, research on dysregulations in stress-sensitive neurobiological systems, including the HPA-axis and serotonin and dopamine systems is also needed to further elucidate the key neurobiological processes underlying the link between CM and EDs (Kaye et al. Reference Kaye, Wierenga, Bailer, Simmons and Bischoff-Grethe2013a , Reference Kaye, Wierenga, Bailer, Simmons, Wagner and Bischoff-Grethe b ; Nemeroff, Reference Nemeroff2016).

Strengths and limitations

In addition to the use of meta-analytic techniques, this study has as strength that it is based on large sample sizes of diagnosed ED subjects and controls and the identification of the type and severity of CM. In addition, unlike previous meta-analyses, rates of CM were also examined in a non-ED psychiatric comparison groups and dose-response associations were assessed at the predictor and outcome level.

A limitation of our work is that prospective studies were not available and mediators of the associations of interest could not be assessed by means of meta-analysis. Moreover, counting as limitation is that we were not able to model the effect of CM exposure on ED course (e.g. duration of illness). A factor that might limit the generalizability of our findings is that most studies solely included women. Hence our results may not generalize to men with EDs. Furthermore, our findings may not generalize to the child and adolescent ED population.

Clinical implications

Our results may have clinical implications. CM exposure obviously conveys important information on who is at risk for developing ED although between-study and between-subject variability in risk-estimates exist. Public health efforts will do well by prioritizing a reduction of CM prevalence and as such decrease the burden caused by the EDs (Fairburn & Harrison, Reference Fairburn and Harrison2003; Arcelus et al. Reference Arcelus, Mitchell, Wales and Nielsen2011; Mischoulon et al. Reference Mischoulon, Eddy, Keshaviah, Dinescu, Ross, Kass, Franko and Herzog2011; Mitchell et al. Reference Mitchell, Mazzeo, Schlesinger, Brewerton and Smith2012; Smink et al. Reference Smink, Van Hoeken and Hoek2013; Yao et al. 2016) by depression (Nanni et al. Reference Nanni, Uher and Danese2012), substance abuse, suicide, and anxiety (Norman et al. Reference Norman, Byambaa, De, Butchart, Scott and Vos2012; Mason et al. 2014; Yao et al. 2016). Moreover, the importance of CM exposure should also be addressed in the clinical setting. As shown above, reported CM is indicative of a more severe clinical profile. Moreover, one study that we know of accounted for the impact of CM on ED treatment outcome and found that ED patients exposed to CM improve less as compared to those who were not exposed (Vrabel et al. Reference Vrabel, Hoffart, Rø, Martinsen and Rosenvinge2010; Norman et al. Reference Norman, Byambaa, De, Butchart, Scott and Vos2012), a phenomenon that is well-documented in the treatment of depression (Nanni et al. Reference Nanni, Uher and Danese2012). Standard assessment of CM exposure and additional treatment interventions for those reporting CM, such as enhancing emotion regulation strategies and/or focused trauma treatment may increase the rather low treatment responses in ED patients.

Conclusions

We performed the most comprehensive meta-analysis of the association between CM exposure and EDs to date. We found that CM, regardless of whether it is emotional, sexual and/or physical in nature, was strongly associated with the presence of all types EDs and severity parameters that characterize these illnesses, often in a dose dependent manner. Notwithstanding strong and consistent associations, unexplained between-study heterogeneity in outcomes remained for most of the associations that we observed. This implies between-study and probably between-subject variability in risk estimates that need to be explained in future work on this important topic.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/S0033291716003561.

Acknowledgements

This work is supported by Leiden University, The Netherlands (Molendijk). The funder had no role in the design and conduct of the study, the interpretation of the data, approval of the manuscript, and the decision to submit the manuscript for publication. We thank Majorie van Seggelen, Marit Oosterom, and Vera Kempen for their assistance with data collection.

Declaration of Interest

None.

