Reading Journal

Are Eating Disorders Culture-Bound Syndromes? Implications for Conceptualizing Their Etiology

Pamela K. Keel Harvard University

Kelly L. Klump Michigan State University

The authors explore the extent to which eating disorders, specifically anorexia nervosa (AN) and bulimia nervosa (BN), represent culture-bound syndromes and discuss implications for conceptualizing the role genes play in their etiology. The examination is divided into 3 sections: a quantitative meta-analysis of changes in incidence rates since the formal recognition of AN and BN, a qualitative summary of historical evidence of eating disorders before their formal recognition, and an evaluation of the presence of these disorders in non-Western cultures. Findings suggest that BN is a culture-bound syndrome and AN is not. Thus, heritability estimates for BN may show greater variability cross-culturally than heritability estimates for AN, and the genetic bases of these disorders may be associated with differential pathoplasticity.

The influence of culture on the development of eating disorders such as anorexia nervosa (AN) and bulimia nervosa (BN) has been long appreciated. These syndromes are more prevalent in indus- trialized and often Western cultures and are far more common among females than males, mirroring cross-cultural differences in the importance of thinness for women (Miller & Pumariega, 2001; Pate, Pumariega, Hester, & Garner, 1992). Furthermore, eating disorders seem to have become more common among younger females during the latter half of the twentieth century, during a period when icons of American beauty (Miss America contestants and Playboy centerfolds) have become thinner and women’s mag- azines have published significantly more articles on methods for weight loss (Garner, Garfinkel, Schwartz, & Thompson, 1980; Owen & Laurel-Seller, 2000; Rubinstein & Caballero, 2000; Wise- man, Gray, Mosimann, & Ahrens, 1992). Patients with eating disorders often describe intense preoccupation with weight and dieting as the first stages of their illness, and the American Psy- chiatric Association’s (2000) diagnostic criteria for AN and BN require body image disturbance. The cognitive– behavioral theory of eating disorders places weight concerns and dieting at the center of their etiology (Fairburn, Cooper, & Shafran, 2002). These patterns indict current cultural beauty ideals in the etiology and maintenance of eating disorders.

Yet, a recent trend in eating disorders research has been to focus on their genetic bases and deemphasize the role of culture (cf.

DeAngelis, 2002). Briefly, data from community-based twin stud- ies have suggested that the heritability of eating disorders is greater than 50% (Bulik, Sullivan, & Kendler, 1998; Bulik, Sullivan, Wade, & Kendler, 2000; Kendler et al., 1991, 1995; Klump, Miller, Keel, McGue, & Iacono, 2001; Kortegaard, Hoerder, Joergensen, Gillberg, & Kyvik, 2001; Wade, Bulik, Neale, & Kendler, 2000), and several recent reports have emerged indicating specific genetic loci for susceptibility to eating disorders (Grice et al., 2002; Koronyo-Hamaoui et al., 2002; Ricca et al., 2002; Westberg et al., 2002). Indeed, one eating disorders researcher has equated the shift from psychosocial to biological theories of eti- ology in eating disorders to the shift that occurred within schizo- phrenia research (DeAngelis, 2002). However, unlike schizophre- nia, eating disorders have been characterized as culture-bound syndromes.

Prince (1985) defined a culture-bound syndrome as “a collection of signs and symptoms (excluding notions of cause) which is restricted to a limited number of cultures primarily by reason of certain of their psychosocial features” (p. 201). With this defini- tion, he proposed that AN might represent a culture-bound syn- drome. However, he cautioned that “the decision hinges on the empirical question of whether or not the syndrome occurs in non-Western cultures or segments of them which are not markedly influenced by Western cultures” (Prince, 1985, p. 201). In contrast, Swartz (1985) argued that evidence of AN in non-Western cultures would not decrease the extent to which it represents a culture- bound syndrome because AN cannot be understood separated from its cultural context:

The identical symptoms now may mean different things from what they may have meant in the late 19th century when the disorder was first documented in the form we now recognize. This implies that it may be valid to relate symptoms now occurring to social factors which may not have existed in the same form years ago. (p. 727)

Thus, the same disorder could be attributed to different social factors (the aspect of the syndrome that is culture bound) despite having existed in different cultural contexts. Swartz’s distinction between disorder and culturally embedded syndrome resembles

Pamela K. Keel, Department of Psychology, Harvard University; Kelly L. Klump, Department of Psychology, Michigan State University.

We acknowledge support from a Radcliffe Institute for Advanced Study Fellowship to Pamela K. Keel and from National Institute of Mental Health Grants R01 61836 and R01 63758 to Pamela K. Keel and R03 65447 to Kelly L. Klump. We also wish to thank Mark Baxter, Kyle Gobrogge, Richard McNally, Natalie Lester, Alyson Zalta, and the 2001–2002 Rad- cliffe Institute Fellows for their contributions in the preparation of this article.

Correspondence concerning this article should be addressed to Pamela K. Keel, who is now at the Department of Psychology, University of Iowa, Iowa City, Iowa 52242. E-mail: [email protected]

Psychological Bulletin Copyright 2003 by the American Psychological Association, Inc. 2003, Vol. 129, No. 5, 747–769 0033-2909/03/$12.00 DOI: 10.1037/0033-2909.129.5.747

747

Kleinman’s (1987) distinction between disease and illness. Ac- cording to Kleinman, disease represents the biological entity for which the pathophysiology is known, and illness represents the culturally shaped manifestation of the disease.

Because research into the genetic diathesis of eating disorders represents a search for their underlying disease, this search would benefit from an examination of whether or not core features of these syndromes exist outside of modern Western influences. We begin by evaluating whether rates of eating disorders have in- creased significantly over recent decades because evidence of a secular increase has been used as the cornerstone of arguments that they are caused by the recent emphasis on thinness (Garner et al., 1980; Russell, 1985). However, this evidence has been challenged by several investigators (Fombonne, 1995, 1996; Lucas, 1992; Pawluck & Gorey, 1998; Van’t Hof & Nicolson, 1996), sparking considerable debate on this point. We then examine the cross- historical and cross-cultural evidence for eating disorders. The purpose of this article is to explore the extent to which eating disorders, specifically AN and BN, represent culture-bound syn- dromes and to discuss the implications of these findings for un- derstanding the role genes play in their etiology.

Method

In order to examine the extent to which eating disorders are represented outside of their current sociohistorical context, this review was subdivided into three sections: epidemiology since the recognition of the modern syndromes, historical evidence of the syndromes prior to their formal recognition, and cross-cultural evidence of the syndromes.

Meta-analyses were conducted to investigate whether AN and BN have become more common during the twentieth century by examining secular trends in incidence rates (newly identified cases per 100,000 population per year) according to the methods described in Rosenthal and Rosnow (1991). Specific details of this method are described within the Epidemiology of AN in the Twentieth Century section in the Data coding subsection. As noted by Hoek (1993; Hoek, Van Hoeken, & Katzman, 2003), epidemio- logical studies of AN tend to use incidence, and epidemiological studies of BN tend to use prevalence (percentage of individuals who currently [point] or have ever [lifetime] met criteria for a disorder regardless of when the disorder first emerged). Because incidence rates track when cases are first identified, only incidence rates are considered in the meta-analyses. Both the cross-historical and cross-cultural sections represent a qualitative re- view, as they necessarily relied more on specific case histories for suffi- cient detail to evaluate the presence of eating disorders.

We restricted our review to published works. Benefits of this approach include the increased scientific rigor of published (particularly peer- reviewed) work compared with unpublished work and permanent avail- ability of published work within the public domain for independent eval- uation by others. Because restriction to published articles has the potential to be biased by the “file-drawer phenomenon” (Rosenthal & Rosnow, 1991), we have included information concerning the number of studies with nonsignificant results that would alter meta-analytic findings and interpretations within the Epidemiology of AN in the Twentieth Century and Epidemiology of BN in the Twentieth Century sections.

Definitions

We sought information for the diagnostic entities of AN and BN as they are conceptualized in the Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM–IV; American Psychiatric Assocation, 1994; 4th ed., text rev.; DSM–IV–TR; American Psychiatric Association, 2000). We did not attempt to limit our review to cases meeting full DSM–IV criteria for AN and BN. Rather the DSM–IV was used to define boundaries that

separate these as distinct syndromes. Although a great deal of recent attention has been paid to binge-eating disorder in the literature, less emphasis has been placed on the role of culture in the etiology of this proposed syndrome, and significant findings on its genetic bases remain limited (Burnet, Smith, Cowen, Fairburn, & Harrison, 1999; Branson et al., 2003; Ricca et al., 2002). We therefore restricted our review to AN and BN.

Using the DSM–IV conceptualization of AN and BN to organize our search introduces a sociohistorical bias. Thus, a brief discussion of both our rationale and our attempts to recognize these biases is warranted. The first issue concerns the overlap of AN and BN. In the Diagnostic and Statistical Manual of Mental Disorders (3rd ed.; DSM–III; American Psychiatric Association, 1980), a diagnosis of AN trumped a diagnosis of bulimia. In the Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; DSM–III–R; American Psychiatric Association, 1987), both diagnoses could be given concurrently. In the DSM–IV, a diagnosis of AN again trumps a diagnosis of BN. Based on data concerning age of onset, natu- ralistic course, treatment response, and mortality, the DSM–IV conceptu- alization has considerable empirical support (Keel & Mitchell, 1997; Keel, 2003). Moreover, women with concurrent symptoms of AN and BN do not differ from women with the restricting subtype of AN, and all women with AN differ significantly from women with BN alone on naturalistic course, treatment response, and mortality (Herzog et al., 1999; Keel et al., 2003). The second issue concerns the combination of binge eating with inappro- priate compensatory behavior in defining BN. In the DSM–III, inappropri- ate compensatory behavior in the form of self-induced vomiting was included as a possible, but not necessary, feature of bulimia. Subsequent work comparing individuals with binge eating alone versus binge eating and inappropriate compensatory behavior, most often purging, shows sig- nificant differences in gender ratios, etiologic factors, clinical correlates, and course (Halmi, 2003; Keel, 2003). Moreover, the DSM–IV definition of BN has demonstrated predictive validity in being distinguished from both AN and binge-eating disorder (Keel, Mitchell, Miller, Davis, & Crow, 2000). Despite evidence supporting the validity of the DSM–IV conceptu- alizations of AN and BN, there remain several definitional shortcomings. The impact of these shortcomings on evidence of these disorders outside of their current sociohistorical context is considered and discussed throughout the review. Particular attention is given to the cognitive features of body image disturbance that define these disorders.

One aspect of this review that became challenging was the need to define the construct of a “non-Western” culture. In reading other reviews, it became clear that what was considered non-Western varied considerably. Indeed, the Cassell Dictionary of Modern Politics (East & Joseph, 1994) describes the West as “a misleading term in a variety of usages” (p. 319). For the purposes of this article, we selected the definition for non-Western used by Karan (in press):

The area whose cultures developed essentially apart from the Greco– Judaic–Christian tradition of the Western culture. Thus, it includes East Asia (China, Japan, and Korea), Southeast Asia, the Indian subcontinent, the Middle East, and Sub-Saharan Africa. Latin Amer- ica, Russian Asia, and Oceania are excluded. The indigenous cultures of Latin America and Oceania are not western, but they are also not non-Western . . . the contemporary cultural pattern in Latin America and Oceania, unlike the non-Western cultures of Asia and Africa, is based dominantly on western traditions and values. (p. 1)

Although clear cultural differences exist among countries such as Bul- garia, Portugal, and the United States and may seem greater than cultural differences among Hong Kong, Japan, and the United States, Karan’s (in press) definition of non-Western is influenced less by recent political and economic changes that essentially coincide with the recognition of eating disorders within Western cultures. Further, this definition acknowledges aspects of culture that are derived from historically unique evolutions of civilization.

748 KEEL AND KLUMP

Literature Search

Literature reviews were conducted using computer databases (PsycINFO, 1887–June 2002; MEDLINE, 1966 –June 2002; RLG’s Eureka Anthropological Literature, 1900 –June 2002; and ABC-CLIO Historical Abstracts, 1954 –June 2002) and hand searches of the reference sections of all articles, books, and book chapters retrieved by computer and hand search. The following search terms were used: eating disorder, anorexia nervosa, bulimia, epidemiology, incidence, prevalence, history, culture, cross-cultural, and the names of every nation and continent in the world with the explode option to detect relevant permutations of terms. For the historical search, the terms binge, purge, vomit, and starvation were also included using the explode option. Pamela K. Keel reviewed references for the epidemiology and cross-historical sections. Titles and abstracts re- trieved for non-Western nations (n � 1,190) were evaluated by Kelly L. Klump for potential relevance, and articles were obtained to determine whether cases of AN or BN were reported in non-Western nations. Al- though several articles provided evidence of elevated levels of disordered eating behaviors and attitudes in non-Western cultures as assessed by standardized questionnaires (Al-Subaie, 2000; Al-Subaie et al., 1996; Hooper & Garner, 1986; A. M. Lee & Lee, 1996; S. Lee, 1993; S. Lee & Lee, 2000; S. Lee, Lee, & Leung, 1998; S. Lee, Lee, Leung, & Yu, 1997; S. Lee, Leung, Lee, Yu, & Leung, 1996), we sought specific instances of clinical eating disorders in these regions. Because we were specifically interested in evidence of eating disorders outside of our current sociohis- torical context, an attempt was made to include information from articles regardless of language. Thus, some information is from English transla- tions of work originally published in another language and abstracts written in English for articles written in non-English languages, as well as reviews of these non-English works by other authors. Much of the historical information prior to the mid-seventeenth century was collected through secondary sources because English translations were unavailable for many original sources. Although this section of the review is vulnerable to the interpretations of the secondary source, many of the cases were reviewed by multiple authors with differing theses, thus allowing a reasonably comprehensive presentation.

Anorexia Nervosa

According to the DSM–IV–TR, AN is defined by the following diagnostic criteria:

A. Refusal to maintain body weight at or above a minimally normal weight for age and height (e.g., weight loss leading to maintenance of body weight less than 85% of that expected; or failure to make expected weight gain during period of growth, leading to body weight less than 85% of that expected).

B. Intense fear of gaining weight or becoming fat, even though underweight.

C. Disturbance in the way in which one’s body weight or shape is experienced, undue influence of body weight or shape on self- evaluation, or denial of the seriousness of the current low body weight.

D. In postmenarcheal females, amenorrhea, i.e., the absence of at least three consecutive menstrual cycles. (A woman is considered to have amenorrhea if her periods occur only following hormone, e.g., estro- gen, administration.) (American Psychiatric Association, 2000, p. 589)

Attempts to characterize the presence of AN across time or across cultures have been marked by debates concerning the core features of the disorder. Although there is universal agreement that AN represents a disorder marked by starvation, some (Banks, 1992, 1997; Katzman & Lee, 1997; S. Lee, 1995; Palmer, 1993) have

argued that weight phobia (Criterion B) represents an aspect of the culturally bound illness that is not necessarily related to the disease that underlies AN. Others (Beumont, 1988; Habermas, 1989) have argued that weight phobia is a necessary motivating force behind food refusal to distinguish AN from “other social (deprivation), physical or psychiatric causes of undernutrition” (Beumont, 1988, p. 171). Similarly, there have been debates over whether or not amenorrhea (Criterion D) should be retained as a diagnostic cri- terion or viewed as a consequence of malnutrition (Cachelin & Maher, 1998). Because the definition of AN continues to be contested, we included studies using various definitions of AN and noted how these definitions differed from that presented in the DSM–IV–TR.

Epidemiology of AN in the Twentieth Century

Table 1 provides reported incidence rates for AN from studies evaluating secular trends, organized by study cohort and presented in ascending order of year (r values and probability values were calculated by Pamela K. Keel as described below). Attempts to determine changes in incidence rates of AN have been plagued by the low base rate associated with the disorder and diagnostic uncertainty (Fombonne, 1995). Indeed, differences in diagnostic criteria have contributed to incidence rates differing more across studies than across time (see Table 1). Further, the definition of AN has changed over recent decades. Thus, studies relying solely on diagnoses recorded in case registers (e.g., Jones, Fox, Babigian, & Hutton, 1980; Kendell, Hall, Hailey, & Babigian, 1973; Møller- Madsen & Nystrup, 1992; Nielsen, 1990; Shinkwin & Standen, 2001; Williams & King, 1987) may be tracking decreasing strin- gency of diagnostic criteria for AN as well as secular trends for true incidence. To guard against this particular confound, several investigators (e.g., Eagles, Johnston, Hunter, Lobban, & Millar, 1995; Hall & Hay, 1991; Hoek et al., 1995; Joergensen 1992; Lucas, Beard, O’Fallon, & Kurland, 1988, 1991; Lucas, Crowson, O’Fallon, & Melton, 1999; Pagsberg & Wang, 1994; Szmukler, McCance, McCrone, & Hunter, 1986; Willi & Grossmann, 1983; Willi, Giacometti, & Limacher, 1990) have ensured application of uniform diagnostic criteria across years by directly evaluating medical records, interviewing clinicians and patients, or training clinicians to assess eating disorders.

Data coding. Across the epidemiological studies included in Table 1, 10 articles reported a statistically significant secular increase in AN incidence, 9 reported a failure to find a significant increase, and one reported a significant decrease. There was a clear relationship between study finding and inclusion of statistics re- quired for meta-analyses. Only 2 studies (Lucas et al., 1988; Turnbull, Ward, Treasure, Jick, & Derby, 1996) that failed to find a significant secular increase provided specific data concerning either effect size or probability value. Thus, findings from 7 studies (Hall & Hay, 1991; Hoek et al., 1995; Joergensen, 1992; Nielsen, 1990; Szmukler et al., 1986; Willi et al., 1990; Williams & King, 1987) indicating at least one nonsignificant secular trend would be excluded from a meta-analysis of reported statistics. To avoid this bias, original data concerning annual incidence per 100,000 were recorded from each study from a combination of tables and figures. Tabulations of these raw data are available on request from Pamela K. Keel. For articles reporting annual inci- dence based on averaged data over a period of several years, this

749CULTURE, GENES, AND EATING DISORDERS

annual incidence was assigned to the year representing the mid- point of that period.

