|Home | About | Journals | Submit | Contact Us | Français|
To examine correlates of childhood maltreatment in women with binge-eating disorder (BED).
Semistructured interviews evaluated 137 women with BED for psychiatric disorders and eating psychopathology, and self-reported childhood maltreatment was assessed.
Emotional abuse was reported by 52% of participants, physical abuse by 28%, sexual abuse by 31%, emotional neglect by 66%, and physical neglect by 48%. Maltreatment categories were not associated with most lifetime psychiatric diagnoses, although specific associations were observed for dysthymic disorder, posttraumatic stress disorder, and alcohol use disorders. Few associations were noted with eating pathology, but most forms of childhood maltreatment were negatively associated with self-esteem.
Women with BED report rates of childhood maltreatment comparable to those for clinical groups, and much higher than community samples. Although prevalent in women with BED, childhood maltreatment is not generally associated with variability in eating pathology or with psychiatric comorbidity, but is associated with lower self-esteem.
Individuals with eating disorders frequently report histories of childhood abuse or neglect, and this has been especially observable in community and clinical samples of women.1–8 The significance of these reports, however, has remained uncertain. Most such studies have focused on anorexia nervosa and bulimia nervosa—with some investigators suggesting that childhood maltreatment may be prevalent among individuals with eating disorders and with disordered eating in general,4,7 others concluding that childhood maltreatment is associated mostly with bulimia nervosa and with binge eating,2,5,8 and others still concluding that childhood maltreatment occurs no more frequently among women with eating disorders than among women with other forms of psychiatric illness.1,3,6
Relatively fewer data exist with regard to childhood abuse and neglect specifically in binge-eating disorder (BED). Yanovski and colleagues9 reported that rates of sexual abuse in a self-referred group of obese subjects did not differ between those with and without BED. In a community-based, case-control study of women, Fairburn and colleagues10 found that risk factors for BED included a wide range of childhood experiences and environmental conditions reflecting various forms of abuse and neglect. They found rates of physical and sexual abuse in subjects with BED—21% and 29%, respectively—that exceeded those in healthy controls but which did not differ significantly from those in their general psychiatric controls. Grilo and Masheb11 examined childhood maltreatment in a group of men and women with BED and found that 59% reported emotional abuse, 36% reported physical abuse, 30% reported sexual abuse, 69% reported emotional neglect, and 49% reported physical neglect. These investigators reported that most specific forms of childhood maltreatment were unrelated to eating pathology—although, among women, they did find that physical neglect was associated with dietary restraint. They also reported that emotional abuse was associated with greater body dissatisfaction, more depression, and lower self-esteem. Grilo and Masheb11 concluded that childhood maltreatment may be associated with increased psychological distress in general, but not specifically with weight or eating disorder symptomatology.
Indeed, against the backdrop of observations that other psychiatric conditions—especially mood, anxiety, and substance use disorders—may have significant diagnostic overlap with BED,9,12,13 some have suggested that these comorbidities tend to confound the relationships between childhood maltreatment and the frequency or clinical characteristics of eating disorders.1 In a study of women with anorexia nervosa and bulimia nervosa, Herzog and colleagues3 noted that subjects with more psychiatric comorbidity also reported more childhood sexual abuse. Richardson and colleagues14 studied childhood abuse among women with bulimia-spectrum disorders, and found that a “high-comorbidity” group—characterized by co-occurrence of major depressive disorder, anxiety disorders, and substance use disorders—reported more childhood physical and sexual abuse than their “low-comorbidity” group. On the other hand, the study by Yanovski and colleagues9—which observed no differences in sexual abuse rates between obese subjects with and without BED—found higher levels of psychiatric comorbidity in the former group compared to the latter. It is worth noting, however, that this study included both male and female patients. Along this line, MacMillan and colleagues15 utilized a large community sample to demonstrate that, while childhood physical and sexual abuse is associated with lifetime psychopathology, this association is stronger for women than it is for men.
The aim of this study was to explore further the associations between childhood maltreatment and lifetime psychiatric comorbidity in women with BED. We sought to examine the broad spectrum of childhood abuse and neglect as well as a broad range of axis I disorders. Also, in order to bring these relationships into better focus, we concurrently examined the associations between childhood maltreatment and measures of eating pathology and associated psychological functioning.
Participants were a nearly-consecutive series of 137 treatment-seeking women who met strict DSM-IV16 research criteria for BED and the additional inclusion/exclusion criteria noted below. They ranged in age from 20 to 59 years (M = 43.9, SD = 9.1). A majority of participants (85%) were Caucasian, and most (87%) had either attended or graduated from college.