References

Anderson, KP, Laporte, D, Brandt, H, Crawford, S (1997). Sexual abuse and bulimia: response to inpatient treatment and preliminary outcome. Journal of Psychiatric Research 31, 621–633.Google Scholar
Allison, KC, Grilo, CM, Masheb, RM, Stunkard, AJ (2007). High self-reported rates of neglect and emotional abuse, by persons with binge eating disorder and night eating syndrome. Behaviour Research and Therapy 45, 2874–2883.Google Scholar
APA (1980). Diagnostic and Statistical Manual of Mental Disorders, 3th edn. American Psychiatric Association: Washington, DC.Google Scholar
APA (2000). Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Association: Washington, DC.Google Scholar
APA (2013). Diagnostic and Statistical Manual of Mental Disorders, 5th edn. American Psychiatric Association: Washington, DC.Google Scholar
Arcelus, J, Mitchell, AJ, Wales, J, Nielsen, S (2011). Mortality rates in patients with anorexia nervosa and other eating disorders: a meta-analysis of 36 studies. Archives of General Psychiatry 68, 724–731.CrossRefGoogle ScholarPubMed
Backholm, K, Isomaa, R, Birgegård, A (2013). The prevalence and impact of trauma history in eating disorder patients. European Journal of Psychotraumatology 4, 1–8.Google Scholar
Bardone-Cone, AM, Maldonado, CR, Crosby, RD, Mitchell, JE, Wonderlich, SA, Joiner, TE Jr. et al. (2008). Revisiting differences in individuals with bulimia nervosa with and without a history of anorexia nervosa: eating pathology, personality, and maltreatment. International Journal of Eating Disorders 41, 697–704.CrossRefGoogle ScholarPubMed
Basurte, E, Díaz-Marsá, M, Martín, O, Carrasco, JL (2004). Traumatic childhood background, impulsiveness and hypothalamus-pituitary-adrenal dysfunction in eating disorders; a pilot study. Actas Españolas Psiquiatricas 32, 149–152.Google ScholarPubMed
Becker, DF, Grilo, CM (2011). Childhood maltreatment in women with binge-eating disorder: associations with psychiatric comorbidity, psychological functioning, and eating pathology. Eating and Weight Disorders 16, 113–120.Google Scholar
Borenstein, M, Hedges, LV, Higgins, JPT (2009). Introduction to Meta-Analyses. Chichester Wiley: NH, USA.CrossRefGoogle Scholar
Brewerton, TD (2007). Eating disorders, trauma, and comorbidity: focus on PTSD. Eating Disorders 15, 285–304.Google Scholar
Brewerton, TD, Rance, SJ, Dansky, BS, O'Neil, PM, Kilpatrick, DG (2014). A comparison of women with child-adolescent versus adult onset binge eating: results from the National Women's Study. International Journal of Eating Disorders 47, 836–843.CrossRefGoogle ScholarPubMed
Brown, L, Russell, J, Thornton, C, Dunn, S (1997). Experiences of physical and sexual abuse in Australian general practice attenders and an eating disordered population. Australian and New Zealand Journal of Psychiatry 31, 398–404.Google Scholar
Carter, JC, Bewell, C, Blackmore, E, Woodside, DB (2006). The impact of childhood abuse in anorexia nervosa. Child Abuse Neglect 30, 257–269.CrossRefGoogle ScholarPubMed
Casper, RC, Lyubomirsky, S (1997). Individual psychopathology relative to reports of unwanted sexual experiences as predictor of a bulimic eating pattern. International Journal of Eating Disorders 21, 229–236.Google Scholar
Castellini, G, Lo Suaro, C, Lelli, L, Godini, L, Vignozzi, L, Rellini, AH, Faravelli, C, Maggi, M, Ricca, V (2013). Childhood sexual abuse moderates the relationship between sexual functioning and eating disorder psychopathology in anorexia nervosa: a 1-year follow-up study. International Society of Sexual Medicine 10, 2190–2200.Google Scholar
Claes, L, Vandereycken, W (2007). Is there a link between traumatic experiences and self-injurious behaviors in eating-disordered patients? Treatment and Prevention 15, 305–315.Google Scholar
Corstorphine, E, Waller, G, Lawson, R, Ganis, C (2007). Trauma and multiple-impulsivity in the eating disorders. Eating Behaviour 8, 23–30.Google Scholar
