INTRODUCTION

In recent years, there have been an increasing number of reports concerning serious traumatic experiences, especially physical and sexual abuse in general psychiatric patients1 and in anorexia nervosa and bulimia nervosa patients2–5 in Western countries. Results of recent studies suggest that the link between abuse histories and eating disorder is not specific nor straightforward, but such histories are risk factors for bulimia nervosa with significant comorbidity.6–10

Although anorexia nervosa and bulimia nervosa have been considered a prototype for investigations of cross‐cultural differences and psychiatric diagnoses,11 only Pope et al. compared abuse rates reported by normal‐weight bulimics among countries (USA, Austria, and Brazil), but failed to find significant differences.12 Moreover, these rates did not appear to be any greater than those reported in comparable studies of women in the general population. However, American and Brazilian society mainly consists of immigrants from European countries, so that these societies have relatively similar cultures. Conversely, Japan is a highly industrialized country, but has a very different culture from that of Western countries. Few Japanese patients with eating disorders have reported traumatic experiences during their childhood,13 although their clinical features show a similar degree of severity compared with patients in Western countries with regard to prevalence14 and stress‐coping strategies.15 Furthermore, anecdotal evidence also suggests that Japanese male patients with alcoholism frequently report being physically abused in childhood. Therefore, a cross‐cultural study, especially in very different cultures, is essential to improve understanding of eating disorder psychopathology and etiology.

This study compares the prevalence of sexual and physical abuse histories in Japanese and USA patients with eating disorders to examine the hypothesis that Japanese patients were less frequently sexually abused, but might have been physically abused more frequently.

METHOD

RESULTS

Table 1 shows the demographics of the subjects. Data were analyzed by two‐way ANOVA. Patients in Japan and the USA were similar in age, however, patients from the USA had significantly earlier onset and had been ill for a significantly longer period compared with Japanese patients. Conversely, Japanese patients showed a significantly lower BMI, however, bulimic patients in both countries showed a similar BMI. Among Japanese patients, 15 AN‐R, 18 AN‐BP and 18 BN patients had a history of inpatient treatment due to eating disorders or were admitted our unit.

All Japanese patients were outpatients and all patients from the USA were inpatients, thus we compared demographics, EDI scores, physical and sexual abuse history between Japanese patients with and without inpatient treatment (including our unit) history. However, there are no significant differences in sexual or physical abuse history in any diagnostic subgroup such as AN‐R, AN‐BP or BN, or in the total Japanese patient group. For example, physical punishment history (as defined below) in Japanese total eating disorder patients, with and without inpatient treatment history were 10 (20%) versus 14 (13%) (P = 0.34 by Fisher’s exact test). Victims of rape among these Japanese patients, with and without inpatient treatment history were 0 versus 4 (4%), respectively (P = 0.31 by Fisher’s exact test). However, among demographics values, Japanese AN‐BP patients with an inpatient treatment history were significantly older (26.9 ± 5.2 vs 23.5 ± 3.9, t = 2.5 P = 0.015), had a longer duration of illness (8.7 ± 4.5 vs 4.1 ± 3.4, t = 4.0 P < 0.001), lower minimal BMI after menarche (12.0 ± 1.9 vs 13.7 ± 2.1, t = 2.7, P = 0.01), and showed significantly lower drive for thinness (7.7 ± 5.8 vs 12.3 ± 6.1, t = 2.2, P = 0.03) and bulimia (4.5 ± 4.8 vs 10.5 ± 6.7, t = 3.0, P = 0.006) scores on EDI compared to those without such history. Therefore, Japanese patients with and without inpatients treatment history were combined into one group in analyzing abuse histories.

