The possible relationship between sexual abuse and the development of an eating disorder has gained attention over the last few years. Researchers have attempted to clarify this potential link using a variety of population samples and research methodologies. As will be shown, the results of these investigations are rather diverse and sometimes inconclusive. In the following review of the literature, the complex relationship between sexual abuse and eating disorders will be examined while also discussing the methodological limitations of the various designs.
Steiger and Zanko (1990) compared rates of incestuous abuses (sexual contacts with family members) and extrafamilial abuses (sexual traumata involving other perpetrators) among eating disordered women who met DSM-III-R (American Psychiatric Association, 1987) criteria, women with psychiatric disturbances, and normal women. The authors’ interest in the psychological effects of abuse led them to examine psychological defenses which are believed to filter perceptions and affects. Defenses were of interest to the authors for two reasons: (1) incest victims often resort to maladaptive defenses with a self-victimizing quality, in which anger at others is expressed through self-sabotaging acts; and (2) the authors’ previous work suggests that eating disordered women use primitive defenses when compared to normal and psychiatrically disturbed women. This particular study was designed to determine the degree to which traumata like sexual abuse might have effects upon defense-style development.
In order to compare rates across eating disorder subtypes, the eating disordered women were divided into the following groups: (1) Anorexic Restricters (n=16); (2) Anorexic Bingers (n=12); (3) Bulimics with an Anorexic History (n=20); and (4) Bulimics with no prior Anorexia (n=25). To compare rates of sexual traumata among eating disordered women to those among women with other psychiatric disturbances (eating disorders excluded), a group of 21 women in hospital inpatient or outpatient treatments was formed, all within the age range of the eating disordered subjects who were not actively psychotic or heavily medicated. A normal control group contained 24 women consisting of hospital staff, parents, friends, and students comparable in age to the eating disordered subjects.
The Defense Style Questionnaire (DSQ, Bond, Gardner, Christian, & Sigal, 1983; Bond & Vaillant, 1986) was used to assess defense styles. In order to study sexual traumata, the authors used a self-report questionnaire on which subjects indicated the following about sexual traumata during their childhood and adolescence: (1) the perpetrator; (2) the victim; (3) the nature of the abuse; (4) the frequency of the abuse; and (5) their age when the abuse occurred. Also, to determine the presence of eating disorders among the control subjects, the authors chose to use the Eating Attitudes Test (Garner & Garfinkel, 1979).
The authors noted that 30% of the eating disordered women reported sexual traumata; however, such traumata was prevalent among bulimics (particularly those with no anorexic history), but was rare among anorexic restricters. Furthermore, subjects showing mixed anorexic and bulimic symptoms, showed prevalence of sexual trauma between that of the restricters and the bulimics. The authors also reported that the bingers were highly comparable to the psychiatrically disturbed subjects in regard to past sexual abuse. Finally, the authors concluded that sexual abuse and incest should not be viewed as variables causing the development of eating disorders, but may be markers of other family and developmental features which may have pathogenic effects on children which, in turn, may lead to the development of eating disorder symptoms.
Welch and Fairburn (1994) designed a case control study of the relationship between sexual abuse and the subsequent development of bulimia nervosa. The study had three main objectives. First, the authors sought to determine whether sexual abuse increases the risk of developing bulimia nervosa by using 50 subjects with bulimia nervosa and 100 normal comparison subjects. The second objective was to see whether any increase in risk identified was specific to bulimia nervosa or whether it reflected an increased risk of psychiatric disorders in general by examining differences in the 50 bulimic subjects and 50 subjects with other psychiatric disorders. Finally, the authors sought to determine whether a clinic group of patients with bulimia nervosa differed from a community group in terms of their histories of sexual abuse by comparing 50 community subjects with bulimia nervosa and 50 clinic cases of bulimia nervosa.
The authors found that only a minority of subjects reported a history of sexual abuse. Furthermore, they also found that a history of sexual abuse was just as common in the group with psychiatric disorders as in the bulimic women. Thus, they concluded that sexual abuse may be a risk factor for psychiatric disorders in general, but should not be considered a specific risk factor for the development of bulimia nervosa.
In a study using 786 women identified as bulimics (n=30, subclinical bulimics (n=37), and normal subjects (n=719) by scores on the Bulimia Test--Revised (BULIT-R Thelan et al., 1991) Hastings and Kern (1994) sought to determine if a significant association exists between bulimia and childhood sexual abuse (CSA). In addition to investigating this association, the authors also examined the relationships between family environment, CSA, and bulimia. The authors report that subjects were given three questionnaires in counterbalanced order to assess bulimia, CSA, and family environment. These questionnaires were: (1) the Bulimia Test--Revised (BULIT-R Thelan et al., 1991); (2) the Child Sexual Abuse Questionnaire (Walters, Smolak, & Sullins, 1987); and (3) the Family Environment Scale (FES), Form R (Moos & Moos, 1984).