Cumella, EJ, Kally, Z (2008). Comparison of middle-age and young women inpatients with eating disorders. Eating and Weight Disorders 13, 183–190.Google Scholar
Caslini, M, Bartoli, F, Crocamo, C, Dakanalis, A, Clerici, M, Carrà, G (2016). Disentangling the association between child abuse and eating disorders: a systematic review and meta-analysis. Psychosomatic Medicine 78, 79–90.Google Scholar
Chen, LP, Murad, MH, Paras, ML, Colbenson, KM, Sattler, AL, Goranson, EN, Elamin, MB, Seime, RJ, Shinozaki, G, Prokop, LJ, Zirakzadeh, A (2010). Sexual abuse and lifetime diagnosis of psychiatric disorders: a systematic review and meta-analysis. Mayo Clinic Proceedings 85, 618–629.Google Scholar
Cochrane Community Handbook (2016) (http://community.cochrane.org/handbook). Accessed 10 February 2016).Google Scholar
Deep, AL, Lilenfeld, LR, Plotnicov, KH, Pollice, C, Kaye, WH (1999). Sexual abuse in eating disorder subtypes and control women: the role of comorbid substance dependence in bulimia nervosa. International Journal of Eating Disorders 25, 1–10.Google Scholar
Duarte, C, Ferreira, C, Pint-Gouveia, J (2016). At the core of eating disorders: overvaluation, social rank, self-criticism and shame in anorexia, bulimia and binge eating disorder. Comprehensive Psychiatry 66, 123–131.Google Scholar
Dunkley, DM, Masheb, RM, Grilo, CM (2010). Childhood maltreatment, depressive symptoms, and body dissatisfaction in patients with binge eating disorder: the mediating role of self-criticism. International Journal of Eating Disorders 43, 274–281.Google Scholar
Duval, S, Tweedie, RL (2000). A nonparametric trim and fill method of accounting for publication bias in meta-analysis. Journal of the American Statistical Society 95, 89–98.Google Scholar
Egger, M, Davey Smith, G, Schneider, M, Minder, C (1997). Bias in meta-analysis detected by a simple, graphical test. British Medical Journal 315, 629–634.CrossRefGoogle ScholarPubMed
Fairburn, CG, Welch, SL, Doll, HA, Davies, BA, O'Conner, ME (1997). Risk Factors for Bulimia Nervosa. Archives of General Psychiatry 54, 509–517.CrossRefGoogle ScholarPubMed
Fairburn, CG, Doll, HA, Welch, SL, Hay, PJ, Davies, BA, O'Connor, ME (1998). Risk factors for binge eating disorders: a community-based, case control study. Archives of General Psychiatry 55, 425–432.CrossRefGoogle ScholarPubMed
Favaro, A, Santonastaso, P (1997). Suicidality in eating disorders: clinical and psychological correlates. Acta Psychiatrica Scandinavia 95, 508–514.Google Scholar
Favaro, A, Grave, D, Santonastaso, P (1998). Impact of a history of physical and sexual abuse in eating disordered and asymptomatic subjects. Acta Psychiatrica Scandinavia 97, 358–363.CrossRefGoogle ScholarPubMed
Favaro, A, Tenconi, E, Santonastaso, P (2010). The interaction between perinatal factors and childhood abuse in the risk of developing anorexia nervosa. Psychological Medicine 40, 657–665.CrossRefGoogle ScholarPubMed
Feldman, MB, Meyer, IH (2007). Childhood abuse and eating disorders in gay and bisexual men. International Journal of Eating Disorders 49, 418–423.Google Scholar
Folsom, V, Krahn, D, Nairn, K, Gold, L, Demitrack, MA, Silk, KR (1993). The impact of sexual and physical abuse on eating disordered and psychiatric symptoms: a comparison of eating disordered and psychiatric inpatients. International of Journal Eating Disorders 13, 249–257.Google Scholar
Friedman, MA, Wilfley, DE, Welch, RR, Kunce, JT (1997). Self-directed hostility and family functioning in normal-weight bulimics and overweight binge eaters. Addictive Behaviour 22, 367–375.CrossRefGoogle ScholarPubMed
Fullerton, DT, Wonderlich, SA, Gosnell, BA (1995). Clinical characteristics of eating disorder patients who report sexual or physical abuse. International Journal of Eating Disorders 17, 243–249.3.0.CO;2-Z>CrossRefGoogle ScholarPubMed
Fairburn, CG, Harrison, PJ (2003). Eating disorders. Lancet 361, 407–416.Google Scholar