Results of sexual abuse histories are summarized in Table 2. Some Japanese patients and controls reported histories of confronting exhibitionism, although no patient from the USA reported this. Likewise, Japanese AN‐R patients reported minor sexual abuse history (sexual abuse history other than rape) significantly frequently compared with the AN‐R patients from the USA, and similarly Japanese AN‐BP and BN patients. However, there were no AN‐R patients in either country who had been raped or victimized by incest. Similar substantial numbers of patients with AN‐BP and BN in both countries were sexually abused, although two major differences between the countries were identified. First, most of the perpetrators of Japanese patients were strangers, while in the patients from the USA, perpetrators were known to some degree. This difference between the two countries reached significance in the AN‐R and AN‐BP groups. Second, victims of rape or incest, including attempts, were significantly more frequent among AN‐BP and BN patients from the USA than among their Japanese counterparts.

Histories of physical punishment are presented in Table 3. Only responses of ‘often’ were counted as a history of abuse in Table 3. Similar to results of sexual abuse histories, none of the Japanese or American patients with AN‐R were physically punished by their parents in their childhood. Japanese patients with AN‐R seemed less likely to be physically punished, even compared with controls, although there were no controls from the USA for statistical comparison. There were no significantly differences between USA and Japanese patients for any item or in any diagnostic subgroup (P = 0.39–1.00 by Fisher’s exact test). We provisionally defined physical punishment histories as any of the following: often locked in a closet, often hit with the hands, often kicked, or often hit with something other than the hands. Surprisingly, a very similar percentage of AN‐BP and BN patients from both countries reported to have had a physical punishment history. AN‐BP and BN groups in Japan and the AN‐BP group in the USA experienced significantly more frequently these physical punishment histories than AN‐R patients.

DISCUSSION

To our knowledge, this is the first study to directly compare the sexual and physical abuse histories in eating disorder patients of Western and Asian countries.

Although methodologies varied considerably, the study demonstrated a childhood sexual abuse rate of approximately 30% in eating disordered individuals in clinical settings.3, 23 Therefore, the results from both countries are consistent with previously reported percentages of patients with such sexual abuse histories. However, there are two major differences between two countries. First, is the high rate of victimization by rape or incest among the AN‐BP and BN patients from the USA. Second, is the significantly higher rate of minor sexual abuse by strangers among Japanese patients, especially in Japanese patients with AN‐R and AN‐BP, compared to the rate among patients from the USA. However, this issue has been discussed elsewhere as a ‘Chikan’ (a Japanese word meaning a person who commits minor sexual crimes such as among rush‐hour subway users).13 In addition, Japanese people might be more shameful of sexual events and difficulties, although our results did not agree with Schmidt’s hypothesis24 that patients with anorexia nervosa had significantly more pudicity events before onset than BN patients or controls.

Rates of rape or incest in AN‐BP or BN patients from the USA were significantly higher than those among their Japanese counterparts (around fivefold higher). There were no controls from the USA in the present study, however, we used questions similar to Finkelhor et al.‘s telephone interview survey,25 and victimization by rape or incest (including attempts) has been reported to be 14.6% among the general population of the USA, and this rate is consistent with the results of a recent study.26 Therefore, victimization involving unwanted sexual intercourse among the general population of the USA is obviously higher than among the Japanese control group in this study (only 1%), although the sample size of the Japanese controls was too small and sampling was biased (all nursing school students) to represent the Japanese general population. Victimization by rape or incest among AN‐BP or BN patients from the USA appears to be comparable with the rate in the general American population.

These discrepancies between two countries cannot be explained by the high criminal rate in the USA because all perpetrators in the USA were known to the person. In addition, the majority of perpetrators were strangers to the Japanese patients. These discrepancies may be explained by cultural attitudes toward sexual matters and the high‐divorce rate in USA.27 Our results clearly show that sexual abuse history is not a necessary and sufficient causal variable for eating disorders, although these events might be related to refractory and chronicity of illness.