Pearson product–moment correlation coefficients were then cal- culated between incidence rates and years of observation within each cohort. These correlations along with their associated one- tailed probability values appear in Table 1. Cohorts rather than articles were selected as the unit of analysis because one article (Kendell et al., 1973) included three separate cohorts, two of which were followed later in subsequent articles (Jones et al., 1980; Szmukler et al., 1986). Similarly, continued follow-up has been reported in subsequent articles for two other cohorts (e.g., Hoek, 1991, Hoek et al., 1995; Lucas et al., 1988, 1991, 1999). Finally, incidences for overlapping periods have been reported for one cohort in three separate articles (Møller-Madsen & Nystrup, 1992; Munk-Jørgensen, Møller-Madsen, Nielsen, & Nystrup, 1995; Nielsen, 1990). Thus, conducting analyses by cohort rather than article ensured independence of observations.

One challenge in conducting the meta-analysis was variations across studies in how incidence was calculated. Williams and King (1987) demonstrated that the increased incidence for AN over time observed in their study was due to an increase in the proportion of the population comprising adolescent and young adult women. After age and sex were controlled for, no significant trend was found. Thus, age- and sex-adjusted incidence rates provide the best measure of secular trends, and studies were weighted according to their method of calculating incidence. A coefficient of 3 was given to studies that controlled for both sex and age. Studies controlling only for sex were given a coefficient of 2, and studies reporting

incidence without controlling for age or sex were given a coeffi- cient of 1. Because the significance of correlation values was influenced by the number of years of observation (which improves reliability of observed time trends) and not by the size of the population from which data were drawn, population size was used to weight values in the meta-analysis on the following scale: 1 � a population of 10,001 to 100,000; 2 � 100,001 to 1,000,000; and 3 � greater than 1,000,000. Thus, the meta-analysis favored find- ings from larger populations because they are less vulnerable to minor fluctuations in numbers of cases.

Results. Combining effect sizes across cohorts according to the equations presented in Rosenthal and Rosnow (1991) resulted in a medium effect for increased AN incidence across cohorts over time (r � .35). Combining probability values across cohorts ac- cording to the equations presented by Rosenthal and Rosnow indicated a significant increase in incidence over time ( p � .00005, one-tailed). According to a formula presented in Rosenthal and Rosnow, approximately 205 nonsignificant studies would be required to alter the result of this meta-analysis.

Theander (1970) was the first author to publish evidence of increased incidence rates for AN; however, he noted many meth- odological reasons his findings might not reflect an increase in the true incidence of AN. These included “above all, increasing ten- dency to arrange medical care and hospitalisation for patients with anorexia nervosa” (Theander, 1970, p. 30). Of interest, during the period in which a significant increase in AN incidence was ob- served in Monroe County, New York (Jones et al., 1980; Kendell et al., 1973), a significant increase was observed for the number of

Table 1 Anorexia Nervosa Incidence

Midpoint Range Location Criteria Reported incidence r p Study

1945 1931–1960 Sweden — 0.24 .948 .002 Theander (1970) 1962 1935–1979 Rochester, Minnesota DSM–III,

DSM–III–R 7.3a .404 .109 Lucas et al. (1988)b

1935–1984 8.2a Lucas et al. (1991)c

1985–1989 8.3a Lucas et al. (1999)c

1968 1960–1969 New York — 0.37 .501 .058 Kendell et al. (1973)c

1960–1976 0.47 Jones et al. (1980)c

1968 1965–1971 London — 0.66 .886 .004 Kendell et al. (1973)c

1970 1956–1975 Zurich, Switzerland — 0.38–1.12 .822 .0005 Willi & Grossmann (1983)c

1983–1985 1.43 Willi et al. (1990)b

1974 1966–1969 Scotland Russell (1970) 1.6 .804 .000004 Kendell et al. (1973)c

1965–1982 4.06 Szmukler et al. (1986)b

1977 1972–1981 England — — .648 .0215 Williams & King (1987)b

1978 1965–1991 Northeast Scotland ICD 8 — .708 .00001 Eagles et al. (1995)c

1980 1970–1989 Bornholm County, Denmark ICD 10 1.6–6.8 .815 .0925 Pagsberg & Wang (1994)c

1981 1977–1985 Ireland ICD 8, ICD 9 4.18 �.935 .0001 Shinkwin & Standen (2001)d

1982 1970–1989 Denmark ICD 8 0.42–1.17 .179 .087 Møller-Madsen & Nystrup (1992)c

1973–1987 1.04 Nielsen (1990)b

1970–1993 — Munk-Jørgensen et al. (1995)c

1982 1977–1986 Fyn County, Denmark DSM–III–R 11.0e .004 .496 Joergensen (1992)b

1982 1977–1986 Wellington, New Zealand DSM–III 5.0 �.230 .261 Hall & Hay (1991)b

1987 1985–1989 The Netherlands DSM–III–R 8.1 .674 .212 Hoek et al. (1995)b

1989 1978–1992 Yamagata Prefecture, Japan DSM–III–R — .805 .202 Nadaoka et al. (1996) 1991 1988–1993 England and Wales DSM–IV 4.2a .128 .405 Turnbull et al. (1996)b

Note. DSM–III � Diagnostic and Statistical Manual of Mental Disorders (3rd ed.); DSM–III–R � Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.); ICD 8 � International Classification of Diseases (8th ed.); ICD 10 � International Classification of Diseases (10th ed.); ICD 9 � International Classification of Diseases (9th ed.); DSM–IV � Diagnostic and Statistical Manual of Mental Disorders (4th ed.). a Age- and sex-adjusted incidence rate. b Reported no significant change. c Reported significant increase in incidence. d Reported significant decrease in incidence. e Incidence for females ages 10 –24 years.

750 KEEL AND KLUMP

adolescents receiving psychiatric treatment in this region. Specif- ically, the incidence of AN rose from 0.35 to 0.64 per 100,000 population per year from 1960 –1969 to 1970 –1976 (Jones et al., 1980). Similarly, the incidence of children and adolescents using psychiatric services rose from 348 to 662 per 100,000 population per year from 1960 to 1976 (Roghmann, Babigian, Goldberg, & Zastowny, 1982). This suggests that cases of AN represented approximately 0.1% of children and adolescents receiving treat- ment from 1960 to 1976 with the apparent increase in AN inci- dence potentially reflecting the increase in adolescents being en- tered into the psychiatric case register. Supporting this possibility, Willi and Grossmann (1983) found that the percentage of those diagnosed with AN first hospitalized in psychiatry versus pediat- rics or internal medicine increased steadily from 0% to 31.6% between 1956 and 1975. These authors attributed this trend to changes in how emaciation was interpreted in adolescent girls: “In the past, the cause of the disease was more frequently assumed to be organic (hypopituitarism), whereas in recent years, the diag- nostic view of anorexia nervosa has changed from somatic to psychic causes” (Willi & Grossmann, 1983, p. 565). Secular increases in psychiatric contact for patients with AN have been reported by others (George, Weiss, Gwirtsman, & Blazer, 1987; Lucas et al., 1999). These patterns may explain why Szmukler et al. (1986) observed a secular increase in AN incidence ascertained through the Aberdeen Psychiatric Case Register in Scotland but observed no significant change in AN incidence ascertained through the general hospital register. Restricting our meta-analysis to just those investigations using general medical records indicates a small effect size (r � .25) that is statistically significant ( p � .03) for an increase in AN incidence over the 20th century. Approximately 23 nonsignificant studies would be required to alter the results of this meta-analysis. Thus, a modest increase in AN incidence has coincided with increasing idealization of thinness (Garner et al., 1980; Owen & Laurel-Seller, 2000; Rubinstein & Caballero, 2000; Wiseman et al., 1992).

Introduction of AN to the medical nomenclature. Although popular and professional attention to AN has increased dramati- cally over recent decades (perhaps out of proportion with the actual increase in AN incidence), AN is not a “new” disorder that first appeared when Western culture emphasized the thin ideal. The term anorexia nervosa was first introduced in the medical literature by William Gull in 1874 to describe four cases of adolescent girls with deliberate weight loss, among whom three went on to achieve full weight recovery (Gull, 1874). Independent descriptions of a similar syndrome appeared under the labels l’anorexie hysterique (Laségue, 1873) and anoressia (Brugnoli, 1875, as cited in Habermas, 1992b; Ruggerio, Prandin, & Mantero, 2001) throughout Europe at that time. Similarly, at the end of the nineteenth century, American physicians were beginning to differ- entiate anorexia as a syndrome distinct from the larger category of hysteria (Vandereycken & Lowenkopf, 1990), and German physi- cians were differentiating AN from the larger category of neuras- thenic disorders (Vandereycken, Habermas, Van Deth, & Meer- mann, 1991). A paper presented at a meeting of the South Australian Branch of the British Medical Association in 1882 documented two cases of anorexia in young women in Australia (Vandereycken & Beumont, 1990), and Kissyel reported a case of severe hysterical anorexia in an 11-year-old girl in Russia in 1894 (as cited in DiNicola, 1990b).

Both Gull (1868, 1874, 1888) and Laségue (1873) acknowl- edged the presence of several cases of anorexia prior to their providing the syndrome with a specific name. Laségue (1873) commented, “I wish to treat of a symptomatic complexus too often observed [italics added] to be a mere exceptional occurrence” (p. 265). Similarly, Gull (1888) stated his case was “an illustration of most of these cases” (p. 517) and in an earlier article referred to “young women emaciated to the last degree through hysteric apepsia” (Gull, 1868, p. 175). Thus, both the language used in these articles and the widespread attention and confirmation these works received from other physicians support that the condition was a familiar entity, although previously known by other names or viewed within a larger category of mental disorders.

The features of Gull’s (1888) AN included significant weight loss, slow pulse rate, skin changes, and loss of menstruation with “perversions of the ‘ego’ being the cause and determining the course of the malady” (p. 517). Gull (1874) specifically noted the following: “The want of appetite is, I believe, due to a morbid mental state. I have not observed in these cases any gastric disorder to which the want of appetite could be referred” (p. 25). In a similar vein, Gull (1874) advised physicians to treat the illness with feeding despite the likely protests of the patient. Laségue (1873) viewed l’anorexie hysterique as a form of hysteria in which psychological distress was converted into food refusal motivated by “disgust” or “uneasiness after food, vague sensations of full- ness, suffering, and gastralgia post-prandium” that “although hy- pothetical, is dreaded in advance” (p. 265). Laségue’s description, based on eight cases, resembled present-day AN in that his patients were primarily women between the ages of 15 and 20 years. Both Gull’s (1874) and Laségue’s formulations of the syndrome focused on the food refusal itself; emaciation (Criterion A) and amenorrhea (Criterion D) were viewed as consequences of the syndrome. This resembles the current view of bradycardia and constipation as sequelae of AN rather than core features of the syndrome. Another distinction between the formulation of the illness by both Gull (1888) and Laségue versus present-day AN was the absence of mention of weight or shape concern.

Cross-Historical Evidence of AN

Historical cases of anorexia. Prior to the formal recognition of AN as a distinct syndrome, several reports in the medical literature suggest the presence of AN throughout the nineteenth century. Marce (1860) described a form of hypochondria in which “young girls, who at the period of puberty and after a precocious physical development, become subject to inappetency carried to the utmost limits” (p. 264). He noted that patients actively refused to eat in the absence of true gastric incapacity (“the stomach is well able to digest, and suffers only from want of food”; Marce, 1860, p. 265) and that the result is that “all trace of adipose tissue had disap- peared, and the patients were reduced to skeletons” (p. 264). Marce emphasized that deliberate food refusal and resistance to treatment could lead to death in some cases; however, he also reported that patients could be cured if forcibly refed. From 1812 to 1917, W. L. L. Parry-Jones (1985) found evidence of 40 cases of AN among 36,000 hospital admissions in England, suggesting that 0.1% of psychiatric inpatients may have suffered from AN (similar to the rate reported from 1960 to 1976 in Monroe County, New York). Two cases were described in greater detail. One involved a young woman who was admitted to the Warneford

751CULTURE, GENES, AND EATING DISORDERS

Asylum in 1831 who was emaciated from chronic food refusal. She was noted to be “delirious from debility rather than insane” (W. L. L. Parry-Jones, 1985, p. 97), and she died after 12 days from starvation. The second case involved a 26-year-old woman admitted in 1862 who was emaciated from food refusal. She was described as “sensible of all that is said to her” (W. L. L. Parry- Jones, 1985, p. 99), and she was ultimately discharged following a 7-month treatment of refeeding. In an 1840 text on women’s diseases, Imbert characterized anorexie nerveuse by loss of appe- tite, refusal to eat, and emaciation (as cited in Vandereycken & Van Deth, 1994). Chipley (1860) described cases of extreme emaciation among females who actively refused to eat (sitomania) despite pangs of hunger and attributed the deliberate self- starvation to a desire for attention and notoriety.

Chipley’s (1860) conclusion concerning the motivation behind sitomania overlaps with the motivations attributed to so-called fasting girls of the eighteenth and nineteenth centuries. Throughout this period, cases of fasting girls gained great attention in the popular media throughout America and Europe (Bemporad, 1996; Brumberg, 1989; Vandereycken & Van Deth, 1994). Typically, these cases involved adolescent girls who abruptly refused to eat. The cases drew a mixture of medical concern and religious awe. Particularly notorious cases attracted national and international attention with details of the fast reported in newspapers and skepticism about the miraculous fasts expressed by medical pro- fessionals in scholarly journals. Thus, in all recorded cases, food refusal leading to emaciation was associated with the secondary gain of fame. And, for some, continued food refusal was a lucra- tive occupation. Not surprisingly, this secondary gain attracted fraud. Cases involving proven deception include Anne Moore, Margaretha Weiss, Barbara Kremers, and Eva Vliegen (Bemporad, 1996; Brumberg, 1989; Vandereyeken & Van Deth, 1994). Some fasting girls were acclaimed as miracles for their ability to eat nothing yet remain completely healthy (including maintaining a healthy weight). Conversely, other fasting girls became extremely emaciated, in some cases leading to their deaths. Among these were Lina Finch (1886), Kate Smulsey (1885), and Lenora Eaton (1881), who all reportedly died of starvation before the age of 22 as a result of their food refusal (Brumberg, 1989).

Sarah Jacobs (1869), the “Welsh Fasting Girl,” represents a particularly tragic case of self-starvation leading to death in a fasting girl (Bemporad, 1996; Brumberg, 1989; Vandereycken & Van Deth, 1994). She began to fast in 1867 at the age of 12. Her parents publicized her behavior with the support of a local rever- end who confirmed the authenticity of the claims. She became a tourist attraction for the curious and a source of inspiration for religious pilgrims. Sarah Jacobs’s fame attracted concern and skepticism from the medical profession. To resolve ongoing debate over her case, nurses from Guy’s Hospital were dispatched to keep watch on the girl. As a condition of the watch, her parents insisted that she not be offered unsolicited food. Over the course of the week, Sarah Jacobs grew feeble and lost the ability to regulate her own body temperature. The nurses and supervising doctors at- tempted to end the watch and recommended refeeding, but her parents refused to end or modify the conditions of the test, and Sarah never requested food. After 10 days, she died of starvation (Brumberg, 1989; Vandereycken & Van Deth, 1994).

According to Brumberg (1989), Anglo American girls during the Victorian era were well-acquainted with the religious fasting of medieval saints, and St. Catherine of Siena’s biography was in-

cluded in inspirational books for girls. Indeed, the actions of Sarah Jacobs’s parents seem to reflect a deep faith in a divine source of their daughter’s food refusal. However, the growing field of psy- chiatry viewed these individuals as suffering from “nervous” con- ditions. Thus, fasting girls of the eighteenth and nineteenth centu- ries embraced the continuity between the religious medieval fasting of saints and their own behaviors. Meanwhile, psychiatrists endorsed the continuity between the extreme fasts leading to death among adolescent girls in the United States, England, France, and Germany and the newly identified syndrome of AN.

Between the seventeenth and eighteenth centuries, “miraculous maids” were girls between the ages of 14 and 20 who engaged in self-starvation, modeling themselves explicitly after ascetic medi- eval saints (Bemporad, 1996). From 1685 to 1710, Bliss and Branch (1960) found seven medical doctoral theses on anorexia, including one authored by Hardenus in 1703 that attributed the disorder to dysfunction of the gastric nerves as well as distur- bances in the spirit and feelings (p. 8). Morton (1689) described nervous atrophy or a nervous consumption characterized by loss of appetite, extreme emaciation, amenorrhea, overactivity, and indif- ference to the condition (as cited in Morton, 1694, English trans- lation). Morton (1694) detailed two cases in the first chapter of his thesis on consumption. Mr. Duke’s 18-year-old daughter experi- enced amenorrhea and significant weight loss “like a Skeleton only clad with skin” (Morton, 1694, p. 9) that could not be attributed to tuberculosis or chlorosis. The patient engaged in “continual poring upon Books, to expose her self both Day and Night to the injuries of the Air” (Morton, 1694, p. 8). After initial compliance with treatment she “quickly tired with Medicines, she beg’d that the whole Affair might be committed again to Nature, whereupon consuming every day more and more, she was after three Months taken with a Fainting Fit and dyed” (Morton, 1694, p. 9). The cause of death is unclear; however, refeeding syndrome seems to be one plausible explanation (see Walsh, Wheat, & Freund, 2000, for a review of treatment complications in AN). The second case involved a 16-year-old boy whose loss of appetite was also attrib- uted to “studying too hard,” although he responded to treatment with a “Milk Diet,” “Riding,” and “Country Air” (Morton, 1694, p. 10).

For possible cases of AN in the sixteenth century, McSherry (1985) suggested that Mary, Queen of Scots (1542–1587), suffered from AN. However, too little information is included in the short report to examine the supposition. Conversely, a great deal has been written of fasting medieval religious ascetics (Bell, 1985; Bynum, 1987; Carroll, 1998; Rampling, 1985), and cases of self- starvation were recorded by hagiographers from the eighteenth century back to the twelfth century.

Bell (1985) reviewed the vitae of 261 saints who lived in the Italian peninsula from A.D. 1200 on for evidence of what he termed holy anorexia. Holy anorexia, like AN, involved food refusal resulting in emaciation overtly motivated by the belief that this reflected divine intervention. Of the 261 saints, approximately 170 had adequate information for determination of eating pathology, and Bell argued that over half of these 170 individuals demon- strated a pattern of holy anorexia, such as that presented by St. Catherine of Siena and St. Veronica.