Garfinkel, PE, Lin, E, Goering, P, Spegg, C, Goldbloom, DS, Kennedy, S Kaplan, AS, Woodside, D. (1995). Bulimia nervosa in a Canadian community sample: prevalence and comparison of subgroups. American Journal of Psychiatry 152, 1052–1058.Google Scholar
Grilo, CM, Masheb, RM (2001). Childhood psychological, physical, and sexual maltreatment in outpatients with binge eating disorder: frequency and associations with gender, obesity, and eating related psychopathology. Obesity Research 9, 320–325.Google Scholar
Grilo, CM, Masheb, RM (2002). Childhood maltreatment and personality disorders in adult patients with binge eating disorder. Acta Psychiatrica Scandinavia 106, 183–188.Google Scholar
Groleau, P, Joober, R, Israel, M, Zeramdini, N, DeGuzman, R, Steiger, H (2014). Methylation of the dopamine D2 receptor (DRD2) gene promotor in women with bulimia-spectrum disorder: associations with borderline personality disorder and exposure to childhood abuse. Journal of Psychiatric Research 48, 121–127.Google Scholar
Hartt, J, Waller, G (2002). Child abuse, dissociation, and core beliefs in bulimic disorders. Child Abuse and Neglect 26, 923–938.Google Scholar
Herzog, Db, Staley, JE, Carmody, S, Robbins, WM, van der Kolk, BA (1993). Childhood sexual abuse in anorexia nervosa and bulimia nervosa: a pilot study. Journal of the American Academy of Child Psychiatry 32, 962–966.Google Scholar
Henderson, M, Freeman, CP (1987). A self-rating scale for bulimia. The ‘BITE’. The British Journal of Psychiatry 150(1), 18–24.Google Scholar
Hepp, U, Spindler, A, Schnyder, U, Kraemer, B, Milos, G (2007). Post-traumatic stress disorder in women with eating disorders. Eating and Weigth Disorders 12, 24–27.Google Scholar
Jaite, C, Schneider, N, Hilbert, A, Pfeiffer, E, Lehmkuhl, U, Salbach-Andrae, H (2012). Etiological role of childhood emotional trauma and neglect in adolescent anorexia nervosa: a cross-sectional questionnaire analysis. Psychopathology 45, 61–66.Google Scholar
Haidich, AB (2010) Meta-analysis in medical research. Hippokratia 14S, 29.Google Scholar
Hedges, LV, Olkin, I (1985). Statistical Method for Meta-Analysis. Princeton, NJ, USA. Academic Press.Google Scholar
Higgins, JPT, Thompson, SG (2002). Quantifying heterogeneity in meta-analysis. Statistics in Medicine 21, 1539–1558.Google Scholar
Hill, AB (1965). The environment and disease: association or causation? Proceedings of the Royal Society of Medicine 58, 295–300.Google Scholar
Karwautz, A, Rabe-Hesketh, S, Hu, X, Zhao, J, Sham, P, Collier, DA, Treasure, JL (2001). Individual-specific risk factors for anorexia nervosa: a pilot study using a discordant sister-pair design. Psychological Medicine 31, 317–329.Google Scholar
Kaye, WH, Wierenga, CE, Bailer, UF, Simmons, AN, Bischoff-Grethe, A (2013 a). Nothing tastes as good as skinny feels: the neurobiology of anorexia nervosa. Trends in Neuroscience 36, 110–120.Google Scholar
Kaye, WH, Wierenga, CE, Bailer, UF, Simmons, AN, Wagner, A, Bischoff-Grethe, A (2013 b). Does a shared neurobiology for foods and drugs of abuse contribute to extremes of food ingestion in anorexia and bulimia nervosa? Biological Psychiatry 73, 836–842.Google Scholar
Kern, J, Hastings, T (1995). Differential family environment of bulimics and victims of childhood sexual abuse: achievement orientation. Journal of Clinical Psychology 51, 499–506.Google Scholar
Kong, S, Bernstein, K (2009). Childhood trauma as a predictor of eating psychopathology and its mediating variables in patients with eating disorders. Journal of Clinical Nursing 18, 1897–1907.Google Scholar
Kugu, N, Akyuz, G, Dogan, O, Ersan, E, Izgic, F (2006). The prevalence of eating disorders among University students and the relationship with some individual characteristics. Aus NZ Journal Psychiatry 40, 129–135.CrossRefGoogle ScholarPubMed
Kent, A, Waller, G (2000). Childhood emotional abuse and eating psychopathology. Clinical Psychology Review 20, 887–903.Google Scholar
Keski-Rahkonen, A, Mustelin, L (2016). Epidemiology of eating disorders in Europe: prevalence, incidence, comorbidity, course, consequences, and risk factors. Current Opinion in Psychiatry 29, 340–345.Google Scholar