Participants responded to media advertisements soliciting individuals with concerns about binge eating and weight, for participation in treatment studies within an urban medical school setting. In order to be included in the study, participants had to be between 18 and 60 years of age, and had to meet full research diagnostic criteria for BED. Individuals were excluded if they were receiving ongoing professional treatment for eating or weight problems—or if they had certain medical conditions that may influence eating or weight (e.g., diabetes, thyroid disease), or had a severe psychiatric illness that could interfere with the assessment process (e.g., psychosis, bipolar disorder). Assessments were administered by trained doctoral-level research clinicians, who were monitored to maintain reliability over time. Full IRB review and approval were obtained. After complete explanation of the study procedures, written informed consent was obtained from all participants.
The Structured Clinical Interview for DSM-IV Axis I Disorders – Patient Edition (SCID-I/P)17 was administered to assess psychiatric disorders, including BED. Kappa coefficients for interrater reliability of psychiatric diagnoses ranged from 0.57 to 1.0; kappa for current BED diagnosis was 1.0. Final research diagnoses were established by the “best estimate” method,18 based on the structured interviews and on any additional relevant clinical data.
The Eating Disorder Examination (EDE)19 was administered to assess the attitudinal and behavioral features of eating disorder psychopathology, and to confirm the BED diagnosis. The EDE is a semistructured interview that assesses the core and associated psychopathology of eating disorders. This instrument focuses on the preceding 28 days, with the exception of diagnostic items for which DSM-IV stipulates specific time-frames or duration criteria. Included in the EDE interview is a module that assesses the DSM-IV research criteria for BED. The EDE assesses the frequency of different forms of overeating, including objective overeating episodes—defined as eating an unusually large quantity of food while experiencing subjective loss of control—which corresponds to the DSM-IV definition of a binge eating episode. An EDE global score incorporates subscale scores reflecting attempts at dietary restraint and concerns about eating, shape, and weight. The EDE is a well-established method for assessing eating disorder psychopathology,20,21 and has demonstrated good interrater and test-retest reliability in diverse groups, including BED.22,23
Height and weight were measured during the evaluation process, and body mass index (BMI) was calculated as weight (in kilograms) divided by the square of height (in meters). Structured clinical interviews inquired about obesity-related historical variables, including age at obesity onset, age at dieting onset, and age at binge eating onset.
The Childhood Trauma Questionnaire (CTQ), 28-item version,24 a self-report instrument, was used to assess childhood maltreatment in five areas: emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect. Participants rate statements about childhood experiences (defined as prior to age 18) on five-point Likert-type scales (“never true” to “very often true”). Most items are phrased in objective terms (e.g., “When I was growing up, some one touched me in a sexual way or made me touch them”), although some items require subjective evaluation (e.g., “When I was growing up, I believe I was sexually abused”). Reliability and validity of the CTQ, including stability over time, discriminant and convergent validity—with structured trauma interviews and with corroborating independent data—have been documented.24–26 The CTQ can be used to produce both dimensional (higher scores reflect greater maltreatment) and categorical results for each form of maltreatment. For our categorical analyses, we used previously established cut-points.24,26
The Beck Depression Inventory (BDI), 21-item version,27 is a well-established and widely used self-report measure of the symptoms of depression—and, more generally, of negative affect (Beck et al., 1988).28 Higher scores reflect greater depressive features.
Phi coefficient, an effect size measure for contingency tables, was used to examine the strengths of associations between the presence of psychiatric disorders and categorical reported childhood maltreatment for each maltreatment domain. These associations were examined for all psychiatric disorder groups that were sufficiently frequent to permit statistical analysis. When a statistically significant association was found between a psychiatric disorder category and at least one of the five maltreatment domains, then specific disorders within that category were examined further. It was intended that posttraumatic stress disorder would be examined—irrespective of whether the anxiety disorder category showed significant associations with one or more maltreatment areas—since this disorder is, by definition, associated with traumatic exposure. Pearson’s product-moment correlation analyses, examined the associations between reported childhood maltreatment scores and the continuous variables describing eating pathology (age at obesity onset, age at dieting onset, age at binge eating onset, current BMI, binge frequency, and EDE global score) and psychological functioning (BDI and RSES).
Table 1 provides categorical frequencies of, and dimensional scores for, all five types of reported childhood maltreatment. The most frequent of these types was emotional neglect, being reported by two-thirds of participants; both emotional abuse and physical neglect were reported by approximately half of participants; sexual abuse and physical abuse were each reported by slightly less than one-third of all participants.