St. Catherine of Siena entered a pattern of self-starvation at around 16 years of age that continued until her death in 1380 (at age 32 or 33). Her death was brought on by her refusal to consume food or water. Because St. Catherine was a prolific writer, more is

752 KEEL AND KLUMP

known about her internal experience of self-starvation than about any other historical case. She portrayed herself as feeling unable to eat and claimed that “I prayed continually and I pray to God and will pray that he will grace me in this matter of eating so that I may live like other creatures” (St. Catherine, as cited in Bell, 1985, p. 23). As with patients treated by Gull, (1874), Laségue (1873), and Marce (1860) St. Catherine refused to eat because she viewed herself as afflicted by an inability to eat. St. Veronica (Veronica Giuliani) began a pattern of self-starvation at age 18 that may have represented a relapse from a previous episode at age 15. Fellow nuns reported seeing St. Veronica sneaking into the kitchen and gorging herself on food when she thought no one else was around, and she was placed repeatedly in the infirmary where she was forced to eat and prevented from binge eating. Although St. Veronica wrote several diaries, little information is given concern- ing the motivation behind her food refusal. The closest revelation on this point is that she felt she was “in a race against all the other novices to show who loved God the most” (Bell, 1985, p. 71). In contrast to St. Catherine, St. Veronica ultimately achieved recov- ery sometime between her 30s and 50s and lived until the age of 67 (death 1727).

Despite the acceptance of divine intervention as the reason for self-starvation in these cases, Bell (1985) described in detail the repeated attempts of peers and superiors to induce eating in the fasting saints to help them avoid the sin of vainglory, an inability to engage in holy responsibilities, or the sin of suicide. According to Bell, the Reformation functionally removed self-starvation as a means of achieving holiness for women such that St. Clare of Assisi (death 1253) and St. Catherine of Siena (death 1380) were revered for their self-starvation whereas Sister Domenica (death 1553) and Catherine Vannini (death 1606) were viewed with significantly greater suspicion. Bynum (1987) argued for a plural- ity of meaning for self-starvation among the fasting saints through- out this period such that extreme asceticism was alternately viewed as a mark of God’s grace, demonic possession, fraud, and illness within the lifetimes of all fasting saints, not just those following the Reformation. Bynum’s (1987) text includes the following quote from St. Catherine’s confessor: “If food was ever forced down her throat, intense pain followed, no digestion took place, and all that had been violently forced down was violently forced back again” (p. 168). The use of forced feeding emphasizes the ambivalence that St. Catherine’s fasting met. According to Rug- giero et al. (2001), psychiatric interpretations of self-starvation in young women were published in Italy by Pietro di Abano (1270 – 1316), Alessandro Benedetti (1533), Simone Porta (1551), and Becarri (1745). Thus, self-starvation in pursuit of a religious ideal was not unambiguously sanctioned in medieval times any more than self-starvation in pursuit of a thin ideal is sanctioned in modern times.

Purported cases of AN prior to the twelfth century include St. Wilgefortis, who allegedly engaged in self-starvation resulting in emaciation and lanugo (fine, downy body hair usually restricted to the fetal stage of development) sometime between A.D. 700 to 1000 (Lacey, 1982). Lacey (1982) argued that the popularity of this legend reflected the presence of AN throughout the European countries in which the bearded saint was known by various names (St. Ontkommena, St. Kummernis, St. Livrade, St. Leberata, and St. Uncumber). Hajal (1982) reviewed the case of self-starvation in al-Mut’tazz billah, son of Khalifah, as a case of AN successfully treated by Bukhtishu’ibn Jibrail (870). Cases of self-starvation

attributed to demonic possession or cured by exorcism were doc- umented during the 5th and 8th centuries, and early Christianity offers a possible case of AN from the late 4th century in Blessila in which an individual died from self-starvation at the age of 20 (Bemporad, 1996; Lacey, 1982). However, these earliest cases are not without controversy. For example, Bynum (1987) character- ized the description of St. Wilgefortis as having AN as a “bizarre communication to a British medical journal” (p. 194). Indeed, the story of St. Wilgefortis recounted by Lacey is open to numerous interpretations, among which AN is one possibility. Moreover, prior to the twelfth century, details become sparse, increasing the difficulty of interpreting these potential cases.

Interpretation of historical cases of anorexia. Although all of the above cases have been included in reviews of the history of AN, controversy exists concerning the validity of retrospectively diagnosing these women with AN. Much of the debate centers on the motivation behind food refusal, and limited information on this point is the greatest challenge to interpretation. Habermas (1989, 1996) has argued that modern-day AN extends to historical cases presented in the latter half of the nineteenth century by Gull (1874, 1888), Laségue (1873), and Marce (1860) but not to fasting found among medieval religious ascetics or Victorian-era females with hysteria or fasting girls. He argued that fasting among religious ascetics and Victorian females with hysteria was not AN because (a) fasting was used to heighten spirituality along with other forms of self-mortification, (b) fasting was associated with convales- cence and paralysis, and (c) fasting was interpreted in religious terms both by the fasting women and by the clergymen and doctors who wrote of these cases: “if some of the fasting women had been worried about their body weight, it would have been a concept accessible also to the reporting clergymen and doctors” (Haber- mas, 1989, p. 261). Conversely, Habermas (1989) interpreted early cases of AN (Gull, 1888) and l’anorexie hysterique (Laségue, 1873) as being motivated by weight phobia despite a lack of documentation of this feature. He concluded that “it is plausible that weight concerns have been present in AN from its beginning on but have been overlooked by most German and British medical men” (Habermas, 1989, p. 269) because early accounts of AN provided no “plausible motivation” (p. 269) for self-starvation and patients with AN tend to hide their goal of losing weight and give other explanations for their refusal to eat such as loss of appetite, inability to eat, and stomachache. According to Habermas (1989), weight loss was the “secret motivation” (p. 269), allowing it to be a necessary, if hidden, feature for defining the presence of the disorder since its introduction in the latter half of the nineteenth century. The absence of this motivation in earlier cases makes them invalid in his view. However, Habermas’s (1989) discussion leaves unclear why the presence of religious motivation would negate the presence of weight phobia (or vice versa) or why weight concerns would be a more accessible concept for medieval cler- gymen and doctors than it was for Gull (1874) or Laségue.

The presence of religious motivation does not preclude the presence of weight concerns as evidenced by modern cases of “spiritual starvation” (Banks, 1992, 1997; Bynum, 1988; Katzman & Lee, 1997; Morgan, Marsden, & Lacey, 2000). Cases presented by Banks (1992, 1997) involve women who fasted to engender spiritual purity and thus bear a striking resemblance to historical cases of fasting performed by medieval religious ascetics. In one case, the syndrome resulted in death despite numerous hospital- izations. In the second case, the syndrome remitted and full recov-

753CULTURE, GENES, AND EATING DISORDERS

ery was achieved. In terms of outcome, these cases are not unlike those of St. Catherine and St. Veronica, respectively. Morgan et al. (2000) reported on three cases of spiritual starvation. Although the authors indicated that DSM–IV AN was diagnosed in two of these cases, religious themes appeared to provide primary motivation for deliberate self-starvation in all three cases (Morgan et al., 2000).

It seems unlikely that Gull (1874) and Laségue (1873) over- looked weight concerns as a motivation for food refusal because contemporary authors noted these features in fasting girls. Win- slow (1880) provided the following quote attributed to Erasmus Darwin from the end of the eighteenth century:

Some young ladies I have observed to fall into this general debility but so as but just to be able to walk about, which I have sometimes ascribed to their voluntary fasting, when they believed themselves too plump; and who have thus lost both health and beauty by too great abstinence, which could never be restored. (p. 281)

Similarly, Charcot emphasized the role of weight phobia in the etiology of self-starvation among his patients (as cited in Haber- mas, 1989; Vandereycken & Van Deth, 1994). The work of both Winslow and Charcot would have been available to Gull and Laségue, suggesting that both men could be aware of this possible explanation for self-starvation. However, neither seemed to regard this information as a compelling explanation for food refusal in their patients. In a later publication, Habermas (1996) elaborated his argument that weight phobia represents the central feature of AN on the basis that diagnostic criteria have to differentiate syndromes in such a way that a syndrome represents a homoge- neous group distinct from other syndromes. Although weight pho- bia achieves this end, it does not appear to be the only means to this end.

Given both current (as well as historical) plurality in motiva- tions behind self-starvation, Rieger, Touyz, Swain, and Beumont (2001) recommended requiring ego-syntonic emaciation rather than weight phobia to differentiate AN from other conditions that lead to weight loss. The authors noted that cases of AN are uniform in the extent to which self-starvation is deliberate and the re- sponses to resulting emaciation range from indifference to pride. Their proposition offers a solution to differentiating fasting to convey religious devotion that is common to most religions (Bliss & Branch, 1960) and that displayed in holy anorexia. Moreover, ritualistic religious fasting is time-limited and tends to focus on avoiding specific foods entirely or not eating during specific time intervals and thus does not typically result in emaciation. It lacks the persistent and pervasive nature of food refusal that character- izes the form of fasting observed in medieval religious ascetics or modern-day AN. In addition, weight loss that results from other physical and psychological conditions is often distressing to the affected individual (Rieger et al., 2001).

Despite the utility of Rieger et al.’s (2001) recommendation, there are other contemporary cases in which ego-syntonic emaci- ation would not represent an eating disorder or any other form of psychopathology. For example, deliberate self-starvation may be used as an effective means of protest. Although women with AN can be seen as using self-starvation as a means of protest (Brum- berg, 1989), this can be differentiated from nonpathological forms of deliberate self-starvation. Individuals who engage in hunger strikes do so as a means to an end. When the end is reached, the self-imposed starvation ceases. Mogul (1980) wrote, “What dis- tinguishes adaptive asceticism from pathological states is not so

much its extent, or even the subjective experience of gratification from it, but the degree to which the asceticism becomes an end in itself” (pp. 159 –160). Food refusal has been used as a means to different ends in different periods (moral purity, fame, attention, thinness), but in all periods it also becomes an end in itself. What seems to unite fasting saints and women with AN is the paradox that the starvation is both deliberate and nonvolitional. That is, across historical contexts women deliberately refuse to eat food that they require for sustenance. Yet, they do not appear to be able to stop their pattern of food refusal in response to reward or punishment. This characterization resembles Shafran, Cooper, and Fairburn’s (2002) depiction of clinical perfectionism as the core pathology of AN. The internal motivation to avoid eating overrides all internal and external drives to eat, and for many adolescent girls throughout history this means becomes their ultimate end.

Although motivations for food refusal may have differed across periods (in many cases this information is simply lacking), pur- ported motivations may not represent the true causes of self- starvation. Instead, they may represent culturally meaningful at- tempts to understand an affliction that leaves women feeling unable and unwilling to eat. The extent to which fear of fat is viewed as causing AN may be, in part, an illusion (Wegner, 2002). Thus, if the core feature of AN is taken to be an intentional yet nonvolitional self-starvation, then evidence of AN appears to trace back to early medieval times. Furthermore, across the various historical contexts in which AN has emerged, this disorder dem- onstrates a particular affinity for affecting females beginning in their adolescence and thus satisfies Habermas’s (1996) call for defining the core features of a syndrome such that they retain descriptive specificity.

Cross-Cultural Evidence of AN

Although several reports have suggested that AN was nonexist- ent in non-Western regions such as sub-Saharan Africa (German, 1972), Northern Sudan (Elsarrag, 1968), and Southeast Asia (Neki, 1973), our review of the literature indicated that AN has been observed in every non-Western region of the world. Table 2 summarizes features of reports of AN in non-Western nations, including the country, diagnostic criteria, number or prevalence of cases, presence of weight concern among these cases, and level of Western influence. These publications included case reports, let- ters to journal editors, and clinical and epidemiological investigations.

Data coding. Presence of weight concerns was evaluated be- cause of the evidence reviewed above that this symptom varies both across and within historical periods. For studies indicating that all cases met diagnostic criteria that require weight concerns and for which details for specific case histories were not provided, we concluded that all cases (e.g., 100%) must have reported weight concern. Unfortunately, this did not preclude an inconsistent or idiosyncratic application of diagnostic criteria in those studies. By contrast, for cases in which weight concerns were explicitly denied by the patient but were perceived by the author to be the secret motivation, we considered the feature to be absent. Some of these cases may resemble patients with AN who have denied fears of becoming fat only for the fears to emerge as weight restoration begins. However, inferring the presence of weight concerns in the absence of any confirming evidence prevents awareness of cases in which this feature is genuinely absent. Further, it runs the risk of

754 KEEL AND KLUMP

relying on an ethnocentric conception of the core features of AN (S. Lee, Ho, & Hsu, 1993).

For characterization of Western influence, we took authors’ estimation of the degree of Western influence when possible. When authors made no comment on this issue, we used all avail- able information to code the study appropriately. We developed six categorizations in this regard: (a) No—at least one reported case where the author specifically noted that the case lacked exposure to Western influence (e.g., Abou-Saleh, Younis, & Karim, 1998, described a case of AN in an 18-year-old nomadic woman from the Empty Quarter in the United Arab Emirates and specifically noted the absence of Western influence in this case); (b) Unlikely—at least one case likely meets all of the following criteria: comes from a nonurban region, is of low socioeconomic status, resides within a traditional family, has not surpassed secondary education, and is non-English speaking; (c) Probable—all cases likely meet at least one of the following criteria: reside in an urban setting, are of high socioeconomic status, are English-speaking, or have higher edu- cational attainment (i.e., baccalaureate or higher); (d) Definite—all cases spent considerable time in a Western nation before develop- ing eating pathology (e.g., Ford, 1992, reported a case of BN in an Egyptian individual who developed the disorder after spending ages 5–17 in Canada); (e) Indeterminable—not enough informa- tion was provided to determine the degree of Western influence (e.g., data come from patients treated in an urban medical center; however, it is not possible to determine if any patients would fit above criteria for unlikely or no Western influence); and (f) Unknown—the original article was not written in English and information on Western influence was not included in the English abstract or English review article from which data were taken. Both the presence of weight concerns and Western influence were coded by consensus between Pamela K. Keel and Kelly L. Klump.

Results. Although food refusal leading to emaciation was re- ported for all cases in Table 2, the presence of weight concerns as a motivating factor for food refusal does not appear to be universal. S. Lee et al. (1993) described 70 cases of AN in Hong Kong among whom 59% did not report weight concerns. As with the cases described by Laségue (1873) and Gull (1874), digestive discomfort was the most frequent reason for not eating. These data suggest that Westernization and industrialization bring about cer- tain aspects of AN (weight concerns) but are not necessary for producing a self-starvation syndrome. Alternatively, as has been argued for historical cases, weight concerns may have been present but denied by some patients. S. Lee et al. (1993) predicted that Western influence in China would result in increased presentation of weight-phobic AN. Consistent with this hypothesis S. Lee (2000) and Lai (2000) have both described an increasing preva- lence of AN characterized by body dissatisfaction in the East. Further, S. Lee and Lee (2000) found that weight concerns were greatest in Hong Kong and least prominent in Hunan, suggesting an association between Westernization and body image disturbance.

Although some authors have indicated that AN cases were limited to the social elite or higher socioeconomic classes (Buchan & Gregory, 1984; Chadda, Malhotra, Asad, & Bambery, 1987), several reports suggested the presence of AN in the absence of Western influence (Abou-Saleh et al., 1998; Chandra et al., 1995; Fahy, Robinson, Russell, & Sheinman, 1988; Gandhi, Appaya, & Machado, 1991). When Western exposure could be determined,

42% (8/19) of articles reported at least one case with no or unlikely exposure to Western ideals.

Epidemiological data not only support the existence of AN in non-Western countries but suggest that its prevalence may be similar to that in Western nations. The lifetime prevalence of DSM–III AN in Korea did not differ from estimates produced by the Epidemiological Catchment Area study for lifetime prevalence of AN in New Haven or St. Louis (C. K. Lee et al., 1987, 1990). A recent study (Nobakht & Dezhkam, 2000) reported the lifetime prevalence of AN to be 0.9% in Iranian schoolgirls, a value that is higher than the lifetime prevalence reported in the DSM–IV–TR, and the incidence and prevalence of AN in Hong Kong (Chen et al., 1993; S. Lee, Chiu, & Chen, 1989) and Japan (Azuma & Henmi, 1982; Kuboki, Nomura, Ide, Suematsu, & Araki, 1996) are similar to those reported in Western nations (e.g., Kendell et al., 1973; C. K. Lee et al., 1987; Møller-Madsen & Nystrup, 1992; Theander, 1970, Willi & Grossmann, 1983). According to reports from Malaysia (Buhrich, 1981; Goh, Ong, & Subramaniam, 1993) and Egypt (Okasha, Kamel, Sadek, Lotaif, & Bishry, 1977), AN constituted between 0.05% and 0.19% of psychiatric cases. These figures are roughly comparable to those reported for Norway from 1990 –1994 (Götestam, Eriksen, Heggestad, & Nielsen, 1998). Thus, excluding the criterion of weight concerns, AN appears to represent a similar proportion of the general and psychiatric pop- ulations in several Western and non-Western nations.

Summary of Findings for AN

It does not appear to be the case that AN is a culture-bound syndrome. Although cultural factors such as the increasing ideal- ization of thinness may influence rates of AN, such factors seem neither sufficient nor necessary. Moreover, evidence of a secular increase in AN incidence is modest after controlling for factors such as changes in population distribution, population size, and ascertainment of cases through psychiatric registers. Cases of deliberate yet nonvolitional self-starvation, sometimes resulting in recovery, sometimes resulting in death, and primarily affecting young adolescent girls, exist in numerous historical periods. Sim- ilarly, specific cases of AN that cannot be attributed to the influ- ence of Western ideals have been reported in the Middle East, the Indian subcontinent, and East Asia. Like historical cases that emerged in the absence of the “modern ‘cult of thinness’ ” (Rus- sell, 1997, p. 23), many of these non-Western cases lack weight concerns. Our results suggest that weight concerns may be a culturally bound phenomenon, restricted to sociohistorical con- texts that idealize thinness and denigrate fatness. However, this is not entirely accurate either because a plurality of motivations for food refusal has been described both across and within sociohis- torical contexts.