Klump, KL, Wonderlich, S, Lehoux, P, Lilenfeld, LRR, Bulik, CM (2002). Does environment matter? A review of non-shared environment and eating disorders. International Journal of Eating Disorders 31, 118–132.Google Scholar
Léonard, S, Steiger, H, Kao, A (2003). Childhood and adulthood abuse in bulimic and nonbulimic women: prevalence and psychological correlates. International Journal of Eating Disorders 33, 397–405.Google Scholar
Lockwood, R, Lawson, R, Waller, G (2004). Compulsive features in the eating disorders: a role for trauma. Journal of Nervous and Mental Disease 192, 247–249.Google Scholar
Luce, KH, Crowther, JH, Pole, M (2008). Eating Disorder Examination Questionnaire (EDE-Q): norms for undergraduate women. International Journal of Eating Disorders 41, 273–276.Google Scholar
Mahon, J, Bradley, NS, Harvey, PK, Winston, AP, Palmer, RL (2001). Childhood trauma has dose effect relationship with dropping out from psychotherapeutic treatment for bulimia nervosa: a replication. International Journal of Eating Disorders 30, 138–148.CrossRefGoogle ScholarPubMed
Machado, BC, Gonçalves, SF, Martins, C, Hoek, HW, Machado, PP (2014). Risk factors and antecedent life events in he development of anorexia nervosa: A Portuguese case-control study. European Eating Disorders Reviews 22, 243–251.Google Scholar
Mangweth-Matzek, B, Rupp, CI, Hausmann, A, Gusmerotti, S, Kemmler, G, Biebl, W (2010). Eating disorders in men: current features and childhood factors. Eating and Weight Disorders 15, 15–22.Google Scholar
Monteleone, AM, Monteleone, P, Serino, I, Scognamiglio, P, Di Genio, M, Maj, M (2015). Childhood trauma and cortisol awakening response in symptomatic patients with anorexia nervosa and bulimia nervosa. International Journal of Eating Disorders 48(6), 615–621.Google Scholar
Matsunaga, H, Kaye, WH, McConaha, CRN, Plotonicov, K, Pollice, CBA, Rao, R, Stein, D (1999). Psychopathological characteristics of recovered bulimics who have a history of physical or sexual abuse. Journal of Nervous and Mental Disease 187, 472–477.Google Scholar
Michopoulos, V, Powers, A, Moore, C, Villarreal, S, Ressler, KJ, Bradley, B (2015). The mediating role of emotion dysregulation and depression on the relationship between childhood trauma exposure and emotional eating. Appetite 91, 129–136.CrossRefGoogle ScholarPubMed
Mintz, LB, O'halloran, MS (2000). The eating attitudes test: validation with DSM-IV eating disorder criteria. Journal of Personality Assessment 74, 489–503.Google Scholar
Mischoulon, D, Eddy, KT, Keshaviah, A, Dinescu, D, Ross, SL, Kass, AE, Franko, DL, Herzog, DB (2011). Depression and eating disorders: treatment and course. Journal of Affective Disorders 130, 470–477.Google Scholar
Mitchell, KS, Mazzeo, SE, Schlesinger, MR, Brewerton, TD, Smith, BN (2012). Comorbidity of partial and subtreshold PTSD among men and women with eating disorders in the national comorbidity survey-replication. International Journal of Eating Disorders 45, 307–315.Google Scholar
Moher, D, Liberati, A, Tetzlaff, J, Altman, DG (2009) Preferred reporting items for systematic reviews and meta-analyses: the prisma statement. Annals of Internal Medicine 151, 264–269.Google Scholar
Moulton, SJ, Newman, E, Power, K, Swanson, V, Day, K (2015). Childhood trauma and eating psychopathology: a mediating role for dissociation and emotion dysregulation? Child Abuse and Neglect 39, 167–174.Google Scholar
Muehlenkamp, JJ, Claes, L, Smits, D, Peat, CM, Vandereycken, W (2011). Non--suicidal self-injury in eating disordered patients: a test of a conceptual model. Psychiatry Research 188, 102–108.Google Scholar
Nagata, H, Kaye, WH, Kiriki, N, Rao, R, Plotnicov, KH (2001). Physical and sexual abuse histories in patients with eating disorders: A comparison of Japanese and American patients. Psychiatry and Clinical Neuroscience 55, 333–340.Google Scholar