After conducting a series of chi-square analyses and one-way analyses of variance (ANOVAs), the authors found that bulimia was associated with self-reports of significant sexual abuse during childhood and/or adolescence. Furthermore, they reported that the severity of the abuse was related to the severity of the bulimia. As for their investigation of family environment, the authors’ results indicated that CSA and family variables combined in an additive manner to increase the risk of bulimia. However, the authors suggest that biological, personality, and contextual factors are still important factors in the etiology of bulimia as 38% of the normals reported CSA and/or dysfunctional family environments without reporting symptoms of bulimia.
Psychiatric Inpatient Samples
Zlotnick et al. (1996) sought to examine the relationship between eating pathology and sexual abuse by looking at the nature of this relationship in a clinical population. They hypothesized that patients with histories of sexual abuse would show increased severity of pathological eating behaviors compared to a control group that had not been sexually abused. Of the 134 female psychiatric patients that participated in the study, 92 reported a history of childhood sexual abuse (CSA), while 42 did not report such a history. Subjects were given a self-report questionnaire with behaviorally defined items describing sexual abuse as well as the Eating Disorder Inventory (EDI; Garner, Olmstead, & Polivy, 1983).
The authors suggest that their findings support an association between sexual abuse and an overall pattern of symptoms associated with eating disorders. Patients with histories of CSA had scores greater than or equal to mean scores for eating-disordered groups (obtained from Garner et al., 1985) on the EDI subscales on Ineffectiveness, Interpersonal Distrust, Interoceptive Awareness, and Body Dissatisfaction. Furthermore, subjects who had histories of CSA without having a diagnosis of an eating disorder still reported core features of both anorexia nervosa and bulimia nervosa. The authors contend that their results support and association between eating pathology and sexual abuse, with patients with histories of CSA being more likely to present with a greater degree of eating pathology than nonabused patients.
Using college undergraduates is a common and convenient method of sample selection. Such samples have several limitations which will later be discussed; however, results from such studies are nonetheless important in determining the nature of the relationship between childhood sexual abuse and the development of eating disorders. Miller et al. (1993) used a sample of 144 female undergraduates to determine possible relationships between CSA and adolescent onset of bulimia nervosa while also investigating family functioning in women diagnosed with bulimia. The authors’ primary hypothesis was that adolescents diagnosed as bulimic would report a higher incidence of CSA than nonbulimic adolescents. The following research protocols were administered to each subject: (1) Background information sheet on the subject and the subject’s family (2) Bulimic Investigatory Test, Edinburgh (BITE; Henderson & Freeman, 1987); (3) Sexual Life Events Questionnaire (SLEQ; Finkelhor, 1979); (4) Dissociative Experiences Scale (DES; Bernstein & Putnam, 1986); and (5) Childhood Family Mealtime Questionnaire (CFMQ; Miller et al., 1993).
The authors’ main hypothesis was tentatively supported with the bulimic group reporting more incidents of sexual abuse with a family member or relative after the age of 12 while also reporting more incidents of sexual abuse with an adult before the age of 12. The authors suggest that elevated sexual abuse rates among bulimic women may explain why bulimics tend to report negative feelings regarding their bodies. The authors also found that the bulimic who reported sexual abuse dissociate significantly more than nonabused bulimic women; thus, the authors further suggest that binge-purge cycles may serve to provide a form of psychological dissociation and/or emotional numbness. As in most studies, not all bulimics reported histories of sexual abuse and some nonbulimics did report sexual abuse; thus, it appears that there are other operating factors outside the experience of sexual abuse that ultimately lead to the development of an eating disorder. The authors also examined events that occurred during mealtimes and found that bulimia may be more likely to develop in women who have experienced a combined pattern of sexual abuse and food-related stress and control.
Using data from 130 undergraduate females on the Eating Attitudes Test (Garner & Garfinkel, 1979, 1982) and the Sexual Events Questionnaire (derived from Russell, 1983), Calam and Slade (1989) hypothesized that the reporting of unwanted sexual experience would be associated with higher scores on the measure of eating problems and that dieting, not bulimia, would be associated with reporting of sexual experience with a family member.