Bulimia Nervosa

The most recent edition of the DSM (DSM–IV–TR) defines BN with the following diagnostic criteria:

A. Recurrent episodes of binge eating. An episode of binge eating is characterized by both of the following:

(1) eating, in a discrete period of time (e.g., within any 2-hour period), an amount of food that is definitely larger than most people would eat during a similar period of time and under similar circumstances

755CULTURE, GENES, AND EATING DISORDERS

Table 2 Studies of Anorexia Nervosa (AN) in Non-Western Nations

Article Place Diagnostic

criteria % weight concern Number–prevalence

Western influence?

Africa

Norris (1979) Johannesburg, South Africa

Study specific 100 54 patients with AN; 8 Afrikaans-speaking and 4 other non-English-speaking patients

Indeterminable

Nwaefuna (1981) Nigeria Not given 100 1 case described in a Black Nigerian woman Indeterminable Famuyiwa (1988) Nigeria Not given 0 2 suspected cases reported; symptoms were loss

of � 30% body weight, denial of seriousness of low weight, and amenorrhea

Case 1: definite; Case 2: indeterminable

Buchan & Gregory (1984)

Zimbabwe Feighner Questionable casea

1 case reported in Black Zimbabwean Definite

Fahy et al. (1988) Ethiopia Not given 100 1 case reported in Ethiopian torture victim Probable

Middle Eastern–Arab cultures

Okasha et al. (1977) Cairo, Egypt Not given Not given 2 cases of AN; 0.19% of psychiatric cases Probable Abou-Saleh et al. (1998) United Arab

Emirates DSM–III–R

for 3 of 5 patients

80 5 cases reported, among which diagnoses of DSM–III–R AN were confirmed by interview for 3 cases; 1 case (20%) lacked exposure to Western influences

No

Nobakht & Dezhkam (2000)

Iran DSM–IV 100 0.9% of 3,100 female high school students age 15–18

Probable

Yager & Smith (1993) Pakistan Not given 100 1 case reported No

Indian subcontinent

Neki, Mohan & Sood (1977)

India Study specific Not given 2 cases who were monozygotic twins Indeterminable

Chakraborty (1985) India Not given Absent in 8–10 cases

15–17 cases reported Indeterminable

Chadda et al. (1987) India DSM–III 100 1 case of AN from the upper social class in India Probable Gandhi et al. (1991) India Not given 60 5 cases reported No Khandelwal et al. (1995) India Modified

Feighner 20 5 cases reported Unlikely

Chandra et al. (1995) India Modified DSM–III–R

33 3 cases reported No

Southeast Asia

Ong et al. (1982) Malaysia Not given 71 7 cases reported—1 case with no formal education from lowest social class

Unlikely

Buhrich (1981) Malaysia Study specific Not given 30 cases reported; 19 Chinese, 8 Indian, 1 Malaysian, 2 Eurasian; no change reported in prevalence over 15- to 20-year period; 0.05% of psychiatric referrals

Indeterminable

Krahl et al. (1981) Malaysia Unknown Unknown 2.5% of cases presenting to a community child guidance clinic

Unknown

Krahl (1983)b Malaysia Unknown Unknown 2 cases reported Unknown Goh et al (1993) West Malaysia Study specific 60 15 cases of partial AN in 9,000 female patients

from 1970–1988; 0.16% of female patients Indeterminable

Ung, Lee, & Kua (1997) Malaysia Not given 33 patients with AN; 36% partial syndrome patients with “lack of amenorrhea in anorexic range” (p. 334)

Indeterminable

East Asia

S. Lee, Chiu, & Chen (1989)

Hong Kong Modified DSM–III–R

67 Reported � 10 cases treated during 5 years (incidence � 0.4 per 100,000 person years); 3 cases described in detail

Indeterminable

Chiu (1989) Hong Kong DSM–III 100 1 case reported No S. Lee (1991) Hong Kong Modified

Feighner 19 16 cases reported Unlikely

S. Lee, Leung et al. (1991)

Hong Kong Study specific 0 2 cases reported Indeterminable

S. Lee, Ho, & Hsu (1993)c

Hong Kong Study specific 41 70 cases reported Indeterminable

756 KEEL AND KLUMP

(2) a sense of lack of control over eating during the episode (e.g., a feeling that one cannot stop eating or control what or how much one is eating)

B. Recurrent inappropriate compensatory behavior in order to pre- vent weight gain, such as self-induced vomiting; misuse of lax- atives, diuretics, enemas, or other medications; fasting; or excessive exercise.

C. The binge eating and inappropriate compensatory behaviors both occur, on average, at least twice a week for 3 months.

D. Self-evaluation is unduly influenced by body shape and weight.

E. The disturbance does not occur exclusively during episodes of Anorexia Nervosa. (American Psychiatric Association, 2000, p. 594)

Although the definition of BN has been revised since its introduc- tion to the psychiatric nomenclature, the above definition bears a

striking resemblance to that given by Russell (1979) in his seminal description of the disorder. Slightly earlier, independent reports of a binge–purge syndrome in normal-weight young women were published in German and Spanish (Ziolko, 1996) and have since been translated into English by the original authors (Doerr-Zegers, 1994; Ziolko, 1994). Notably, among these first descriptions of BN (Doerr-Zegers, 1994; Russell, 1979; Ziolko, 1994), no author described a large awareness of this syndrome before 1960. Based on BN’s recent appearance and dramatically increasing incidence, Russell (1997) concluded, “Bulimia nervosa is a new disorder” (p. 23).

Epidemiology of BN in the Twentieth Century

Table 3 provides incidence rates for BN from studies evaluating secular trends, organized by study cohort and presented in ascend-

Table 2 (continued )

Article Place Diagnostic

criteria % weight concern Number–prevalence

Western influence?

East Asia (continued)

Chen et al. (1993) Hong Kong DSM–III 100 0.03 lifetime prevalence of AN in 7,229 males and females surveyed

Probable

Lai, Pang, & Wong (1995)

Hong Kong Not given 100 1 case reported Probable

Kam & Lee (1998) Hong Kong Study specific 50 1 case reported Indeterminable S. Lee, Lee, & Leung

(1998) Hong Kong “Broadly

diagnosed” 58 26 cases of AN from case series treated from

1990–1996 Probable

Lai (2000) Hong Kong Lask (1992) 88 16 cases reported Probable S. Lee (2001) Hong Kong Modified

DSM–IV 67 48 cases reported Indeterminable

Song & Fang (1990)d China DSM–III Unknown 9 cases reported Unknown Tseng et al. (1989)e Taiwan DSM–III Unknown 7 cases reported Unknown Hung & Cheng (1992)f Taipei, Taiwan DSM–III–R 100 2 cases of AN reported among 12,435 females

ages 12–25 Indeterminable

Sonoda et al. (1974)e Japan Unknown Unknown 1 case reported in a 10-year-old girl Unknown Nogami & Yabana

(1977)g Japan Not given Not given 9 cases reported Indeterminable

Azuma & Henmi (1982)h Japan Unknown Unknown 0.2% urban areas; 0.05% rural areas Unknown Mizushima & Ishii

(1983) Japan Study specific 100 16 girls and 1 boy with AN among 36,403 girls

and 37,520 boys in secondary school Unknown

Suematsu et al. (1985) Japan Study specific 75% of 1,011

2,392 patients treated in 1980–1981; detailed data collected from 1,011 cases

Indeterminable

Suematsu et al. (1986) Japan Study specific 100 7 cases reported Indeterminable Kuboki et al. (1996) Japan Modified

DSM–III–R 100 2.9–3.7/100,000 general population 1985; 3.6–

4.5/100,000 general population 1992 Indeterminable

Nadaoka et al. (1996) Japan DSM–III–R 100 59 cases of AN treated at Yamagata University Hospital between 1978 and 1992

Probable

Nakamura et al. (2000) Japan DSM–IV 100 0.005% of female population; 0.003% of general population

Indeterminable

Takei et al (2000)e Japan DSM–IV 100 223 students with AN from 335 college health administration centers across Japan

Unknown

C. K. Lee et al. (1987) Korea DSM–III 100 0.02% of general population Indeterminable C. K. Lee et al. (1990) Korea DSM–III 100 0.03% of general population Indeterminable

Note. Not given � authors did not provide the information in the article; Unknown � authors may have provided information in the non-English publication, but the information was not provided in the English abstract or the English review article summarizing it; Study specific � diagnostic criteria were specific to the study indicated; Indeterminable � exposure to Western influence could not be determined from article; DSM–III–R � Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.); DSM–IV � Diagnostic and Statistical Manual of Mental Disorders (4th ed.); DSM–III � Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). a Report states “score on the Eating Attitudes Test (Garner & Garfinkel, 1979) was 78 — clear evidence of her terror of becoming overweight” (Buchan & Gregory, 1984, pp. 328 –329). However, initial weight loss does not seem intentional, and case was characterized by many psychotic and dissociative symptoms. b As reported in DiNicola (1990b). c 16 (23%) participants also took part in the S. Lee (1991) study. d As reported in S. Lee, Ho, and Hsu (1993). e Data were taken from an English abstract of a non-English publication. Overlap of cases across reports from a single country is unknown. f As reported in Tsai (2000). g As reported in Davis and Yager (1992). h As reported in Dolan (1991).

757CULTURE, GENES, AND EATING DISORDERS

ing order of year. Among these studies, all reported a significant rise in cases over time (Hall & Hay, 1991; Hoek et al., 1995; Joergensen, 1992; Nadaoka et al., 1996; Pagsberg & Wang, 1994; Soundy, Lucas, Suman, & Melton, 1995; Turnbull et al., 1996). Of interest, several of these studies found a significant increase in BN incidence despite finding no significant change in AN incidence (Hall & Hay, 1991; Hoek et al., 1995; Turnbull et al., 1996). Combining results across the seven studies that provide BN inci- dence data over time (Hall & Hay, 1991; Hoek et al., 1995; Joergensen, 1992; Nadaoka et al., 1996; Pagsberg & Wang, 1994; Soundy et al., 1995; Turnbull et al., 1996) weighted by method of calculating incidence and population size (as described above for the meta-analysis of AN incidence data), we found a large effect size (r � .89) for a secular trend of increasing BN incidence that is statistically significant ( p � .0000001). Approximately 62 non- significant studies would be required to alter the results of this meta-analysis. Restricting our meta-analysis to include only those studies that used general medical records also results in a large effect size (r � .90) that is statistically significant ( p � .00002). Approximately 28 nonsignificant results would be required to alter the results of this meta-analysis. Thus, studies tracking a secular trend in BN incidence from 1970 to 1993 suggest a significant increase.

Fombonne (1996) argued that the patterns seen for BN fre- quency follow that expected for a newly defined disorder but do not necessarily represent the emergence of a previously nonexist- ent syndrome. Specifically, one would expect to see a significant increase in incidence immediately following official recognition of a disorder because incidence rates often reflect when individuals first seek treatment for a disorder. Thus, incidence rates are in- flated initially by the combination of new onset cases and cases with onset prior to disorder recognition. This description matches changes in incidence rates reported by Soundy et al. (1995). Incidence of BN increased steeply from 7.4 to 49.7 per 100,000 females from 1980 to 1983 and then remained relatively stable around 30 per 100,000 females from 1984 to 1990. According to Fombonne (1996), the decline from 49.7 to 30 cases per 100,000 females per year reflected a decline in cases with onset before the disorder had been defined rather than a true decrease in disorder incidence. Conversely, Soundy et al. noted that the particularly high rate in 1983 was inflated by a BN treatment study that year for which participants were being actively recruited from the community. Unfortunately, there are no incidence data for BN prior to 1970. Thus, one cannot use incidence rates to determine

how the frequency of this disorder changed in relation to the current cultural idealization of thinness as was possible for AN.

Despite the ever-present cultural ideal of thinness, BN preva- lence estimates were reduced from the DSM–III to the DSM–III–R. Early studies of BN point prevalence suggested that the disorder affected 8% (Pyle, Halvorson, Neuman, & Mitchell, 1986; Zuck- erman, Colby, Ware, & Lazerson, 1986) to 19% (Halmi, Falk, & Schwartz, 1981) of university women whereas later studies sug- gested a lower point prevalence of BN among college women with estimates ranging from approximately 1% (Schotte & Stunkard, 1987) to 3% (Drewnowski, Yee, & Krahn, 1988). Although these results suggest a decline in the point prevalence of BN running counter to the continued idealization of thinness for this time (Wiseman et al., 1992), several methodological factors, unrelated to changing prevalence, may account for the observed differences.

First, earlier studies used the DSM–III criteria (Halmi et al., 1981; Pyle et al., 1986; Zuckerman et al., 1986) whereas later studies used DSM–III–R criteria (Drewnowski et al., 1988; Schotte & Stunkard, 1987). Between the DSM–III and DSM–III–R, the diagnostic criteria for BN grew more stringent. For example, recurrent inappropriate compensatory behavior was required in the DSM–III–R but was not strictly required in the DSM–III. In addi- tion to being generally less stringent, the DSM–III was also less specific. For example, although the DSM–III criteria required “recurrent” binge-eating episodes, it did not specify a minimum frequency required for the episodes to be considered recurrent. Zuckerman et al. (1986) demonstrated that the simple stipulation of bingeing more than once per week versus just once per week reduced the point prevalence estimate of bulimia from 8% to 4% for college women in 1982. Similarly, Pyle et al. (1986) reported that the point prevalence of DSM–III bulimia in college women was approximately 8% in both 1980 and 1983; however, when weekly binge eating and purging were required, the point preva- lence dropped to 1% in 1980 and 3% in 1983. Of interest, the stricter definition resulted in a threefold increase in the point prevalence of bulimia among college women from 1980 to 1983— mirroring results of incidence studies for this period. Thus, had DSM–III–R criteria been used in the earliest studies, prevalence estimates might have ranged from 1% to 4%.

Second, different methods of assessment across studies over time may have contributed to the perceived decrease in prevalence rates. Early prevalence studies (Halmi et al., 1981; Pyle et al., 1986; Zuckerman et al., 1986) more often used survey assess- ments, and later studies more often used interview assessments

Table 3 Bulimia Nervosa Incidence

Midpoint Range Location Criteria Reported incidence r p Study

1980 1970–1989 Bornholm County, Denmark ICD 10 0.7–3.0 .905 .0475 Pagsberg & Wang (1994) 1982 1977–1986 Fyn County, Denmark DSM–III 5.5a .618 .0285 Joergensen (1992) 1982 1977–1986 Wellington, New Zealand DSM–III 6.0–44.0 .869 .0005 Hall & Hay (1991) 1985 1980–1990 Rochester, Minnesota DSM–III–R 13.5b .347 .1475 Soundy et al. (1995) 1987 1985–1989 The Netherlands DSM–III–R 11.5 .828 .042 Hoek et al. (1995) 1989 1978–1992 Yamagata Prefecture, Japan DSM–III–R — .863 .1685 Nadaoka et al. (1996) 1991 1988–1993 England and Wales DSM–IV 12.2 .968 .001 Turnbull et al. (1996)

Note. ICD 10 � International Classification of Diseases (10th ed.); DSM–III � Diagnostic and Statistical Manual of Mental Disorders (3rd ed.); DSM–III–R � Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.); DSM–IV � Diagnostic and Statistical Manual of Mental Disorders (4th ed.). a Incidence for females ages 10 –24 years. b Age- and sex-adjusted incidence rate.

758 KEEL AND KLUMP

(Bushnell, Wells, Hornblow, Oakley-Browne, & Joyce, 1990; Schotte & Stunkard, 1987). Self-report surveys tend to produce higher estimates of BN prevalence than structured clinical inter- views (Keel, Crow, Davis, & Mitchell, 2002). Bulimic symptoms such as binge eating are particularly vulnerable to high false- positive rates on self-report surveys compared with interviews because lay definitions tend to be more inclusive than clinical definitions (Whitehouse, Cooper, Vize, Hill, & Vogel, 1992). In addition, women appear to be more willing to admit to bulimic symptoms on self-report surveys compared with interviews be- cause the increased anonymity of questionnaires increases candor (Keel et al., 2002). Both patterns could contribute to an artificial decline in BN prevalence over time.

Third, differences in the samples evaluated have led to varying prevalence rates (Fairburn & Beglin, 1990). Few studies have used random sampling, and inconsistent sampling strategies across stud- ies could influence the extent to which individuals with eating disorders are over- or underrepresented in study groups. Moreover, epidemiological samples have varied in response rates, socioeco- nomic status, geographic location, ethnic representation, and age. Thus attempts to determine time-related changes in prevalence could be confounded by any of these differences. For example, women ages 18 to 24 (approximately college-aged) have rates of BN that are 2 to 10 times greater than those reported for women ages 25 to 64 (Bushnell et al., 1990).

An alternative to examining prevalence rates over different time periods is to measure lifetime prevalence rates across birth cohorts at one time. If a disorder has a stable incidence and onset occurs throughout life, then lifetime prevalence rates increase as birth cohorts age. If a disorder has a stable incidence but onset is limited to early ages (as it is for eating disorders), then lifetime prevalence rates would remain stable across birth cohorts that had passed through the period of risk. However, as noted above, lifetime BN prevalence rates are actually higher among younger birth cohorts compared with older birth cohorts (Bushnell et al., 1990; Kendler et al., 1991). These findings suggest either that the risk of BN has increased over recent years or that recall of past episodes of BN diminishes with age. Of interest, lifetime prevalence of AN has not differed significantly across birth cohorts (Robins et al., 1984; Wade, Heath, Abraham, & Treloar, 1996). Similarly, in an exam- ination of eating disorders among relatives of AN probands, rates of BN were higher among sisters compared with mothers (4.1% vs. 1.0%), consistent with lifetime prevalence rates for a disorder with increasing incidence rates (Strober, Lampert, Morrell, Bur- roughs, & Jacobs, 1990). Conversely, rates of AN were 3 times lower in sisters than in mothers (2.0% vs. 6.2%), consistent with lifetime prevalence rates for a disorder with stable incidence rates (Strober et al., 1990). Unfortunately, data on lifetime prevalence rates do not inform us regarding when the disorder developed relative to the cultural ideal of thinness.