Nickel, C, Simek, M, Moleda, A, Muehlbacher, M, Buschmann, W, Fartacek, R Bachler, E, Egger, C, Rother, WK, Loew, TH, Nickel, MK (2006). Suicide attempts versus suicidal ideation in bulimic female adolescents. Pediatrics International 48, 374–381.Google Scholar
Nanni, V, Uher, R, Danese, A (2012). Childhood maltreatment predicts unfavorable course of illness and treatment outcome in depression: a meta-analysis. American Journal of Psychiatry 169, 141–151.CrossRefGoogle ScholarPubMed
Nemeroff, CB (2016). Paradise lost: the neurobiological and clinical consequences of child abuse and neglect. Neuron 89, 892–909.Google Scholar
Norman, RE, Byambaa, M, De, R, Butchart, A, Scott, J, Vos, T (2012). The long-term health consequences of child physical abuse, emotional abuse, and neglect: a systematic review and meta-analysis. PLoS Medicine 9, e1001349.Google Scholar
Olivardia, R, Pope, HG, Mangweth, B, Hudson, JL (1995). Eating disorders in college men. American Journal of Psychiatry 152, 1279–1285.Google Scholar
Oppenheimer, R, Howells, K, Palmer, RL, Chaloner, DA (1985). Adverse sexual experience in childhood and clinical eating disorders: a preliminary description. Journal of Psychiatric Research 19, 357–361.CrossRefGoogle ScholarPubMed
Palmer, RL, Oppenheimer, R, Dignon, A, Chaloner, DA, Howells, K (1990). Childhood sexual experiences with adults reported by women with eating disorders: an extended series. British Journal of Psychiatry 156, 699–703.Google Scholar
Pitts, C, Waller, G (1993). Self-denigratory beliefs following sexual abuse: association with the symptomatology of bulimic disorders. International Journal of Eating Disorders 13, 407–410.3.0.CO;2-7>CrossRefGoogle ScholarPubMed
Pope, HG, Mangweth, B, Brooking Negrão, MA, Hudson, JI, Cordas, TA (1994). Childhood sexual abuse and bulimia nervosa: a comparison of American, Austrian, and Brazilian women. American Journal of Psychiatry 151, 732–737.Google Scholar
Preti, A, Incani, E, Camboni, MV, Petretto, DR, Masala, C (2016). Sexual and eating disorder symptoms: the mediator role of bodily dissatisfaction. Comprehensive Psychiatry 47, 475–481.Google Scholar
Racine, SE, Wildes, JE (2015). Emotion dysregulation and anorexia nervosa: an exploration of the role of childhood abuse. International Journal of Eating Disorders 48, 55–58.CrossRefGoogle ScholarPubMed
Rayworth, BB, Wise, LA, Harlow, BL (2004). Childhood abuse and risk of eating disorders in women. Epidemiology 15, 271–278.Google Scholar
Rind, B, Tromovitch, P, Bauserman, R (1998). A meta-analytic examination of assumed properties of child sexual abuse using college samples. Psychological Bulletin 124, 22–53.Google Scholar
Röhr, S, Dölemeyer, R, Klinitzke, G, Steinig, J, Wagner, B, Kersting, A (2015). Kindesmishandlung bei binge eating Störung: ein systematischer Literaturüberblick. Psychiatrische Praxis 42, 125–132.Google Scholar
Romans, SE, Gendall, KA, Martin, JL, Mullen, PE (2001). Child sexual abuse and later disordered eating: a New Zealand epidemiological study. International Journal of Eating Disorders 29, 380–392.Google Scholar
Rorty, M, Yager, J, Rossotto, E (1994). Childhood sexual, physical, and psychological abuse and their relationship to comorbid psychopathology in bulimia nervosa. International Journal of Eating Disorders 21, 317–334.Google Scholar
Richardson, J, Steiger, H, Schmitz, N, Joober, R, Bruce, KR, Israel, M, Gauvin, L, Anestin, AS, Dandurand, C, Howard, H, de Guzman, R (2008). Relevance of the 5-HTTLPR polymorphism and childhood abuse to increased psychiatric comorbidity in women with bulimia spectrum disorders. Journal of Clinical Psychiatry 69, 981–990.Google Scholar
Root, MPP, Fallon, P (1988). The incidence of victimization experiences in a bulimic sample. Journal of Interpersonal Violence 3, 161–173.Google Scholar
Sanci, L, Coffey, C, Olsson, C, Reid, S, Carlin, JB, Patton, G (2008). Childhood sexual abuse and eating disorders in females. Findings from the Victorian adolescent health cohort study. Archives of Pediatric and Adolescent Medicine 162, 261–267.Google Scholar