Conversely, rates of physical punishment history in every diagnostic subgroup were very similar between the two countries. The impact of a physically abusive childhood is profound. These include psychopathology such as post‐traumatic stress disorder, depression, and alcoholism;28–31 and the intergenerational transmission of impaired parenting.32

Few studies have examined the relationship between eating disorders and a physical abuse history. Rorty et al. found that women with bulimia nervosa reported significantly more physical punishment than did controls.33 McCarthy et al. reported that dissociation appeared to be linked to physical rather than sexual abuse in bulimic patients.34 However, Folsom et al. reported that there were no significant differences in the rates of physical and sexual abuse between eating disordered patients and the general psychiatric group.7 We found a significantly higher prevalence of physical punishment in bulimic anorexic and bulimic patients than in the restricting anorexic sample in both countries. Although there were no controls from the USA, AN‐BP and BN patients tended to have been physically punished more frequently than controls in Japan. Whether childhood sexual abuse is more linked to bulimic than restricting anorexics currently remains controversial (among controlled studies, two support35, 36 and two do not37, 38). However, there are no available data on whether childhood physical abuse is more linked to bulimic than restricting anorexics. Our study is the first one to show that physical abuse history tends to be linked with bulimic behaviors even in very different cultures.

However, it would be inappropriate to conclude that physical abuse as a traumatic event causes the bulimia. Because physical forms of child abuse are generally intrafamilial, it may be argued that many of the apparent associations between childhood physical abuse and later adjustment reflect the social and family context within which the abuse occurs rather than the direct traumatic effects of abuse on individual adjustment.39 Even sexual abuse is not randomly distributed through the community but is more likely to be found in disrupted and disturbed families and in those families subject to economic and social disadvantage.40 Therefore, it seems to be natural for physical abuse to be clearly linked with family disadvantages. Indeed, in the bulimia nervosa group, increased levels of physical punishment were associated with greater global family pathology.33 The elevated rates of adjustment problems in physically abused subjects may have been largely or wholly due to the social environment and context in which physical punishment/maltreatment occurred.41 Even in terms of sexual abuse, there were data that adverse family background may be a more important etiological factor for bulimia nervosa than the sexual abuse itself.42 Therefore, physical abuse may only be an indicator of family background. Moreover, physical abuse may be related to a factor that decides whether eating disorder patients are restricting or bulimic, rather than a causal factor of eating disorders.

There were some limitations in the present study. First, the relatively small number of subjects from the USA, especially normal‐weight bulimics and absence of American control subjects were serious shortcomings of this study. Most interesting is how prevalent abuse and eating disorders are in the general populations of the two countries. Further study needs to clarify this point.

Second, inpatients usually have a more severe form of illness than outpatients in regard to some clinical symptoms, and all the patients from the USA in our study were inpatients. Chronicity of the subjects from the USA was probably due to this difference, because Japanese AN‐BP patients with inpatient treatment history showed a degree of chronicity similar to that of patients from the USA. However, there were no significant differences in physical or sexual abuse history between Japanese in‐ and outpatients. Similarly, Favaro et al. found that the prevalence of sexual and physical abuse did not differ significantly between the in‐ and outpatients.43

Third, the degree to which a self‐reporting questionnaires can accurately identify childhood histories of abuse remains unclear. Many factors could potentially lower the rate of reported abuse (e.g. the degree of honesty, dissociation or repression of traumatic memory). Many investigators believe patients commonly do not report histories of abuse, especially sexual abuse, until they are well into their psychotherapy. However, Waller found no difference between clinical interview and self‐reports in eliciting a history of sexual abuse in eating disordered subjects,36 similar to our results in a small number of inpatients. The number of patients admitting to stealing in self‐reporting questionnaire also did not change after 2 months of intensive treatment.44 In addition, other groups in Japan45 reported that there were no differences in reporting sexual abuse history for female eating disorder patients whether the interviewer was male or female.

Taken as a whole, differences in sexual abuse histories suggest that these two cultures and two societies are very different. Despite these differences, we found a very similar physical punishment history between patients in these countries. It would be oversimplifying to conclude that physical abuse is the essential trauma causing bulimic behaviors because we found that only around one‐fifth of the bulimic patients had such childhood histories, even though we included mild forms of physical punishment. Therefore, physical abuse might be related to whether patients are restricting or bulimic, even in different cultures, rather than a risk factor for developing eating disorders. Further cross‐cultural study, including family environment, childhood disruptive behaviors and temperament using general population sample, will help us better understand etiology and psychopathology of eating disorders.