There is evidence of changing rates of BN during the latter half of the 20th century; however, the causes of these changes are unclear. Incidence data suggest a significant increase in BN inci- dence from 1970 to 1990 but do not allow evaluation of BN incidence prior to 1970. Prevalence studies in college and primary care samples suggest a decreasing prevalence from the 1980s to the 1990s; however, inconsistent methods across prevalence stud- ies obfuscate the meaning of this observation. Finally, lifetime prevalence data across birth cohorts support an increasing inci- dence. Taken together, these data suggest that BN emerged as a

new syndrome during the latter half of the twentieth century. However, it is possible that perceived trends resulted from the formal recognition of a disorder that had existed as a hidden form of psychiatric morbidity in earlier periods.

Cross-Historical Evidence of BN

Historical cases of bulimia. Several authors have reviewed historical cases of BN prior to its formal recognition (Habermas, 1989; B. Parry-Jones, 1991; B. Parry-Jones & Parry-Jones, 1991; Pope, Hudson, & Mialet, 1985; Stein & Laakso, 1988; Ziolko, 1996). The case histories of Ellen West (Binswanger, 1994, as cited in Stein & Laakso, 1988; Binswanger, 1958, as cited in Pope & Hudson, 1988) and Laura (Lindner, 1954, as cited in Stein & Laakso, 1988) have been interpreted as examples of BN in the earlier half of the twentieth century. However, several experts in the field of eating disorders have reviewed Ellen West as repre- senting AN, and it is unclear that binge eating and purging ever occurred at normal weight in her case (DiNicola, 1990a; Haber- mas, 1992a). Cases described by Wulff (1932, as cited in Haber- mas, 1989 and Stunkard, 1990), Bergmann, (1932, as cited in Habermas, 1992a), and Feuchtinger (1942, as cited in Habermas, 1992a) in the German psychiatric literature have been character- ized by binge eating and purging. According to Habermas (1989), Wulff’s cases included four women and one man, all of whom experienced periods of compulsive eating at normal weight. In three cases, age of onset was during puberty. In two cases, vom- iting was present in addition to dieting. Finally, in three cases, body image disturbance was recorded. However, a different pic- ture emerges from Stunkard’s (1990) translation of Wulff’s work. First, Wulff described marked obesity in two of four cases (Cases A and C). Further, in three of the four cases (Cases A, B, and C), binge eating and fasting or purging occurred in distinct phases. In Case A, for example, the periods of binge eating, hypersomnia, and lack of personal hygiene would persist for months without attempts to counteract the episodes of overeating. These periods are contrasted with periods in which Patient A would eat very little and experience improved mood and increased energy (Stunkard, 1990). Stunkard agreed with Habermas’s (1989) conclusion that these cases represent early instances of “bulimia,” but his use of the term bulimia rather than bulimia nervosa appears to be a deliberate distinction between the syndrome defined in the DSM– III versus that defined in the DSM–III–R. Given that the DSM–III category encompassed what are currently recognized as BN and binge-eating disorder, it is unclear whether Wulff’s cases better represent binge-eating disorder. Echoing this diagnostic uncer- tainty, Habermas (1992a) commented, “Stunkard’s (1990) hypoth- esis of a historical continuity of bulimia probably is not true if it is restricted to bulimia nervosa” (p. 357). Only Patient D was char- acterized by a pattern in which she experienced a binge–purge cycle within one period; these periods alternated with short periods of abstinence and fasting (Wulff, 1932, as translated by Stunkard, 1990).

Pope et al. (1985) reviewed four individuals presented by Pierre Janet in 1903 who appeared to satisfy DSM–III criteria for bulimia on the basis of recurrent binge-eating episodes. However, of these four cases, the combination of binge eating with purging is pre- sented in only one case. This case involved a 17-year-old man, “Ron,” who experienced episodes of “voracious appetite,” “never felt satiety,” (Pope et al., 1985, p. 741) and engaged in self-induced

759CULTURE, GENES, AND EATING DISORDERS

vomiting after the periods of heavy food consumption. The re- maining three cases provide inadequate information to make a differential diagnosis of present-day BN versus AN— binge–purge subtype or binge-eating disorder. Habermas (1991) presented Lud- wig Binswanger’s description of Irma, published in 1909, as “the first known report on a case of bulimia nervosa at normal body weight” (p. 361). The 22-year-old patient engaged in recurrent binge eating and fasting and experienced fear of becoming over- weight. Thus, according to Habermas’s (1989, 1991, 1992a) re- views, cases of BN existed during the first half of the 20th century.

Following their review of Janet’s cases, Pope et al. (1985) concluded that the major psychiatric texts of the nineteenth cen- tury, including those of Esquirol (1838), Briquet (1859), and Laségue (1871; all as cited in Pope et al., 1985), included no apparent cases of DSM–III bulimia. Conversely, Habermas’s (1989) review of Briquet revealed a case of apparent BN in a woman who “ate well” but vomited everything she had eaten and maintained a normal weight (Habermas, 1989, p. 267). Van Deth and Vandereycken (1995) reviewed cases of hysterical vomiting and noted that some cases occurred in individuals of normal weight who also engaged in binge eating and fasting behaviors. Notably, most of these cases occurred in female adolescents. However, the authors equated these cases more with a modern-day conversion disorder, psychogenic vomiting, or AN rather than BN (Van Deth & Vandereycken, 1995). Rosenvinge and Vanderey- cken (1994) reviewed a case of “hysteria” described by Selmer in 1892 in which a 12-year-old girl refused to eat but maintained normal weight. This apparent contradiction was explained one night when the girl’s mother observed her “eating butter, herrings, potatoes and all the food she was able to find in the house” (Selmer as cited in Rosenvinge & Vandereycken, 1994, p. 280). Thus, this girl appeared to fast during the day and binge eat at night (Ros- envinge & Vandereycken, 1994), although it remains unclear how the parents failed to notice the disappearance of so much food earlier in the course of their daughter’s illness. One 1870 report described a 14-year-old girl who would fast for 18 days and then enter a period during which she ate voraciously (B. Parry-Jones & Parry-Jones, 1991). This case resembles those presented by Wulff (Stunkard, 1990) in that extended periods of binge eating alternate with extended periods of fasting rather than there being a cycle of binge eating coupled with inappropriate compensatory behavior. Remaining cases of purported bulimia in the nineteenth century include one 30-year-old man in 1897 who consumed large quan- tities of food day and night and 2 men whose voracious consump- tion included living animals and human flesh, but none engaged in purging or other forms of inappropriate compensatory behavior (B. Parry-Jones & Parry-Jones, 1991).

For the eighteenth century, B. Parry-Jones (1992) detailed the case of Samuel Johnson from 1784 as meeting DSM–III–R criteria for BN. Johnson engaged in binge-eating episodes as they are defined in the DSM–IV, and these caused him to be significantly overweight. To control his weight, he engaged in fasting and used senna as a purging agent. As in cases reported by Wulff (1932, as cited in Stankard, 1990), Johnson’s use of fasting and purging seemed quite time limited compared with his long-standing pattern of compulsive overeating. In 1764, a 16-year-old boy developed a voracious appetite during the course of illness with typhus fever; however, no inappropriate compensatory behavior was noted (B. Parry-Jones & Parry-Jones, 1993; Silverman, 1987).

For the seventeenth century, a 50-year-old man (1678) experi- enced uncontrollable eating followed by vomiting 20 days each year, and, following the 20-day binge–purge cycle, the man fasted for 20 days and then resumed normal eating throughout the re- mainder of the year (B. Parry-Jones & Parry-Jones, 1991; Ziolko, 1996). Robert Whytt (1714 –1766) provided a description of fames canina originally observed by Richard Lower (1631–1691) in the seventeenth century (Silverman, 1987). According to Whytt (1764), Lower observed “an uncommon hunger” among patients with hypochondria and hysteria that produced “a great craving for food” (as cited in Silverman, 1987, p. 145). “In other cases, however, the morbid matter affecting the nerves of the stomach in hypochondriac and hysteric patients, sometimes occasions a want of appetite and a nausea” (Silverman, 1987, p. 145). Although fames canina is supposed to be a disorder characterized by large food intake followed by vomiting (Stein & Laakso, 1988), Whytt’s review of Lower’s observations does not make a clear connection between the presence of “uncommon hunger” and “want of appe- tite and a nausea” in the same patients (Silverman, 1987). B. Parry-Jones and Parry-Jones (1991) reviewed 12 potential cases of BN from the seventeenth to nineteenth centuries. Five of these cases were described above. Among the remaining 7 cases, none were associated with inappropriate compensatory behavior, and parasitic worms were found in 4 cases. As cited in Ziolko (1996), Forestus (1602) described a nun afflicted with canine appetite (fames canina or kynorexia) who was miraculously cured after several unsuccessful medicinal treatments by physicians. Like fames canina, kynorexia was defined by insatiable appetite, eating that is out of control, and then compulsive vomiting as a result of excessive food intake (Stein & Laakso, 1988; Ziolko, 1996).

From the twelfth to the seventeenth centuries, many of the fasting saints were reported to engage in binge eating (e.g., St. Veronica) and self-induced vomiting (e.g., St. Catherine; Bell, 1985; Rampling, 1985). However, these cases appear to fall within the diagnosis of AN— binge–purge subtype (DSM–IV–TR). It is unclear whether there were cases of binge eating and purging among normal-weight women seeking spiritual purification through fasting. If these cases were common in convents, there is little reason to think that they would go unnoticed given the religious significance of women’s eating during medieval times (Bynum, 1987).

In the eighth century A.D., Avicenna prescribed self-induced vomiting to undo the ill effects of overeating; however, he warned,

To procure vomiting to an undue degree is injurious for the stomach, it is also prejudicial to the thorax and the teeth . . . and may lead to consumption. The custom of some people who eat to excess and then procure vomiting, is one of the things that end in a chronic disorder. (Gruner, 1930, as cited in Nasser, 1993, p. 130)

This suggests awareness of the morbidity associated with chronic self-induced vomiting and that a disorder characterized by over- eating and self-induced vomiting was known in Arabic medicine. However, no further information was provided concerning the individuals with this disorder or whether the disorder was main- tained at normal weight.

For cases prior to the second century A.D., Crichton (1996) speculated as to whether the Roman emperors Claudius (A.D. 41–54) and Vitellius (circa A.D. 69) suffered from BN. Notably, Ziolko (1996) rejected these cases on the basis that excessive food intake appeared to be based in intentional gluttony with self-

760 KEEL AND KLUMP

induced vomiting used as a means to allow continued consump- tion. However, excerpts from Suetonius’s description of Vitellius suggest that the emperor’s excessive food intake may not have been under his control: “He was a man of not only such extreme and impulsive, but also disgusting, gluttony that he could not even curb it during a sacrifice or on a journey” (translated in Crichton, 1996, p. 204). Crichton noted that vomiting distinguished Claudius and Vitellius from their historical peers; however, Seneca is known to have commented, “vomunt ut edant, edunt ut vomant [they vomit that they may eat, they eat that they may vomit]” (as cited in Crichton, 1996, p. 206). This suggests that binge eating and purging may have represented a common behavioral pattern among the elite in the Roman Empire, and, in the case of Vitellius, a not entirely volitional pattern of eating.

Interpretation of historical cases of bulimia. In our review of historical cases of AN, we found numerous examples of young women engaging in self-starvation. The overt behaviors, conse- quences, course, and affected population resembled that for modern-day AN. Thus, the debate centered mostly on the motiva- tion behind food refusal. In contrast, evidence of a binge–purge syndrome outside of its present historical context is quite sparse. Historical accounts of bulimia do not seem to preponderate in adolescent or young adult females. In fact, prior to the nineteenth century, cases involved mostly adult men. Further, most historical cases of bulimia represented the syndrome defined in the DSM–III that did not require inappropriate compensatory behavior. Thus, in our attempt to find historical accounts of BN, we seemed to find numerous examples of binge-eating disorder.

According to the restraint hypothesis (Polivy & Herman, 1985), one would expect that the fasting evident throughout the twelfth to the nineteenth centuries would result in binge eating for a number of young women. Consistent with this hypothesis, there were isolated cases in which individuals seemed to engage in binge eating after a period of restricted food intake. However, the use of purging (or other forms of inappropriate compensatory behaviors) to counteract or undo such episodes was lacking in most reported cases of binge eating. This difference may be explained by the motivation behind food restriction. In modern times, food restric- tion often is intended to achieve weight loss. When a binge episode occurs, purging is motivated by the belief that it will prevent weight gain. Conversely, if fasting is interpreted in a religious framework, then purging cannot prevent the sin of gluttony once the binge episode has occurred (Bynum, 1987). Bynum (1987) related the story of Friderade, who suffered from voracious appe- tite and “grew enormous” (p. 89). After confessing, she was cured and did not eat for the following 3 years (Bynum, 1987). This story provides a template for most of the historical cases of bulimic syndromes—a period of recurrent binge eating is followed by a period of abstinence. However, this behavioral pattern resembles that of a person with binge-eating disorder in which extended periods of binge eating alternate with periods of dieting.

Our review of AN suggested that specific motivations related to weight concerns did not seem to be necessary to produce the syndrome. We raised the question of whether weight concerns truly cause AN or represent an attempt to understand the disorder by both the individual suffering from AN (Wegner, 2002) and her culture. Conversely, our review suggests that a binge–purge syn- drome predominantly affecting normal-weight women may not emerge in the absence of weight concerns.

Cross-Cultural Evidence for BN

Characterizing the actual prevalence of BN outside of a Western context has proven challenging. For example, across several re- views of eating disorders cross-culturally (Davis & Yager, 1992; Dolan, 1991; Miller & Pumariega, 2001; Pate, Pumariega, Hester, & Garner, 1992; Tsai, 2000), references to AN far outnumbered references to BN. This difference is also reflected in the range of countries reporting BN among their citizens; in our review, we found reports of AN in five of five non-Western regions of the world (see Table 2), whereas BN was reported in only three of five non-Western regions (see Table 4). Similar to the construction of Table 2, Table 4 indicates the country in which cases were found, diagnostic criteria used, number or prevalence of cases, and degree of Western influence. We eliminated the column concerning pres- ence of weight concerns because our review of cases indicated that BN was accompanied by weight concerns in all cases. As in the analysis shown in Table 2, weight concern and Western influence were established by consensus.

We found no studies reporting the presence of BN in an indi- vidual with no exposure to Western ideals. All of the BN cases that could be categorized involved individuals who were probably or definitely exposed to Western ideals and values through urbaniza- tion, English-medium schools, previous residence in Western na- tions, and/or higher socioeconomic status or educational attain- ment. If BN only emerges in non-Western countries as a result of Western influences, then this may explain why there does not appear to be a non-weight-concerned form of BN.

With the exceptions of Japan and a recent study in Iran, prev- alence estimates of full BN in non-Western nations were below the range reported for Western nations in the DSM–IV–TR. Degree of Western influence may account for a good deal of variance in BN prevalence estimates. For example, following World War II, Japan has emulated some of the ideals of the United States. Indeed, the Cassell Dictionary of Modern Politics (East & Joseph, 1994) indicates that Japan is sometimes included in the definition of the West. Thus, it is not surprising that BN prevalence is the most similar between the United States and this non-Western nation. Epidemiological data for BN in non-Western nations suggest that BN has a lower prevalence than AN in these countries, a pattern that is the opposite of that observed in Western nations. The point prevalence of AN has been found to be four to five times greater than full-syndromal BN in Japan (Nakamura et al., 2000), and it has been noted that AN is more prevalent than BN in Hong Kong (Chen et al., 1993; S. Lee, Hsu, & Wing, 1992). Thus, even when BN is found in non-Western nations, it is not found in the absence of Western influence, and it seems to be less common than AN.

Summary of Findings for BN

It appears to be the case that BN is a culture-bound syndrome. Although epidemiological data for BN are limited, they support a large and significant increase in BN incidence during the latter half of the twentieth century. Our attempts to find evidence of BN in earlier historical periods were largely unsuccessful. Most historical cases of recurrent binge eating seem to represent binge-eating disorder (because of a lack of inappropriate compensatory behav- ior) or AN— binge–purge subtype (because of the presence of low weight). Finally, although BN does exist in non-Western nations, we were unable to find evidence of the disorder arising in the absence of Western influence.

761CULTURE, GENES, AND EATING DISORDERS

Discussion

Divergent patterns emerged in our examination of epidemiolog- ical, cross-historical, and cross-cultural evidence of AN and BN as culture-bound syndromes. Unlike previous conclusions that eating disorders either are culture bound or are not culture bound (Haber- mas, 1989; S. Lee et al., 1993), our findings suggest that BN is culture bound and that AN is not.

Reasons for the differences in findings between AN and BN could be attributed to our definition of the phenotypes. We sought

evidence of AN and BN according to their conceptualization within the DSM–IV. However, the DSM–IV does not represent “nature carved at its joints.” If we had allowed a more generous definition of BN to include all syndromes characterized by recur- rent binge eating, the conclusions of our review would be signif- icantly altered as the symptom of binge eating appeared in more historical and culture contexts than the syndrome of BN. We chose to follow the DSM–IV conceptualization of these syndromes be- cause it has the greatest empirical support. Of interest, the results of our review seem to add support for distinguishing between the

Table 4 Studies of Bulimia Nervosa (BN) in Non-Western Nations

Article Place Diagnostic

criteria Number–prevalence Cases without

Western influence?