Steiger, H, Zanko, M (1990). Sexual traumata among eating-disordered, psychiatric, and normal female groups. Journal of Interpersonal Violence 5, 74–86.Google Scholar
Steiger, H, Richardson, J, Joober, R, Israel, M, Bruce, KR, Ying Kin, NMK, Howard, H, Anestin, A, Dandurand, C, Gauvein, L (2008). Dissocial behavior, the 5HTTLPR polymorphism, and maltreatment in women with bulimic syndromes. American Journal of Medical Genetics part B (Neuropsychiatric Genetics) 147B, 128–130.Google Scholar
Steiger, H, Richardson, J, Schmitz, N, Israel, M, Bruce, KR, Gauvin, L (2010). Trait-defined eating disorder subtypes and history of childhood abuse. International Journal of Eating Disorders 43, 428–423.Google Scholar
Sachs-Ericsson, N, Keel, PK, Selby, EA, Verona, E, Cougle, JR, Palmer, E (2012). Parental disorders, childhood abuse, and binge eating in a large community sample. International Journal of Eating Disorders 45, 316–325.Google Scholar
Stuart, GW, Laraia, MT, Ballenger, JC, Lydiard, RB (1990). Early family experiences of women with bulimia and depression. Archives of Psychiatric Nursing 4, 43–52.Google Scholar
Schmidt, UH, Tiller, J, Treasure, JL (1993). Setting the scene for eating disorders: childhood care, classification and course of illness. Psychological Medicine 23, 663–672.Google Scholar
Steiger, H, Léonard, S, Kin, NMK, Ladouceur, C, Ramdoyal, D, Young, SN (2000). Childhood abuse and platelet tritiated-paroxetine binding in bulimia nervosa: implications for borderline personality disorder. Journal of Clinical Psychiatry 61, 427–435.Google Scholar
Schmidt, U, Evans, K, Tiller, J, Treasure, J (1995). Puberty, sexual milestones and abuse: how are they related in eating disorder patients. Psychological Medicine 25, 413–417.Google Scholar
Sullivan, PF, Bulik, CM, Carter, FA, Joyce, PR (1995). The significance of a history of childhood sexual abuse in bulimia nervosa. British Journal of Psychiatry 167, 679–682.Google Scholar
Steiger, H, Jabalpurwala, S, Champagne, J (1996). Axis II comorbidity and developmental adversity in bulimia nervosa. Journal of Nervous and Mental Disease 184, 555–560.Google Scholar
Schoemaker, C, Smit, F, van Bijl, R, Vollebergh, WAM (2002). Bulimia Nervosa following psychological and multiple child abuse: support for the self-medication hypothesis in a population-based cohort study. International Journal of Eating Disorders 32, 381–388.CrossRefGoogle Scholar
Striegel-Moore, RH, Dohm, F-A, Pike, KM, Wifley, DE, Fairburn, CG (2002). Abuse, bullying, and discrimination as risk factors for binge eating disorder. American Journal of Psychiatry 159, 1902–1907.Google Scholar
Steiger, H, Gauvin, L, Israël, M, Kin, NMKN, Young, SN, Roussin, J (2004). Serotonin function, personality variations and childhood abuse in women with bulimia-spectrum eating disorders. Journal of Clinical Psychiatry 65, 830–837.Google Scholar
Tasca, GA, Ritchie, K, Zachariades, F, Proulx, G, Trinneer, A, Balfour, L, Demidenko, N, Hayden, G, Wong, A, Bissada, H (2013). Attachment insecurity mediates the relationship between childhood trauma and eating disorder psychopathology in a clinical sample: a structural equation model. Child Abuse and Neglect 37, 926–933.Google Scholar
Smink, FRE, Van Hoeken, D, Hoek, HW (2013). Epidemiology, course and outcome of eating disorders. Current Opinion in Psychiatry 26, 543–548.CrossRefGoogle ScholarPubMed
Smolak, L, Murnen, SK (2002). A meta-analytic examination of the relationship between child sexual abuse and eating disorders. International Journal of Eating Disorders 31, 136–150.Google Scholar
StataCorp (2013). Stata Statistical Software: Release 13. StataCorp LP: College Station, TX.Google Scholar
Trace, SE, Baker, JH, Peñas-Lledó, E, Bulik, CM (2013). The genetics of eating disorders. Annual Review of Clinical Psychology 9, 589–562.Google Scholar