Table 2 summarizes the lifetime frequencies of psychiatric disorder categories and specific psychiatric disorders—and their correlations with categorical reported childhood maltreatment. Among the psychiatric disorder categories, only the depressive disorders, anxiety disorders, and substance use disorders occurred at frequencies sufficiently large to permit statistical analysis. As correlation analysis demonstrated a statistically significant association between the lifetime presence of depressive disorders and categorical reported physical abuse, specific depressive disorders were examined further. No significant associations were observed between the presence of major depressive disorder and any of the maltreatment types. However, a significant association was observed between the presence of dysthymic disorder and emotional abuse. Despite the absence of significant associations between anxiety disorders and any of the maltreatment types, associations with posttraumatic stress disorder were examined further (for the reason given above). The lifetime presence of posttraumatic stress disorder was significantly associated with both sexual abuse and physical neglect. Finally, the presence of substance use disorders was significantly associated with physical abuse; further analysis demonstrated a similar association between alcohol use disorders and physical abuse, but no significant associations between drug use disorders and any of the maltreatment types.
Table 3 summarizes the mean values for the eating pathology and psychological functioning variables, and their correlations with reported childhood maltreatment scores. As is evident from the table, relatively few significant associations were observed between the eating pathology variables and the maltreatment types; the only statistically significant associations among these were age at obesity onset having an association with physical abuse, and age at dieting onset having a negative association with emotional neglect. While no significant associations were observed between BDI score and any of the maltreatment types, statistically significant negative associations were observed between RSES score and emotional abuse, sexual abuse, emotional neglect, and physical neglect.
Our study group of treatment-seeking obese women with BED reported rates of childhood maltreatment that were roughly two to three times higher than those reported by a normative sample of adult women using the same self-report instrument.26 In general, maltreatment rates for our study group were much higher than those reported for healthy women and were comparable to those for other clinical groups of women—including bariatric surgery candidates, eating disorder patients, and general psychiatric patients.10,11,30 Although prevalent in women with BED, we found that childhood maltreatment showed very few statistically significant associations with variability in psychiatric comorbidity or eating pathology. Collectively, these findings are consistent with the view that childhood maltreatment may represent a general risk factor for broad lifetime psychopathology and distress or maladjustment, which may include BED, but does not have specific associations with variability in eating pathology.10,15 Keeping this overall context in mind, we will comment cautiously on specific findings with reference to the empirical literature.
We observed relatively few associations between maltreatment types and lifetime psychiatric disorders. We found that depressive disorders as a group were associated with reported childhood physical abuse. While generally congruent with findings from the study by MacMillan and colleagues15 that childhood physical abuse is associated with major depression and drug use disorders in women, this finding is also consistent with observations from community studies of obesity in women, as well as studies of bariatric surgery candidates and patients with bulimia nervosa, suggesting an association between physical abuse and depression.30–32 Although we did not, in our study group, observe a specific association between major depressive disorder and any of the maltreatment categories, we did find an association between dysthymic disorder and emotional abuse. Similar associations between depression and emotional abuse have been observed in bariatric surgery candidates and patients with BED, and in studies of bulimic symptoms among undergraduate women.11,30,33 Studies of verbal abuse—which can be viewed as one aspect of emotional abuse—suggest that such maltreatment may have deeper effects than other types of abuse because children tend to identify with their parents and to internalize verbally abusive statements throughout their lives.34,35 In this light, it is not unexpected that emotional abuse would have a specific association with dysthymic disorder—which is marked by feelings of low self-esteem and inferiority, and which frequently has its onset early in life.