Middle Eastern–Arab Cultures

Ford (1992) Egypt DSM–III–R 1 case reported; developed BN upon returning to Egypt after living in Canada for 12 years

No

Nasser (1994) Cairo, Egypt Russell (1979) 0.9% full BN; 3.4% partial syndrome BN in 351 high school students

Unlikely

Nobakht & Dezhkam (2000) Tehran, Iran DSM–IV 3.2% of 3,100 female high school students ages 15–18 Unlikely Choudry & Mumford (1992) Pakistan DSM–III–R 0.4% full BN in 271 Urdu medium school girls ages 12–16 Unlikely Mumford et al. (1992) Pakistan DSM–III–R 0.002% full BN, and 0.01% subthreshold BN in 369

English-medium school girls ages 14–16 Unlikely

Southeast Asia

Ong et al. (1982) Malaysia Not given 1 case of nonpurging BN reported in a woman during recovery from AN

Unlikely

Schmidt (1993) Malaysia DSM–III–R 1 case in a male; vomiting–weight concerns began in Malaysia, binge eating in England

Unlikely

Goh et al. (1993) West Malaysia Study specific 1 case reported Indeterminable Ung et al. (1997) Malaysia Not given 16 cases of “bingeing and/or purging behavior in subjects

of normal or above normal weight” (p. 332) Indeterminable

East Asia

Nogami et al. (1984) Japan Unknown 17 individuals with both binge eating and purging of 846 university students ages 18–30

Unlikely

Kiriike et al. (1988) Japan DSM–III and Russell (1979)

2.9% of 456 college women report bingeing and purging at least twice per week

Unlikely

Takeda et al. (1993)a Japan DSM–III–R 1.9% high school girls Unlikely Kuboki et al. (1996) Japan Not given 1.3–2.5 per 100,000 general population in 1992 Indeterminable Nadaoka et al. (1996) Japan DSM–III–R 38 cases treated at Yamagata University Hospital between

1978 and 1992 Unlikely

Nakamura et al. (2000) Japan DSM–IV 1.02 per 100,000 females Indeterminable Takei et al. (2000)b Japan DSM–IV 321 students with BN from 335 college health

administration centers across Japan Unknown

Chun et al. (1992) China DSM–III–R 1.1% of 509 freshman medical college students (401 female)

Unlikely

S. Lee, Hsu, & Wing (1992) Hong Kong DSM–III–R 4 cases treated at Department of Psychiatry, University of Hong Kong, between 1980 and 1988

Unlikely

S. Lee (1993) Hong Kong DSM–III–R 0.3% partial BN syndromes in 1,020 bilingual university students

Unlikely

S. Lee, Lee, & Leung (1998) Hong Kong DSM–III–R 17 consecutive series of patients treated from 1990–1996 Unlikely Lam & Lee (2000) Hong Kong DSM–IV Report on 30 cases of BN between 1984 and 1998;

provided detailed case histories of 3 patients among whom only 1 seemed to have DSM–IV BN

Indeterminable; unlikely for 3 detailed cases

Tseng et al (1989)b Taiwan DSM–III 5 patients reported Unknown Hung & Cheng (1992)a Taipei, Taiwan DSM–III–R 5.6% in 12,435 females ages 12–25 Indeterminable

Note. Not given � authors did not provide the information in the article; Unknown � authors may have provided indicated information in the non-English publication, but the information was not provided in the English abstract or the English review article summarizing it; Study specific � diagnostic criteria were specific to the study indicated; Indeterminable � exposure to Western influence could not be determined from article; DSM–III–R � Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.); DSM–IV � Diagnostic and Statistical Manual of Mental Disorders (4th ed.); DSM–III � Diagnostic and Statistical Manual of Mental Disorders (3rd ed.). a As reported in Tsai (2000). b Data were taken from an English abstract of a non-English publication.

762 KEEL AND KLUMP

two syndromes. Further support comes from a recent study of the influence of Western media exposure on rates of disordered eating in Fiji (Becker, Burwell, Gilman, Herzog, & Hamburg, 2002). In 1995 (within a month of the introduction of television), 7.9% of Fijian girls reported binge eating. In 1998 (3 years after the introduction of television), the percentage had decreased nonsig- nificantly to 4.6%. Conversely, the percentage of girls who re- ported self-induced vomiting increased significantly from 0% to 11.3% over the same period (Becker et al., 2002). Prior to study initiation, one case of AN had been reported in Fiji (Becker et al., 2002). These data suggest that AN and binge eating in the absence of inappropriate compensatory behavior may not be particularly dependent on exposure to Western ideals. However, a syndrome characterized by the combination of binge eating and purging at normal weight may be more culture bound.

Although the evidence leads us to a strong conclusion that AN has existed outside of its current sociohistorical context, the con- clusion for BN is necessarily weaker as there were fewer articles concerning BN than AN. This may represent the relatively longer recognition of AN as a syndrome (and thus greater time to accu- mulate publications on AN); however, the majority of the articles reviewed have been published since the introduction of BN to the psychiatric nomenclature. Further, this publication pattern is the inverse of that observed for the treatment literature in AN versus BN (Peterson & Mitchell, 1999). Thus, it remains unclear whether the recent recognition of BN is the cause or result of limited data concerning its existence. The limited data for BN may not reflect the absence of the disorder because BN may have existed but eluded detection. This seems plausible because there is no overt sign of BN such as the emaciation that characterizes AN. Indeed, Whitehouse et al. (1992) found that 50% of cases of BN deter- mined in the community were unknown to their general practitio- ners, despite referrals made for complications of bulimic pathology in half of these “hidden” cases. No such hidden cases of AN were found, suggesting that the low weight of AN likely makes AN easier to detect outside of the confines of a recognized syndrome. However, there are reasons why concerns about detection of BN may be overstated. Further, these reasons contribute to the under- standing of why BN may represent a more culturally bound syn- drome than AN.

Although self-starvation can occur in any context, binge eating requires large stores of readily edible food. Thus, cases may be validly limited to places and periods with abundant food, such as the palaces of the Roman Empire, places of communal living with shared kitchens, or affluent households in Victorian England. In earlier historical contexts, food consumed a greater proportion of household budgets, and distribution was closely monitored with greater portions going to men who were engaging in manual labor (Bailey & Earle, 1999). Indeed, gender differences in access to food may explain why cases of binge eating prior to the nineteenth century more commonly involved men (B. Parry-Jones & Parry- Jones, 1991). Thus, whereas individuals can deliberately starve themselves regardless of access to food—in some cases refusing even the meager portions allotted to them—individuals cannot binge eat without free access to large quantities of food. If large amounts of food were obtainable, it is difficult to believe that recurrent binge episodes would go unnoticed because food was not obtained in anonymity as it is today. Certainly, the food intake of Samuel Johnson was not only noticed but recorded by several individuals during and following his lifetime (B. Parry-Jones,

1992) as were the binge-eating episodes of St. Veronica (Bell, 1985). Indeed, the presence of binge eating as a symptom has been recorded both cross-historically and cross-culturally with the pri- mary constraint being one of access to food. Similar to difficulties in attempting to binge eat without detection, purging would be a difficult behavior to hide prior to the wide availability of modern plumbing. The ability to flush away evidence of self-induced vomiting, or laxative or diuretic abuse, from the privacy of one’s indoor bathroom would greatly facilitate secretive purging. Cham- ber pots and outhouses would not. Further, accounts of fasting and excessive exercise are plentiful among historical cases of fasting girls, even those of normal weight, suggesting that nonpurging forms of inappropriate compensatory behavior would not remain hidden for long. Although no obvious sign characterizes BN, it seems unlikely that BN would go unnoticed in earlier historical periods. Thus, the relative dearth of evidence of BN outside of its sociohistorical cultural context may reflect its lack of existence. Of interest, an ability to obtain large amounts of food inconspicuously may explain why BN, but not AN, is more common in urban than rural areas (Hoek et al., 1995; Robins et al., 1984).

Binge eating may be constrained by access to large quantities of food, and purging may be relatively limited to a context in which prevention of weight gain is culturally meaningful. Thus, unlike AN, which can occur in numerous contexts, the specific combi- nation of binge eating and purging may be limited by two factors specific to modern Western cultures—access to food and use of purging to prevent weight gain. This would suggest greater cross- historical and cross-cultural representation of the restricting sub- type of AN compared with the binge–purge subtype. Of interest, a clinical difference between S. Lee et al.’s (1993) patients with fat phobic and non-fat phobic AN was the increased presence of binge–purge symptoms in the patients with fat phobia.

Conceptualizing the Genetic Bases of Eating Disorders

Sociocultural and cognitive– behavioral models of eating disor- ders emphasize the role of weight concerns in their etiology, whereas behavioral genetic research emphasizes the role of genes. As we stated in the introduction, the purpose of our research was to understand the extent to which eating disorders represent culture-bound syndromes and the implications of our findings for conceptualizing their genetic bases. Our review suggests that BN may be a culture-bound syndrome, influenced by weight concerns, anonymous access to large quantities of food, and a motivation to prevent the effects of binge eating on weight through the use of inappropriate compensatory behavior. Conversely, our review sug- gests that weight concerns can influence the incidence of AN but that whatever cultural influences contribute to the etiology of AN, they are not particularly limited in their distribution across history or cultures.

Although recent behavioral genetic research has provided sub- stantial evidence of genetic influences on the development of both AN (Klump et al., 2001; Kortegaard et al., 2001; Wade et al., 2000) and BN (Bulik et al., 1998, 2000; Kendler et al., 1991, 1995), heritability estimates are constrained to the cultures from which study samples are drawn. In cultures in which the thin ideal is ubiquitous, these kinds of environmental factors are held rela- tively constant across individuals and thus cannot account for individual differences in eating disorder development. In cultures in which relevant environmental factors vary across individuals,

763CULTURE, GENES, AND EATING DISORDERS

cultural pressures to be thin have the potential to play a much larger role in individual differences in the development of eating pathology. As such, heritability estimates of BN are more likely to vary across cultures. Within a culture in which there is large variability in exposure to the thin ideal and access to food, genetic estimates may be quite low for BN. Conversely, heritability esti- mates of AN are likely to be more stable. Current data do not allow us to examine this hypothesis directly because twin studies of AN and BN have been conducted within industrialized Western na- tions. Future research should seek to examine this question more closely using, whenever possible, twins from non-Western coun- tries or combining samples from Western and non-Western countries.

Because heritability estimates are statistical entities that gauge the extent to which individual differences within a population can be explained by genetic influences, they do not directly index the action of genes. Although these statistical estimates can vary depending on environmental contexts (with the potential for BN to have low heritability in some contexts), this variation does not mean that genes do not influence BN or that they influence BN less than AN. The genetic diathesis for BN may exhibit more patho- plasticity cross-culturally than the genetic diathesis for AN. When the diathesis for a disorder is general or has high pathoplasticity, the disorder may not be evident cross-culturally or cross- historically. That is, although the genetic diathesis may exist throughout history and cross-culturally, when tapped, it presents as a different form of pathology conforming to its current sociohis- torical context. In contrast, when the genetic diathesis for a disor- der is specific or has low pathoplasticity, evidence of the disorder should be available cross-culturally and cross-historically. The disorder may be less common in different sociohistorical periods, but whenever the diathesis is tapped, it should produce the same narrow syndrome.

In past historical periods, the genetic predisposition for what is now BN may have produced different forms of psychopathology, such as hysteria in the Victorian era. The demographic features of those afflicted with these two illnesses are quite similar, and the abrupt emergence of these syndromes is also quite similar. Pope and Hudson (1988) have long viewed eating disorders as lying along an etiologic spectrum with numerous other syndromes. The extent to which BN represents a culturally shaped manifestation of a general genetic diathesis may explain why this disorder, but not AN, responds to the same medications and psychosocial treatments that are used to treat major depressive disorder (MDD) and obsessive– compulsive disorder (OCD; Peterson & Mitchell, 1999). Specifically, both antidepressant medications and cognitive– behavioral therapy have demonstrated efficacy in treat- ing BN but not AN (Peterson & Mitchell, 1999). Possibly the genetic diathesis to BN can be expressed as hysteria, BN, MDD, OCD, or any combination of these depending on a variety of environmental factors. Twin studies have the ability to examine this question indirectly by investigating whether disorders share genetic transmission. If there is a high degree of shared transmis- sion among disorders, then the disorders may represent different expressions of a genetic diathesis with high pathoplasticity. An early twin investigation of BN found shared genetic transmission between this disorder and major depression (Walters et al., 1992). However, when examined with a larger, more powerful multivar- iate model including six psychiatric disorders (BN, phobias, gen- eralized anxiety disorder, panic disorder, major depression, and

alcoholism), BN loaded on a single genetic factor with phobias and panic disorder (Kendler et al., 1995). Findings from this study suggest that BN may indeed share genetic transmission with other forms of “neuroses” and that this shared transmission may be indicative of greater pathoplasticity of the disorder. To date, no study has examined AN using multivariate, genetic analyses. Al- though one study found some common genetic risk factors be- tween AN and major depression (Wade et al., 2000), findings were limited by the inclusion of only one comorbid disorder. Thus, future twin research can evaluate the pathoplasticity associated with the genetic diatheses for AN and BN.

Conclusion

Previous work has posed modern Western idealization of thin- ness as a common etiologic factor for AN and BN. However, our review reveals that the epidemiology, history, and cultural distri- bution of these disorders are distinct. Researchers interested in revealing the pathophysiology of these disorders would do well to attend to these patterns, as they suggest distinct hypotheses con- cerning variability of heritability estimates and pathoplasticity. If our hypotheses are supported in future behavior genetic research, then two lines of evidence would suggest that some or all of the genes contributing to the development of BN differ from those contributing to the development of AN. This would provide the greatest evidence, thus far, for etiologic validity in distinguishing between these two syndromes. The extent to which AN and BN are linked in the medical, academic, and popular press may represent a historical coincidence more than a true relationship between these disorders.

References

References marked with an asterisk indicate studies included in the meta-analysis. Abou-Saleh, M. T., Younis, Y., & Karim, L. (1998). Anorexia nervosa in

an Arab culture. International Journal of Eating Disorders, 23, 207– 212.

Al-Subaie, A. S. (2000). Some correlates of dieting behavior in Saudi schoolgirls. International Journal of Eating Disorders, 28, 242–246.

Al-Subaie, A., Sulaiman, A., Bamgboye, E., Al-Sabhan, K., Sulaiman, A., & Bannah, A. R. (1996). Validity of the Arabic version of the eating attitude test. International Journal of Eating Disorders, 20, 321–324.

American Psychiatric Association. (1980). Diagnostic and statistical man- ual of mental disorders (3rd ed.). Washington, DC: Author.

American Psychiatric Association. (1987). Diagnostic and statistical man- ual of mental disorders (3rd ed., rev.). Washington, DC: Author.

American Psychiatric Association. (1994). Diagnostic and statistical man- ual of mental disorders (4th ed.). Washington, DC: Author.

American Psychiatric Association. (2000). Diagnostic and statistical man- ual of mental disorders (4th ed., text rev.). Washington, DC: Author.

Azuma, Y., & Henmi, M. (1982). A study on the incidence of anorexia nervosa in schoolgirls. Annual Report of Research Group in Eating Disorders, 30 –34.

Bailey, R., & Earle, M. (1999). Home cooking to takeaways. Changes in food consumption in New Zealand during 1880 –1990 (2nd ed.). Pal- merston North, New Zealand: Massey University, Department of Food Technology.

Banks, C. G. (1992). “Culture” in culture-bound syndromes: The case of anorexia nervosa. Social Science & Medicine, 34, 867– 884.

Banks, C. G. (1997). The imaginative use of religious symbols in subjec- tive experiences of anorexia nervosa. Psychoanalytic Review, 2, 227– 236.

764 KEEL AND KLUMP

Becker, A. E., Burwell, R. A., Gilman, S. E., Herzog, D. B., & Hamburg, P. (2002). Eating behaviours and attitudes following prolonged exposure to television among ethnic Fijian adolescent girls. British Journal of Psychiatry, 180, 509 –514.

Bell, R. M. (1985). Holy anorexia. Chicago: University of Chicago Press. Bemporad, J. R. (1996). Self-starvation through the ages: Reflections on

the pre-history of anorexia nervosa. International Journal of Eating Disorders, 19, 217–237.

Beumont, P. J. V. (1988). Bulimia: Is it an illness entity? International Journal of Eating Disorders, 7, 167–176.

Bliss, E. L., & Branch, C. H. H. (1960). Anorexia nervosa. New York: Hoeber.

Branson, R., Potoczna, N., Kral, J. G., Lentes, K.-U., Hoehe, M. R., & Horber, F. F. (2003). Binge eating as a major phenotype of melanocor- tin 4 receptor gene mutations. New England Journal of Medicine, 348, 1096 –1103.

Brumberg, J. J. (1989). Fasting girls: A history of anorexia nervosa. New York: Plume.

Buchan, T., & Gregory, L. D. (1984). Anorexia nervosa in a Black Zimbabwean. British Journal of Psychiatry, 145, 326 –330.

Buhrich, N. (1981). Frequency of presentation of anorexia nervosa in Malaysia. Australian and New Zealand Journal of Psychiatry, 15, 153– 155.

Bulik, C. M., Sullivan, P. F., & Kendler, K. S., (1998). Heritability of binge-eating and broadly defined bulimia nervosa. Biological Psychia- try, 44, 1210 –1218.

Bulik, C. M., Sullivan, P. F., Wade, T. D., & Kendler, K. S. (2000). Twin studies of eating disorders: A review. International Journal of Eating Disorders, 27, 2–20.

Burnet, P. W., Smith, K. A., Cowen, P. J., Fairburn, C. G., & Harrison, P. J. (1999). Allelic variation of the 5-HT2C receptor (HTR2C) in bulimia nervosa and binge eating disorder. Psychiatric Genetics, 9, 101–104.

Bushnell, J. A., Wells, J. E., Hornblow, A. R., Oakley-Browne, M. A., & Joyce, P. (1990). Prevalence of three bulimia syndromes in the general population. Psychological Medicine, 20, 671– 680.

Bynum, C. W. (1987). Holy feast and holy fast. The religious significance of food to medieval women. Berkeley: University of California Press.

Bynum, C. W. (1988). Holy anorexia in modern Portugal. Culture, Med- icine, and Psychiatry, 12, 239 –248.

Cachelin, F. M., & Maher, B. A. (1998). Is amenorrhea a critical criterion for anorexia nervosa? Journal of Psychosomatic Research, 44, 435– 440.

Carroll, L. I. (1998). Holy anorexia revisited: The reputation of fasting in the case of Maria Janis. Psychohistory Review, 26, 115–136.

Chadda, R., Malhotra, S., Asad, A. G., & Bambery, P. (1987). Sociocul- tural factors in anorexia nervosa. Indian Journal of Psychiatry, 29, 107–111.

Chakraborty, A. (1985). [Letter to the editor]. Transcultural Psychiatric Research Review, 22, 210 –211.

Chandra, P. S., Shaw, A., Shenoy, J., Kumar, U., Varghese, M., Bhatti, R. S., & Channabasavanna, S. M. (1995). Family pathology and anorexia in the Indian context. International Journal of Social Psychiatry, 41, 292–298.