Ulfvebrand, S, Birgegård, A, Norring, C, Högdahl, L, Hausswolff-Juhlin, Y (2015). Psychiatric comorbidity in women and men with eating disorders results from a large clinical database. Psychiatry Research 230, 294–299.Google Scholar
Vrabel, KR, Hoffart, A, Rø, Ø, Martinsen, EW, Rosenvinge, JH (2010). Co-occurrence of avoidant personality disorder and child sexual abuse predicts poor outcome in long-standing eating disorder. Journal of Abnormal Psychology 119, 623–629.Google Scholar
Van Gerko, K, Hughes, ML, Hamil, M, Waller, G (2005). Reported childhood sexual abuse and eating-disordered cognitions and behaviors. Child Abuse and Neglect 18, 375–382.Google Scholar
Vaz Leal, FJ, Guisado Marcias, JA, Garcia-Herraiz, MA, López Vinuesa, B, Monge Bautista, M, Bolivar Perálvarez, M (2005). History of sexual abuse in patients with bulimia nervosa: its influence on clinical status. Actas Españolas Psiquiatricas 33, 135–140.Google Scholar
Waller, G (1994). Childhood sexual abuse and borderline personality disorder in the eating disorders. Child Abuse and Neglect 18, 97–101.Google Scholar
Waller, G (1998). Perceived control in eating disorders: relationship with reported sexual abuse. International Journal of Eating Disorders 23, 213–216.Google Scholar
Waller, G (1992). Sexual abuse and the severity of bulimic symptoms. British Journal of Psychiatry 161, 90–93.Google Scholar
Waller, G, Meyer, C, Ohanian, V, Elliott, P, Dickson, C, Sellings, J (2001). The psychopathology of bulimic women who report childhood sexual abuse: the mediating role of core beliefs. Journal of Nervous and Mental Disease 189, 700–708.Google Scholar
Webster, JJ, Palmer, RL (2000). The childhood and family background of women with clinical eating disorders: a comparison with women with major depression and women without psychiatric disorder. Psychological Medicine 30, 53–60.Google Scholar
Wentz, E, Gillberg, IC, Gillberg, G, Råstman, M (2005). Fertility and history of sexual abuse at 1-year follow-up of adolescent-onset anorexia nervosa. International Journal of Eating Disorders 37, 294–298.Google Scholar
Welch, SL, Doll, HA, Fairburn, CG (1997). Life events and the onset of bulimia nervosa. Psychological Medicine 27, 515–522.Google Scholar
Wonderlich, SA, Rosenfeldt, S, Crosby, RD, Mitchell, JE, Engel, SG, Smyth, J, Miltenberger, R (2007). The effects of childhood trauma on daily mood lability and comorbid psychopathology in bulimia nervosa. Journal of Traumatic Stress 20, 77–87.Google Scholar
Wear, RW, Pratz, O (1987). Test-retest reliability for the eating disorder inventory. International Journal of Eating Disorders 6, 767–769.Google Scholar
Wells, GA, Shea, B, O'Connell, D (2016). The Newcastle-Ottawa Scale (NOS) for assessing the quality non-randomised studies in meta-analyses. Ottawa Healthcare Institute (http://www.ohri.ca). Accessed 2 February 2016.Google Scholar
WHO (2016 a). International Classification of Diseases. World Health Organization: Geneva.Google Scholar
WHO (2016 b). Childhood abuse (2016). (http://www.who.int/topics/child_abuse/en/). World Health Organization. Accessed 3 January 2016.Google Scholar
Yao, S, Kuja-Halkola, R, Thorton, LM, Runfola, CD, D'Onofrio, BM, Almqvist, C, Lichtenstein, P, Sjölander, A, Larsson, H, Bulik, CM (2016). Familial liability for eating disorders and suicide attempts: evidence from a population registry in Sweden. JAMA Psychiatry 73, 284–291.Google Scholar
Figure 0

Table 1. Basic characteristics of the included studies

Figure 1

Table 2. Prevalence (in bold) of Child maltreatment (CM) exposure (by type) and estimated 95% CI and diagnostic status (by type)

Figure 2

Table 3. Child maltreatment (CM) exposure (overall and by type) and the odds (in bold) on ED (by type) v. healthy controls by type

Figure 3

Table 4. Child maltreatment exposure, co-morbidity (part a), and eating disorder (ED) characteristics (part b) in ED patients

Supplementary material: File

Molendijk supplementary material

Molendijk supplementary material 1

Download Molendijk supplementary material(File)
File 3 MB