The authors found a pattern of cooccurrence of sexual experiences and high scores of eating disorder symptomatology after the age of 14. Dieting and bulimia were associated with sexual experience involving force; however, bulimia was not associated with sexual experience involving a family member. The authors suggest that dieting or self-starvation may form an overt way of regaining control with a family where sexual abuse had occurred, while compulsive eating or bulimia may occur in situations such as sexual abuse outside the family where the individual was not in a position to gain control over the perpetrator. Once again, sexual abuse should be considered as one of many possible triggers for the development of an eating disorder, with the context of the abuse likely playing a role in determining whether or not an eating disorder develops.
Wonderlich et al. (1996) used 38 women involved in an incest treatment program at a family service center to examine the hypothesis that victims of CSA would show greater levels than control subjects (27 subjects receiving treatment at the same family service center, but with no history of CSA) of each of the DSM-III-R diagnostic criteria for bulimia nervosa. Furthermore, the authors predicted that victims of CSA would present with higher levels of other tension-reducing behaviors (e.g., alcohol and cigarette use, self-mutilation, and suicidal gestures) than t control subjects. Their second hypothesis was that the magnitude of the emotional reaction to CSA is correlated with the development of eating disorders with the level of posttraumatic stress disorder symptomatology being correlated with the presence of bulimic behavior. Subjects were given the following questionnaires: (1) Eating Disorders Questionnaire (EDQ); (2) Eating Disorders Inventory (EDI); and (3) Response to Childhood Incest Questionnaire (RCIQ).
Consistent with their prediction, the authors found that the CSA subjects reported more bulimic behavior than the control subjects. Furthermore, their prediction that CSA subjects would display more tension-reducing behaviors was also supported, with incest victims displaying significantly more substance use, cigarette smoking, self-mutilation, and suicidal gestures than the control subjects. The authors suggest that these findings, along with the increased level of bulimic behavior, are consistent with the idea that CSA individuals engage in many self-destructive behaviors which potentially reduce emotional distress associated with their abuse.
Furthermore, the authors found that the presence of bulimic behavior was also associated with reported symptoms of posttraumatic stress disorder. They suggest that dieting, bingeing, and purging may be efforts to manage emotional states associated with memories of their abuse. The authors further suggest that eating disturbance is linked to the affective reactions, associated with the abuse. They conclude that their data provide support for the notion that bulimic behaviors are involved in affective regulation instead of being reactions to thoughts and memories of abuse.
Although many of the aforementioned studies found associations between sexual abuse and eating disorders (e.g., bulimia nervosa), there are many methodological problems associated with these research designs. Small sample sizes plague many of these studies; it is possible that some of the results found in these studies would have been smaller or larger in degree with larger samples. Furthermore, the samples varied greatly ranging from college students to individuals receiving treatment for clinical diagnoses of eating disorders or other psychiatric disturbances. These sample differences hinder generalizability as it is unknown if there is something unique about each of these groups which would cause the association between CSA and their disorders to be specific only to their particular group. For example, in college samples, even if CSA or eating disorders are present, these individuals may have higher levels of functioning than clinical samples.
Another problem associated with the studies on CSA and eating disorders is the use of many different definitions of sexual abuse and eating disorders. For example, some studies have used DSM-III-R criteria for determining subjects with eating disorders while others have used criteria that were not diagnostic in nature. Thus, it is difficult to determine whether the results found can be compared across studies as it is unknown whether sexual abuse and/or eating disorders are truly being measured.
The gathering of retrospective data is also a problem due to the potential bias in recall. Individuals are often selective in the information they recall about events in the past. Furthermore, most of the data are gathered through the use of questionnaires, some of which have not been evaluated to determine their reliability and validity. Similarly, in cases in which questionnaires are given by interviewers, these individuals may not be blind to the case status of the subjects, thus, creating a potential for biased data.
Finally, causal relationships between histories of sexual abuse and the development of eating disorders often cannot be determined based on the data gathered. There is often a lack of attention paid to the timing of the sexual abuse and the development of the eating disorder so that it is unknown whether there were pre-existing symptoms of eating disorders or if the sexual abuse preceded or exacerbated the development of the eating disorder. Furthermore, there is often little attention paid to other potentially influential factors such as family environment, stressors, and/or other mental disturbances. Such methodological problems do not cause the gathered data to be useless in examining the potential relationship between sexual abuse and eating disorders; however, these problems do hinder the ability to understand the full impact on sexual abuse (or lack thereof) on the development and/or maintenance of eating disorders.