Chen, C.-N., Wong, J., Lee, N., Chan-Ho, M-W., Lau, J. T.-F., & Fung, M. (1993). The Shantin community mental health survey in Hong Kong: II. Major findings. Archives of General Psychiatry, 50, 125–133.

Chipley, W. J. (1860). On sitomania. Journal of Psychological Medicine and Mental Pathology, 13, 266 –270.

Chiu, L. P. W. (1989). Anorexia nervosa in a young Chinese woman in Hong Kong. Canadian Journal of Psychiatry, 34, 162–163.

Choudry, I. Y., & Mumford, D. B. (1992). A pilot study of eating disorders in Mirpur (Pakistan) using an Urdu version of the Eating Attitudes Test. International Journal of Eating Disorders, 11, 243–251.

Chun, Z. F., Mitchell, J. E., Li, K., Yu, W. M., Lan, Y. D., Jun, Z., et al. (1992). The prevalence of anorexia nervosa and bulimia nervosa among

freshman medical college students in China. International Journal of Eating Disorders, 12, 209 –214.

Crichton, P. (1996). Were the Roman emperors Claudius and Vitellius bulimic? International Journal of Eating Disorders, 19, 203–207.

Davis, C., & Yager, J. (1992). Transcultural aspects of eating disorders: A critical literature review. Culture, Medicine, and Psychiatry, 16, 377– 394.

DeAngelis, T. (2002). A genetic link to anorexia. Monitor on Psychol- ogy, 33, 34 –36.

DiNicola, V. F. (1990a). Anorexia multiforme: Self-starvation in historical and cultural context: I. Self-starvation as a historical chameleon. Transcultural Psychiatric Research Review, 27, 165–196.

DiNicola, V. F., (1990b). Anorexia multiforme: Self-starvation in historical and cultural context: II. Anorexia nervosa as a culture-reactive syn- drome. Transcultural Psychiatric Research Review, 27, 245–286.

Doerr-Zegers, O. (1994). About a particular type of oral perversion in the female: Hyperphagia followed by vomiting. International Journal of Eating Disorders, 16, 117–132.

Dolan, B. (1991). Cross-cultural aspects of anorexia nervosa and bulimia: A review. International Journal of Eating Disorders, 10, 67–78.

Drewnowski, A., Yee, D. K., & Krahn, D. D. (1988). Bulimia in college women: Incidence and recovery rates. American Journal of Psychiatry, 145, 753–755.

*Eagles, J. M., Johnston, M. I., Hunter, D., Lobban, M., & Millar, H. R. (1995). Increasing incidence of anorexia nervosa in the female popula- tion of Northeast Scotland. American Journal of Psychiatry, 152, 1266 – 1271.

East, R., & Joseph, T. (1994). Cassell dictionary of modern politics. London: Cassell.

Elsarrag, M. E. (1968). Psychiatry in the Northern Sudan: A study in comparative psychiatry. British Journal of Psychiatry, 114, 945–948.

Fahy, T. A., Robinson, P. H., Russell, G. F. M., & Sheinman, B. (1988). Anorexia nervosa following torture in a young African woman. British Journal of Psychiatry, 153, 385–387.

Fairburn, C. G., & Beglin, S. J. (1990). Studies of the epidemiology of bulimia nervosa. American Journal of Psychiatry, 147, 401– 408.

Fairburn, C., Cooper, Z., & Shafran, R. (2002, April). A new “transdiag- nostic” cognitive behavioral treatment for eating disorders. Paper pre- sented at the International Conference on Eating Disorders, Boston.

Famuyiwa, O. O. (1988). Anorexia nervosa in two Nigerians. Acta Psy- chiatrica Scandinavica, 78, 550 –554.

Fombonne, E. (1995). Anorexia nervosa. No evidence of an increase. British Journal of Psychiatry, 166, 462– 471.

Fombonne, E. (1996). Is bulimia nervosa increasing in frequency? Inter- national Journal of Eating Disorders, 19, 287–296.

Ford, K. A. (1992). Bulimia in an Egyptian student: A case study. Inter- national Journal of Eating Disorders, 11, 407– 411.

Gandhi, D. H., Appaya, M. P., & Machado, T. C. (1991). Anorexia nervosa in Asian children. British Journal of Psychiatry, 159, 591–592.

Garner, D. M., Garfinkel, P. E., Schwartz, D., & Thompson, M. (1980). Cultural expectations of thinness in women. Psychological Reports, 47, 483– 491.

George, D. T., Weiss, S. R., Gwirtsman, H. E., & Blazer, D. (1987). Hospital treatment of anorexia nervosa: A 25 year retrospective study from 1958 to 1982. International Journal of Eating Disorders, 6, 321– 330.

German, G. A. (1972). Aspects of clinical psychiatry in sub-Saharan Africa. British Journal of Psychiatry, 121, 461– 479.

Goh, S. E., Ong, S. B. Y., & Subramaniam, M. (1993). Eating disorders in Hong Kong. British Journal of Psychiatry, 162, 276 –277.

Götestam, K. G., Eriksen, L., Heggestad, T., & Nielsen S. (1998). Preva- lence of eating disorders in Norwegian general hospitals 1990 –1994: Admissions per year and seasonality. International Journal of Eating Disorders, 23, 57– 64.

Grice, D. E., Halmi, K. A., Fichter, M. M., Strober, M., Woodside, D. B.,

765CULTURE, GENES, AND EATING DISORDERS

Treasure, J. T., et al. (2002). Evidence for a susceptibility gene for anorexia nervosa on Chromosome 1. American Journal of Human Ge- netics, 70, 787–792.

Gull, W. W. (1868). The address in medicine delivered before the annual meeting of the British Medical Association, at Oxford. Lancet, ii, 171– 176.

Gull, W. W. (1874). Anorexia nervosa (apepsia hysterica, anorexia hys- terica). Transactions of the Clinical Society of London, 7, 22–28.

Gull, W. (1888). Anorexia nervosa. Lancet, i, 516 –517. Habermas, T. (1989). The psychiatric history of anorexia nervosa and

bulimia nervosa: Weight concerns and bulimic symptoms in early case reports. International Journal of Eating Disorders, 8, 259 –273.

Habermas, T. (1991). The role of psychiatric and medical traditions in the discovery and description of anorexia nervosa in France, Germany, and Italy, 1873–1918. Journal of Nervous and Mental Disease, 179, 360 – 365.

Habermas, T. (1992a). Further evidence on early case descriptions of anorexia nervosa and bulimia nervosa. International Journal of Eating Disorders, 11, 351–359.

Habermas, T. (1992b). Historical continuities and discontinuities between religious and medical interpretations of extreme fasting: The background to Giovanni Brugnoli’s description of two cases of anorexia nervosa in 1875. History of Psychiatry, 3, 431– 455.

Habermas, T. (1996). In defense of weight phobia as the central organizing motive in anorexia nervosa: Historical and cultural arguments for a culture-sensitive psychological conception. International Journal of Eating Disorders, 19, 317–334.

Hajal, F. (1982). Psychological treatment of anorexia: A case from the ninth century. Journal of the History of Medicine & Allied Sciences, 37, 325–328.

*Hall, A., & Hay, P. J. (1991). Eating disorder patient referrals from a population region 1977–1986. Psychological Medicine, 21, 697–701.

Halmi, K. A. (2003). Diagnosis, classification and comorbidities of eating disorders: A review. In M. Mario, K. Halmi, J. J. López-lbor, & N. Sartorius (Eds.), Evidence and experience in psychiatry (Vol. 6, pp. 31–33). Chichester, England: Wiley.

Halmi, K. A., Falk, J. R., & Schwartz, E. (1981). Binge-eating and vomiting: A survey of a college population. Psychological Medicine, 11, 697–706.

Herzog, D. B., Dorer, D. J., Keel, P. K., Selwyn, S. E., Ekeblad, E. R., Flores, A. T., et al. (1999). Recovery and relapse in anorexia and bulimia nervosa: A 7.5-year follow-up study. Journal of the American Academy of Child & Adolescent Psychiatry, 38, 829 – 837.

Hoek, H. W. (1991). The incidence and prevalence of anorexia nervosa and bulimia nervosa in primary care. Psychological Medicine, 21, 455– 460.

Hoek, H. W. (1993). Review of the epidemiological studies of eating disorders. International Review of Psychiatry, 5, 61–74.

*Hoek, H. W., Bartelds, A. I. M., Bosveld, J. J. F., van der Graaf, Y., Limpens, V. E. L., Maiwald, M., & Spaaij, C. J. K. (1995). Impact of urbanization on detection rates of eating disorders. American Journal of Psychiatry, 152, 1272–1278.

Hoek, H. W., Van Hoeken, D., & Katzman, M. A. (2003). Epidemiology and cultural aspects of eating disorders: A review. In M. Mario, K. Halmi, J. J. López-lbor, & N. Sartorius (Eds.), Evidence and experience in psychiatry (Vol. 6, pp. 75–104). Chichester, England: Wiley.

Hooper, M. S. H., & Garner, D. M. (1986). Application of the Eating Disorders Inventory to a sample of Black, White, and mixed race schoolgirls in Zimbabwe. International Journal of Eating Disorders, 5, 161–168.

Hung, C., & Cheng, S. (1992). Body image and eating behavior in young females in Taipei. Journal of the National Public Health Association, Republic of China, 11, 316 –327.

*Joergensen, J. (1992). The epidemiology of eating disorders in Fyn County, Denmark, 1977–1986. Acta Psychiatrica Scandinavica, 85, 30 –34.

*Jones, D. J., Fox, M. M., Babigian, H. M., & Hutton, H. E. (1980). Epidemiology of anorexia nervosa in Monroe County, New York: 1960 –1976. Psychosomatic Medicine, 42, 551–558.

Kam, W. K., & Lee, S. (1998). The variable manifestations and contextual meanings of anorexia nervosa: Two case illustrations from Hong Kong. International Journal of Eating Disorders, 23, 227–231.

Karan, P. P. (in press). The non-Western world: Development, environment and human rights. New York: Routledge.

Katzman, M. A., & Lee, S. (1997). Beyond body image: The integration of feminist and transcultural theories in the understanding of self starvation. International Journal of Eating Disorders, 22, 385–394.

Keel, P. K. (2003). Validity of categorical distinctions for eating disorders: From disorders to symptoms. In M. Mario, K. Halmi, J. J. López-lbor, & N. Sartorius (Eds.), Evidence and experience in psychiatry (Vol. 6, pp. 52–54). Chichester, England: Wiley.

Keel, P. K., Crow, S. J., Davis, T. L., & Mitchell, J. E. (2002). Assessment of eating disorders: Comparison of interview and questionnaire data from a long-term follow-up study of bulimia nervosa. Journal of Psy- chosomatic Research, 53, 1043–1047.

Keel, P. K., Dorer, D. J., Eddy, K. T., Franko, D. L., Charatan, D., & Herzog, D. B. (2003). Predictors of mortality in eating disorders. Ar- chives of General Psychiatry, 60, 179 –183.

Keel, P. K., & Mitchell, J. E., (1997). Outcome in bulimia nervosa. American Journal of Psychiatry, 154, 313–321.

Keel, P. K., Mitchell, J. E., Miller, K. B., Davis, T. L., & Crow, S. J. (2000). Predictive validity of bulimia nervosa. American Journal of Psychiatry, 157, 136 –138.

*Kendell, R. E., Hall, D. J., Hailey, A., & Babigian, H. M. (1973). The epidemiology of anorexia nervosa. Psychological Medicine, 3, 200 –203.

Kendler, K. S., MacLean, C., Neale, M., Kessler, R., Heath, A., & Eaves, L. (1991). The genetic epidemiology of bulimia nervosa. American Journal of Psychiatry, 148, 1627–1637.

Kendler, K. S., Walters, E. E., Neale, M. C., Kessler, R. C., Heath, A. C., & Eaves, L. J. (1995). The structure of the genetic and environmental risk factors for six major psychiatric disorders in women: Phobia, generalized anxiety disorder, panic disorder, bulimia, major depression, and alcoholism. Archives of General Psychiatry, 52, 374 –383.

Khandelwal, S. K., Sharan, P., & Saxena, S. (1995). Eating disorders: An Indian perspective. International Journal of Social Psychiatry, 41, 132– 146.

Kiriike, N., Nagata, T., Tanaka, M., & Nishiwaki, S. (1988). Prevalence of binge-eating and bulimia among adolescent women in Japan. Psychiatry Research, 26, 163–169.

Kleinman, A. (1987). Anthropology and psychiatry: The role of culture in cross-cultural research on illness. British Journal of Psychiatry, 151, 447– 454.

Klump, K. L., Miller, K. B., Keel, P. K., McGue, M., & Iacono, W. G. (2001). Genetic and environmental influences on anorexia nervosa syn- dromes in a population-based twin sample. Psychological Medicine, 31, 737–740.

Koronyo-Hamaoui, M., Danziger, Y., Frisch, A., Stein, D., Leor, S., Laufer, N., et al. (2002). Association between anorexia nervosa and the hsKCa3 gene: A family-based and case control study. Molecular Psy- chiatry, 7, 82– 85.

Kortegaard, L. S., Hoerder, K., Joergensen, J., Gillberg, C., & Kyvik, K. O. (2001). A preliminary population-based twin study of self-reported eating disorder. Psychological Medicine, 31, 361–365.

Krahl, W. (1983, March). Two cases of anorexia nervosa in Malaysia. In Asian Forum on Child and Adolescent Psychiatry: Vol. 4. Prevention of psychological problems in children and adolescents (pp. 267–271). Kuala Lumpur, Malaysia: Malaysian Psychiatric Association.

Krahl, W., Quek, S. L., & Raman, N. (1981). A community child and adolescent guidance clinic in Malaysia. Medical Journal of Malay- sia, 36, 171–173.

Kuboki, T., Nomura, S., Ide, M., Suematsu, H., & Araki, S. (1996).

766 KEEL AND KLUMP

Epidemiological data on anorexia nervosa in Japan. Psychiatry Re- search, 62, 11–16.

Lacey, J. H. (1982). Anorexia nervosa and a bearded female saint. British Medical Journal, 285, 1816 –1817.

Lai, K. Y. C. (2000). Anorexia nervosa in Chinese adolescents—Does culture make a difference? Journal of Adolescence, 23, 561–568.

Lai, K. Y. C., Pang, A. H. T., & Wong, C. K. (1995). Case study: Early-onset anorexia nervosa in a Chinese boy. Journal of the American Academy of Child & Adolescent Psychiatry, 34, 383–387.

Lam, M., & Lee, S. (2000). Three different presentations of bulimia nervosa. Hong Kong Medical Journal, 6, 227–230.

Laségue, C. (1873). On hysterical anorexia. Medical Times and Gazette, 2, 265–266.

Lee, A. M., & Lee, S. (1996). Disordered eating and its psychosocial correlates among Chinese adolescent females in Hong Kong. Interna- tional Journal of Eating Disorders, 20, 177–183.

Lee, C. K., Kwak, Y. S., Rhee, H. Kim, Y. S., Han, J. H., Choi, J. O., & Lee, Y. H. (1987). The nationwide epidemiological study of mental disorders in Korea. Journal of Korean Medical Science, 2, 19 –34.

Lee, C. K., Kwak, Y. S., Yamamoto, J., Rhee, H., Kim, Y. S., Han, J. H., et al. (1990). Psychiatric epidemiology in Korea: I. Gender and age differences in Seoul. Journal of Nervous and Mental Disease, 178, 242–246.

Lee, S. (1991). Anorexia nervosa in Hong Kong: A Chinese perspective. Psychological Medicine, 21, 703–711.

Lee, S. (1993). How abnormal is the desire for slimness? A survey of eating attitudes and behaviour among Chinese undergraduates in Hong Kong. Psychological Medicine, 23, 437– 451.

Lee, S. (1995). Self-starvation in context: Towards a culturally sensitive understanding of anorexia nervosa. Social Science Medicine, 41, 25–36.

Lee, S. (2000). Eating disorders are becoming more common in the East too [Letter to the editor]. British Medical Journal, 321, 1023.

Lee, S. (2001). From diversity to unity: The classification of mental disorders in 21st-century China. Psychiatric Clinics of North Amer- ica, 24, 421– 431.

Lee, S., Chiu, H. F. K., & Chen, C. (1989). Anorexia nervosa in Hong Kong: Why not more Chinese? British Journal of Psychiatry, 154, 683– 688.

Lee, S., Ho, T. P., & Hsu, L. K. G. (1993). Fat phobic and non–fat phobic anorexia nervosa: A comparative study of 70 Chinese patients in Hong Kong. Psychological Medicine, 23, 999 –1017.

Lee, S., Hsu, L. K., & Wing, Y. K. (1992). Bulimia nervosa in Hong Kong Chinese patients. British Journal of Psychiatry, 161, 545–551.

Lee, S., & Lee, A. M. (2000). Disordered eating in three communities of China: A comparative study of female high school students in Hong Kong, Shenzhen, and rural Hunan. International Journal of Eating Disorders, 27, 317–327.

Lee, S., Lee, A. M., & Leung, T. (1998). Cross-cultural validity of the Eating Disorder Inventory: A study of Chinese patients with eating disorders in Hong Kong. International Journal of Eating Disorders, 23, 177–188.

Lee, S., Lee, A. M., Leung, T., & Yu, H. (1997). Psychometric properties of the Eating Disorders Inventory (EDI-1) in a nonclinical Chinese population in Hong Kong. International Journal of Eating Disor- ders, 21, 187–194.

Lee, S., Leung, C. M., Wing, Y. K., Chiu, H. F. K., & Chen, C.-N. (1991). Acne as a risk factor for anorexia nervosa in the Chinese. Australian and New Zealand Journal of Psychiatry, 25, 134 –137.

Lee, S., Leung, T., Lee, A. M., Yu, H., & Leung, C. M., (1996). Body dissatisfaction among Chinese undergraduates and its implications for eating disorders in Hong Kong. International Journal of Eating Disor- ders, 20, 77– 84.

Lucas, A. R. (1992). The eating disorder “epidemic”: More apparent than real? Pediatric Annals, 21, 746 –751.

*Lucas, A. R., Beard, C. M., O’Fallon, W. M., & Kurland, L. T. (1988).