Similarly, it was not unexpected that, although we found no specific associations between anxiety disorders as a group and any of the maltreatment categories, we did find significant associations between posttraumatic stress disorder and both sexual abuse and physical neglect. Maltreatment is, by definition, associated with posttraumatic stress disorder—and the literature on eating disorders and obesity has documented these relationships, as well, especially with respect to childhood sexual abuse in women.36–38
We also found an association between physical abuse and substance use disorders. Although there were no significant associations between drug use disorders and any of the maltreatment categories, we did observe an additional association between physical abuse and alcohol use disorders. These findings are generally consistent with the results of studies suggesting relationships between childhood maltreatment and substance use disorders among community and clinical samples of women.14,39,40 Associations of physical abuse in particular with substance use disorders, and with alcohol use disorders specifically, have been observed among women in the community, as well as among women with bulimia nervosa.15,31
Very few associations were observed between the maltreatment categories and the eating pathology variables. Our results are, therefore, consistent with those of other studies suggesting that childhood maltreatment does not have specific associations with disordered eating behavior.6,10,11,30 We did observe an association between emotional neglect and earlier age at dieting onset, which is generally consistent with an earlier report that growing up under such adverse conditions may lead to extreme dieting behaviors perhaps as a coping method.41 We also observed an association between physical abuse and later age at obesity onset. This interesting finding is consistent with several recent reports linking childhood abuse—and physical abuse, in particular—with obesity during middle age.32,42–44
With respect to our measures of psychological functioning, we observed no associations between BDI score and any of the maltreatment categories. Given that we found some associations with respect to lifetime rates of syndromal depression, this might appear surprising. We note, however, that the BDI assesses a broad range of current depressive and negative affect (dimensionally), whereas the SCID-I/P findings are for lifetime occurrences of mood disorders. In addition, we note that the absence of specific associations with the BDI may simply be due to a “restricted range”—that is, moderately high BDI scores characterized our treatment-seeking study group. We did, however, observe negative associations between several childhood maltreatment categories and RSES score. These findings are generally consistent with results from community and clinical settings suggesting that low self-esteem (i.e., a broad, general form of maladjustment) may be associated with a range of childhood maltreatment types, including emotional neglect, emotional abuse, and sexual abuse.1,11,30,34,37
We note several strengths and limitations of our study as a context for interpreting these results. Strengths include the use of diagnostic and structured interviews, for evaluating psychiatric disorders and eating pathology, that were reliably administered by trained and experienced doctoral-level research clinicians. With respect to study limitations, our findings pertain to women meeting strict research criteria for BED, and to those responding to advertisements seeking participants for treatment studies at a specialty research program in a university-based medical center. Therefore, our results may not generalize to those with sub-threshold binge eating, to community samples, to non-treatment-seekers, or to those who might be uninterested in research participation. Thus, it is possible that our paucity of significant associations might reflect a “restricted range” (i.e., all subjects met full BED criteria with high levels of eating and depressive pathology)—and that different patterns of association might characterize groups comprised of individuals who did not seek treatment, and who therefore may be less distressed, or who sought treatment at general treatment clinics rather than at a university-based specialty program. Due to our exclusion criteria, these results may not generalize to BED patients with co-occurring psychosis or bipolar disorder, although such exclusions were extremely rare. Similarly, our results may not generalize to those with certain co-occurring medical illnesses, such as diabetes or thyroid disease. Also, our results pertain to a broad range of women with BED, but may not pertain to men, or to those under 18 or over 60 years of age. Additionally, we relied on a self-report instrument for assessing childhood maltreatment. Although we acknowledge the limitations inherent in self-report methods for assessing maltreatment, the measure used is well validated,26 and such approaches may remove interpersonal concerns or barriers to disclosing sensitive material.45 Finally, our assessments were cross-sectional, and cannot discern causal or etiologic relationships. Longitudinal studies are needed to understand better the relationships between childhood adversity, disordered eating, and general psychopathology during adulthood.
With these strengths and limitations in mind, as context, some conclusions can be drawn from the data. We found rates of reported childhood maltreatment in our group of women with BED that were similar to those in other clinical groups, but well above those in non-clinical samples. Nonetheless, despite the large number of analyses conducted, there were very few statistically significant associations between maltreatment domains and specific axis I disorders—and even fewer associations between maltreatment domains and indicators of disordered eating. These observations, together with previous studies, support the view that, while childhood maltreatment may represent a general risk factor for broad lifetime psychopathology and maladjustment (e.g., lower self-esteem),1,11,30,34,37 it does not have specific associations with eating pathology.6,10,15
Although there is little evidence that childhood maltreatment specifically results in eating pathology or in BED, future studies should examine further the possibility that, among women with BED, clinical course and the efficacy of treatment may be affected by factors more directly related to maltreatment experiences during childhood. In a study of women undergoing treatment for bulimia nervosa, for example, one group of investigators found that childhood trauma has a dose-effect relationship with drop-out from psychotherapy.46 Others have noted associations—again, in bulimic and bulimia-spectrum patients—between childhood abuse and response to pharmacologic interventions.14,31 It remains unknown, however, whether histories of maltreatment represent a negative prognostic indicator for evidence-based psychological or pharmacological interventions in patients with BED.47,48
This research was supported in part by grants from the National Institutes of Health (R01 DK49587 and K24 DK070052). No additional funding was received for the completion of this work.
This paper was presented in part at the 163rd annual meeting of the American Psychiatric Association, New Orleans, Louisiana; May 22–26, 2010.
Biomedical Support Disclosures: The authors report no commercial or biomedical industry support or conflicts of interest.