In examining the relationship between sexual abuse and the development of eating disorders, the data seem to suggest that there is a potential relationship between sexual abuse and bulimia nervosa. The data presented here do not seem to indicate that sexual abuse has any relationship to anorexia nervosa. Furthermore, despite the fact that an association between sexual abuse and bulimia exists in some form, a causal connection has yet to be determined. It is likely that any traumatic event has the potential to lead to the development of bulimia nervosa given that the abuse occurs in an environment which perpetuates such development. This relationship is a complex one which is likely affected by both individual and family characteristics. There are individuals with histories of sexual abuse that never develop bulimia while, in turn, there are bulimics who have no history of sexual abuse. Clearly, there is more to this relationship than the two factors of abuse and bulimia alone.
In the future, it will be important to further analyze the effects of individual and family characteristics which may prove to exacerbate or moderate the development of bulimia (and possibly other psychological disturbances). Furthermore, it may also be of importance to look at the relationship between other forms of trauma (i.e., physical abuse) and the development of eating disorders. Unfortunately, until the methodological problems mentioned previously are resolved in future research designs, the same problems with assessing the data will be prevalent and the true nature of the relationship between sexual abuse and eating disorders will remain a mystery.
American Psychiatric Association. (1987). Diagnostic and Statistical Manual of Mental
Disorders. (Revised third edition). Washington, DC: Author.
Bernstein, E.M., & Putnam, F.W. (1986). Development, reliability, and validity of a
dissociation scale. Journal of Nervous and Mental Disease, 174, 727-735.
Bond, M., Gardner, S.T., Christian, J., & Sigal, J.J. (1983). Empirical study of self-rated
defense styles. Archives of General Psychiatry, 40, 333-338.
Calam, R.M., & Slade, P.D. (1989). Sexual Experience and Eating Problems in Female
Undergraduates. International Journal of Eating Disorders, 8, 391-397.
Finkelhor, D. (1979). Sexually victimized children. New York: Macmillan.
Garner, D.M., & Garfinkel, P.E. (1979). The Eating Attitudes Test: An index of the
symptoms of anorexia nervosa. Psychological Medicine, 9, 273-279.
Garner, D.M., Garfinkel, P.E., & O’Shaughnessy, M. (1985). The validity of the
distinction between bulimia with and without anorexia nervosa. American Journal of
Psychiatry, 142, 581-587.
Garner, D.M., Olmstead, M.A., & 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.
Hastings, T., & Kern, J.M. (1994). Relationshipd Berween Bulimia, Childhood Sexual
Abuse, and Family Environment. International Journal of Eating Disorders, 15, 103-
Henderson, M., & Freeman, C.P.L. (1987). A self-rating scale for bulimia: The "BITE."
British Journal of Psychiatry, 150, 18-24.
Miller, D.A.F., & McCluskey-Fawcett, K. (1993). The Relationship Between Childhood
Sexual Abuse and Subsequent Onset of Bulimia Nervosa. Child Abuse & Neglect, 17,
Moos, R., & Moos, B. (1984). Family environment scale. Palo Alto, CA: Consulting
Steiger, H., & Zankor, M. (1990). Sexual Traumata Among Eating-Disordered,
Psychiatric, and Normal Female Groups: Comparison of Prevalences and Defense
Styles. Journal of Interpersonal Violence, 5, 74-86.
Thelan, M.H., Farmer, J., Wonderlich, S., & Smith, M. (1991). A revision of the Bulimia
Test: The BULIT-R. Psychological Assessment: A Journal of Consulting and Clinical
Psychology, 3, 119-124.
Walters, C., Smolak, L., & Sullins, E. (1987). Parent-child interactions and severity of
child sexual abuse. Paper presented to the annual convention of the Society for
Research in Child Development, Baltimore.
Welch, S.L., & Fairburn, C.G. (1994). Sexual Abuse and Bulimia Nervosa: Three
integrated case control comparisons. American Journal of Psychiatry, 151, 402-407.
Wonderlich, S., Donaldson, M.A., Carson, D.K., Staton, D., Gertz, L., Leach, L.R., &
Johnson, M. (1996). Journal of Interpersonal Violence, 11, 195-207.
Zlotnick, C., Hohlstein, L.A., Shea, M.T., Pearlstein, T., Recupero, P., & Bidadi, K.
(1996). The Relationship Between Sexual Abuse and Eating Pathology. International
Journal of Eating Disorders, 20, 129-134.
Psychology DepartmentThe Health Psychology Home Page is produced and maintained by David Schlundt, PhD.
Vanderbilt Homepage | Introduction to Vanderbilt | Admissions | Colleges & Schools | Research Centers | News & Media Information | People at Vanderbilt | Libraries | Administrative Departments | Medical
|Return to the Health Psychology Home Page|
|Send E-mail comments or questions to Dr. Schlundt|