Anorexia nervosa in Rochester, Minnesota: A 45-year study. Mayo Clinic Proceedings, 63, 433– 442.

*Lucas, A. R., Beard, C. M., O’Fallon, W. M., & Kurland, L. T. (1991). 50-year trends in the incidence of anorexia nervosa in Rochester, Minn.: A population-based study. American Journal of Psychiatry, 148, 917– 922.

*Lucas, A. R., Crowson, C. S., O’Fallon, W. M., & Melton, L. J. (1999). The ups and downs of anorexia nervosa. International Journal of Eating Disorders, 26, 397– 405.

Marce, L. V. (1860). On a form of hypochondriacal delirium occurring consecutive to dyspepsia, and characterized by refusal of food. Journal of Psychological Medicine and Mental Pathology, 13, 264 –266.

McSherry, J. A. (1985). Was Mary, Queen of Scots, anorexic? Scottish Medical Journal, 30, 243–245.

Miller, M. N., & Pumariega, A. J. (2001). Culture and eating disorders: A historical and cross-cultural review. Psychiatry, 64, 93–110.

Mizushima, N., & Ishii, Y. (1983). The epidemiology of anorexia nervosa in junior and senior high school students in Ishikawa Prefecture. Shin- shin-Igaku, 23, 311–319.

Mogul, S. L. (1980). Asceticism in adolescence and anorexia nervosa. Psychoanalytic Study of the Child, 35, 155–175.

*Møller-Madsen, S., & Nystrup, J. (1992). Incidence of anorexia nervosa in Denmark. Acta Psychiatrica Scandinavica, 86, 197–200.

Morgan, J. F., Marsden, P., & Lacey, J. H. (2000). “Spiritual starvation?”: A case series concerning Christianity and eating disorders. International Journal of Eating Disorders, 28, 476 – 480.

Morton, R. (1694). Pthisiologia: Or, a treatise of consumptions. London: Smith & Walford.

Mumford, D. B., Whitehouse, A. M., Platts, M., & Choudry, I. Y. (1992). Survey of eating disorders in English-medium schools in Lahore, Paki- stan. International Journal of Eating Disorders, 11, 173–184.

*Munk-Jørgensen, P., Møller-Madsen, S., Nielsen, S., & Nystrup, J. (1995). Incidence of eating disorders in psychiatric hospitals and wards in Denmark, 1970 –1993. Acta Psychiatrica Scandinavica, 92, 91–96.

*Nadaoka, T., Oiji, A., Takahashi, S., Morioka, Y., Kashiwakura, M., & Totsuka, S. (1996). An epidemiological study of eating disorders in a northern area of Japan. Acta Psychiatrica Scandinavica, 93, 305–310.

Nakamura, K., Yamamoto, M., Yamazaki, O., Kawashima, Y., Muto, K., Someya, T., et al. (2000). Prevalence of anorexia nervosa and bulimia nervosa in a geographically defined area of Japan. International Journal of Eating Disorders, 28, 173–180.

Nasser, M. (1993). A prescription of vomiting: Historical footnotes. Inter- national Journal of Eating Disorders, 13, 129 –131.

Nasser, M. (1994). Screening for abnormal eating attitudes in a population of Egyptian secondary school girls. Social Psychiatry and Psychiatric Epidemiology, 29, 25–30.

Neki, J. S. (1973). Psychiatry in South-East Asia. British Journal of Psychiatry, 123, 257–269.

Neki, J. S., Mohan, D., & Sood, R. K. (1977). Anorexia nervosa in a monozygotic twin pair. Journal of Indian Medical Association, 68, 98 –100.

*Nielsen, S. (1990). The epidemiology of anorexia nervosa in Denmark from 1973 to 1987: A nationwide register study of psychiatric admis- sion. Acta Psychiatrica Scandinavica, 81, 507–514.

Nobakht, M., & Dezhkam, M. (2000). An epidemiological study of eating disorders in Iran. International Journal of Eating Disorders, 28, 265– 271.

Nogami, Y., & Yabana, F. (1977). On Kibarashi-gui (binge eating). Folia Psychiatrica et Neurologicak Japonica, 31, 159 –166.

Nogami, Y., Yamaguchi, T., Ishiwata, H., Sakai, T., & Kusakabe, Y. (1984). Nichidai igaku zasshi [The prevalence of binge eating among a Japanese university population]. Journal of Nihon University Medical Association, 43, 749 –752.

Norris, D. L. (1979). Clinical diagnostic criteria for primary anorexia nervosa. South African Medical Journal, 56, 987–993.

767CULTURE, GENES, AND EATING DISORDERS

Nwaefuna, A. (1981). Anorexia nervosa in a developing country [Letter to the editor]. British Journal of Psychiatry, 138, 270.

Okasha, A., Kamel, M., Sadek, A., Lotaif, F., & Bishry, Z. (1977). Psychiatric morbidity among university students in Egypt. British Jour- nal of Psychiatry, 131, 149 –154.

Ong, Y. L., Tsoi, W. F., & Cheah, J. S. (1982). A clinical and psychosocial study of seven cases of anorexia nervosa in Singapore. Singapore Medical Journal, 23, 255–261.

Owen, P. R., & Laurel-Seller, E. (2000). Weight and shape ideals: Thin is dangerously in. Journal of Applied Sociology, 30, 979 –990.

*Pagsberg, A. K., & Wang, A.-R. (1994). Epidemiology of anorexia nervosa and bulimia nervosa in Bornholm County, Denmark, 1970 – 1989. Acta Psychiatrica Scandinavica, 90, 259 –265.

Palmer, R. L. (1993). Weight concern should not be a necessary criterion for the eating disorders: A polemic. International Journal of Eating Disorders, 14, 459 – 465.

Parry-Jones, B. (1991). Historical terminology of eating disorders. Psycho- logical Medicine, 21, 21–28.

Parry-Jones, B. (1992). A bulimic ruminator? The case of Dr. Samuel Johnson. Psychological Medicine, 22, 851– 862.

Parry-Jones, B., & Parry-Jones, W. L. (1991). Bulimia: An archival review of its history in psychosomatic medicine. International Journal of Eat- ing Disorders, 10, 129 –143.

Parry-Jones, B., & Parry-Jones, W. L. (1993). Self-mutilation in four historical cases of bulimia. British Journal of Psychiatry, 163, 394 – 402.

Parry-Jones, W. L. L. (1985). Archival exploration of anorexia nervosa. Journal of Psychiatric Research, 19, 95–100.

Pate, J. E., Pumariega, A. J., Hester, C., & Garner, D. M. (1992). Cross- cultural patterns in eating disorders: A review. Journal of the American Academy of Child & Adolescent Psychiatry, 31, 802– 809.

Pawluck, D. E., & Gorey, K. M. (1998). Secular trends in the incidence of anorexia nervosa: Integrative review of population-based studies. Inter- national Journal of Eating Disorders, 23, 347–352.

Peterson, C. B., & Mitchell, J. E. (1999). Psychosocial and pharmacolog- ical treatment of eating disorders: A review of research findings. Journal of Clinical Psychology, 55, 685– 697.

Polivy, J., & Herman, C. P. (1985). Dieting and binging: A causal analysis. American Psychologist, 40, 193–201.

Pope, H. G., & Hudson, J. I. (1988). Is bulimia nervosa a heterogeneous disorder? Lessons from the history of medicine. International Journal of Eating Disorders, 7, 155–166.

Pope, H. G., Hudson, J. I., & Mialet, J.-P. (1985). Bulimia in the late nineteenth century: The observations of Pierre Janet. Psychological Medicine, 15, 739 –743.

Prince, R. (1985). The concept of culture-bound syndromes: Anorexia nervosa and brain-fag, Social Science & Medicine, 21, 197–203.

Pyle, R., Halvorson, P. A., Neuman, P. A., & Mitchell, J. E. (1986). The increasing prevalence of bulimia in freshman college students. Interna- tional Journal of Eating Disorders, 5, 631– 647.

Rampling, D. (1985). Ascetic ideals and anorexia nervosa. Journal of Psychiatric Research, 19, 89 –94.

Ricca, V., Nacmias, B., Cellini, E., Di Bernado, M., Rotella, C. M., & Sorbi, S. (2002). 5-HT2A receptor gene polymorphism and eating dis- orders. Neuroscience Letters, 323, 105–108.

Rieger, E., Touyz, S. W., Swain, T., & Beumont, P. J. V. (2001). Cross- cultural research on anorexia nervosa: Assumptions regarding the role of body weight. International Journal of Eating Disorders, 29, 205–215.

Robins, L. N., Helzer, J. E., Weissman, M. M., Orvaschel, H., Gruenberg, E., Burke, J. D., & Regier, D. A. (1984). Lifetime prevalence of specific psychiatric disorders in three sites. Archives of General Psychiatry, 41, 949 –958.

Roghmann, K. J., Babigian, H. M., Goldberg, I. D., & Zastowny, T. R. (1982). The increasing number of children using psychiatric services: Analysis of a cumulative psychiatric case register. Pediatrics, 70, 790 – 801.

Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research methods and data analysis (2nd ed.). Boston: McGraw-Hill.

Rosenvinge, J. H., & Vandereycken, W. (1994). Early descriptions of eating disorders in the Norwegian medical literature. Acta Paedopsychi- atrica, 56, 279 –281.

Rubinstein, S., & Caballero, B. (2000). Is Miss America an undernourished role model? Journal of the American Medical Association, 283, 1569.

Ruggerio, G. M., Prandin, M., & Mantero, M. (2001). Eating disorders in Italy: A historical review. European Eating Disorders Review, 9, 292– 300.

Russell, G. (1979). Bulimia nervosa: An ominous variant of anorexia nervosa. Psychological Medicine, 9, 429 – 448.

Russell, G. F. M. (1985). The changing nature of anorexia nervosa: An introduction to the conference. Journal of Psychiatric Research, 19, 101–109.

Russell, G. F. M. (1997). The history of bulimia nervosa. In D. M. Garner & P. E. Garfinkel (Eds.), Handbook of treatment for eating disorders (2nd ed., pp. 11–24). New York: Guilford Press.

Schmidt, U. (1993). Bulimia nervosa in the Chinese. International Journal of Eating Disorders, 14, 505–509.

Schotte, D., & Stunkard, A. J. (1987). Bulimia vs. bulimic behaviors on a college campus. Journal of the American Medical Association, 258, 1213–1215.

Shafran, R., Cooper, Z., & Fairburn, C. G. (2002). Clinical perfectionism: A cognitive– behavioral analysis. Behavior Research and Therapy, 40, 773–791.

*Shinkwin, R., & Standen, P. J. (2001). Trends in anorexia nervosa in Ireland: A register study. European Eating Disorders Review, 9, 263– 276.

Silverman, J. A. (1987). Robert Whytt, 1714 –1766, eighteenth century limner of anorexia nervosa and bulimia, an essay. International Journal of Eating Disorders, 6, 143–146.

Song, Y. P., & Fang, Y. K. (1990). A clinical report of 9 cases of anorexia nervosa. Chinese Mental Health Journal, 4, 24 –25.

Sonoda, J., Kiyotake, H., Nozoe, S., & Iwao, T. (1974). A case of anorexia nervosa successfully treated by behavioral therapy in an outpatient clinic. Japanese Journal of Psychiatry, 15, 115–123.

*Soundy, T. J., Lucas, A. R., Suman, V. J., & Melton, L. J. (1995). Bulimia nervosa in Rochester, Minnesota from 1980 to 1990. Psychological Medicine, 25, 1065–1071.

Stein, D. M., & Laakso, W. (1988). Bulimia: A historical perspective. International Journal of Eating Disorders, 7, 201–210.

Strober, M., Lampert, C., Morrell, W., Burroughs, J., & Jacobs, C. (1990). A controlled family study of anorexia nervosa: Evidence of familial aggregation and lack of shared transmission with affective disorders. International Journal of Eating Disorders, 9, 239 –253.

Stunkard, A. (1990). A description of eating disorders in 1932. American Journal of Psychiatry, 147, 263–268.

Suematsu, H., Ishikawa, H., Kuboki, T., & Ito, T. (1985). Statistical studies on anorexia nervosa in Japan: Detailed clinical data on 1,011 patients. Psychotherapy and Psychosomatics, 43, 96 –103.

Suematsu, H., Kuboki, T., & Ogata, E. (1986). Anorexia nervosa in monozygotic twins. Psychotherapy & Psychosomatics, 45, 46 –50.

Swartz, L. (1985). Anorexia nervosa as a culture-bound syndrome. Social Science & Medicine, 20, 725–730.

*Szmukler, G., McCance, C., McCrone, L., & Hunter, D. (1986). Anorexia nervosa: A psychiatric case register study from Aberdeen. Psychological Medicine, 16, 49 –58.

Takeda, A., Suzuki, K., & Matsushita, Y. (1993). Prevalence of bulimia nervosa (DSM–III–R) among male and female high school students. Sishin-Igaku, 35, 1273–1278.

Takei, A., Tamakawa, N., Sato, H., & Okuno, A. (2000). A study of eating disorders in Japanese college students. Seishin Igaku (Clinical Psychi- atry), 42, 877– 882.

768 KEEL AND KLUMP

*Theander, S. (1970). Anorexia nervosa: A psychiatric investigation of 44 female cases. Acta Psychiatrica Scandinavica 214(Suppl.), 1–194.

Tsai, G. (2000). Eating disorders in the Far East. Eating and Weight Disorders, 5, 183–197.

Tseng, M.-C., Lee, M.-B., & Lee, Y.-J. (1989). A clinical study of Chinese patients with eating disorders. Chinese Psychiatry, 3, 17–28.

*Turnbull, S., Ward, A., Treasure, J., Jick, H., & Derby, L. (1996). The demand for eating disorder care: An epidemiological study using the general practice research database. British Journal of Psychiatry, 169, 705–712.

Ung, E. K., Lee, S., & Kua, E. H. (1997). Anorexia nervosa and bu- limia—A Singapore perspective. Singapore Medical Journal, 38, 332– 335.

Van Deth, R., & Vandereycken, W. (1995). Was late-nineteenth-century nervous vomiting an early variant of bulimia nervosa? History of Psy- chiatry, 6, 333–347.

Vandereycken, W., & Beumont, P. J. V. (1990). The first Australian case description of anorexia nervosa. Australian and New Zealand Journal of Psychiatry, 24, 109 –112.

Vandereycken, W., Habermas, T., Van Deth, R., & Meermann, R. (1991). German publications on anorexia nervosa in the nineteenth century. International Journal of Eating Disorders, 10, 473– 490.

Vandereycken, W., & Lowenkopf, E. L. (1990). Anorexia nervosa in 19th century America. Journal of Nervous and Mental Diseases, 178, 531– 535.

Vandereycken, W., & Van Deth, R. (1994). From fasting saints to anorexic girls: The history of self-starvation. New York: New York University Press.

Van’t Hof, S., & Nicolson, M. (1996). The rise and fall of a fact: The increase in anorexia nervosa. Sociology of Health & Illness, 18, 581– 608.

Wade, T. D., Bulik, C. M., Neale, M., & Kendler, K. S. (2000). Anorexia nervosa and major depression: Shared genetic and environmental risk factors. American Journal of Psychiatry, 157, 469 – 471.

Wade, T., Heath, A. C., Abraham, S., & Treloar, S. A. (1996). Assessing the prevalence of eating disorders in an Australian twin population. Australian and New Zealand Journal of Psychiatry, 30, 845– 851.

Walsh, J. M. E., Wheat, M. E., & Freund, K. (2000). Detection, evaluation, and treatment of eating disorders: The role of the primary care physician. Journal of General Internal Medicine, 15, 577–590.

Walters, E. E., Neale, M. C., Eaves, L. J., Heath, A. C., Kessler, R. C., & Kendler, K. S. (1992). Bulimia nervosa and major depression: A study

of common genetic and environmental factors. Psychological Medi- cine, 22, 617– 622.

Wegner, D. M. (2002). The illusion of conscious will. Cambridge MA: MIT Press.

Westberg, L., Bah, J., Rastam, M., Gillberg, C., Wentz, E., Melke, J., et al. (2002). Association between a polymorphism of the 5-HT2C receptor and weight loss in teenage girls. Neuropsychopharmacology, 26, 789 – 793.

Whitehouse, A. M., Cooper, P. J., Vize, C. V., Hill, C., & Vogel, L. (1992). Prevalence of eating disorders in three Cambridge general practices: Hidden and conspicuous morbidity. British Journal of General Prac- tice, 42, 57– 60.

*Willi, J., Giacometti, G., & Limacher, B. (1990). Update on the epide- miology of anorexia nervosa in a defined region of Switzerland. Amer- ican Journal of Psychiatry, 147, 1514 –1517.

*Willi, J., & Grossmann, S. (1983). Epidemiology of anorexia nervosa in a defined region of Switzerland. American Journal of Psychiatry, 140, 564 –567.

*Williams, P., & King, M. (1987, January 24). The “epidemic” of anorexia nervosa: Another medical myth? Lancet, 1, 205–207.

Winslow, L. S. F. (1880). Fasting and feeding: A detailed account of recorded instances of unusual abstinence from food, and of cases illus- trating inordinate appetite. Journal of Psychological Medicine and Men- tal Pathology, 6, 253–299.

Wiseman, C. V., Gray, J. J., Mosimann, J. E., & Ahrens, A. H. (1992). Cultural expectations of thinness: An update. International Journal of Eating Disorders, 11, 85– 89.

Yager, J., & Smith, M. (1993). Restricter anorexia nervosa in a thirteen- year-old sheltered Muslim girl raised in Lahore, Pakistan: Developmen- tal similarities to Westernized patients. International Journal of Eating Disorders, 14, 383–386.

Ziolko, H. U. (1994). Hyperorexia nervosa (1976). International Journal of Eating Disorders, 16, 133–135.

Ziolko, H. U. (1996). Bulimia: A historical outline. International Journal of Eating Disorders, 20, 345–358.

Zuckerman, D., Colby, A., Ware, N. C., & Lazerson, J. S. (1986). The prevalence of bulimia among college students. American Journal of Public Health, 76, 1135–1137.

Received August 30, 2002 Revision received January 24, 2003

Accepted March 27, 2003 �

769CULTURE, GENES, AND EATING DISORDERS