Childhood Adversities Associated With Risk For Eating Disorders or Weight Problems
Childhood Adversities Associated With Risk For Eating Disorders or Weight Problems
Childhood Adversities Associated With Risk For Eating Disorders or Weight Problems
revious research has suggested that childhood adversities may contribute to the development of eating disorders. Individuals with eating disorders are more likely than
those without eating disorders to report a history of childhood maltreatment (13), other chronic or episodic childhood adversities (47), and problematic relationships with
their parents (1, 813). These findings have enabled researchers to generate hypotheses about the role of childhood adversities in the development of eating disorders
(14). However, nearly all of the studies that have examined
the association between childhood adversities and eating
disorders have been cross-sectional case-control investigations. It is problematic to make inferences about risks
associated with the onset of eating disorders among individuals in the general population on the basis of cross-sectional studies of patients with eating disorders.
To investigate potential risk factors that may contribute
to the development of eating disorders, prospective longitudinal research with a sizable community-based sample
is necessary. Potential risk factors must be assessed before
394
Method
search has supported the reliability and validity of these 10 indices of childhood temperament (21).
Parental psychiatric symptoms. Interview items used to assess maternal psychiatric symptoms in 1975, 1983, and 1985
1986 were obtained from the Disorganizing Poverty Interview, the
California Psychological Inventory (22), the Hopkins Symptom
Checklist (23), and instruments that assessed maternal alienation
(24), rebelliousness (25), and other maladaptive traits (26, 27). Paternal alcohol abuse, drug abuse, and antisocial behavior were assessed during the 1975, 1983, and 19851986 interviews by using
the Disorganizing Poverty Interview. In addition, lifetime histories of parental anxiety, depressive, disruptive, and substance use
disorders were assessed during the 19911993 maternal interview
by using items adapted from the New York High Risk Study Family
Interview (28). Additional data were provided by the interviewers
observations of the mothers behavior during the interview. Parental eating disorders were not assessed. Data regarding age at
onset permitted identification of maternal and paternal psychiatric disorders that were evident by 19851986. Computerized diagnostic algorithms were developed by using the items from these
instruments to assess the DSM-IV diagnostic criteria for maternal
personality disorders, paternal antisocial personality disorder,
and maternal and paternal anxiety, depressive, disruptive, and
substance use disorders.
The current analyses were conducted with data from 782 families for whom information was available on childhood adversities and problems with eating or weight during adolescence or
early adulthood. These families are a subset of 976 randomly
sampled families from two upstate New York counties. The
mothers were interviewed in 1975, 1983, 19851986, and 1991
1993 (15, 16). One randomly selected child from each family was
interviewed in 1983, 19851986, and 19911993. The sample was
generally representative of families in the northeastern United
States with regard to socioeconomic status and most demographic variables; the sample reflected the region in having high
proportions of Catholic (54%) and white (91%) participants (16).
The mean age of the 397 male and 385 female offspring in the
sample was 6 years (SD=3) in 1975, 14 years (SD=3) in 1983, 16
years (SD=3) in 19851986, and 22 years (SD=3) in 19911993.
Study procedures were approved according to appropriate institutional guidelines. Written informed consent was obtained after the interview procedures were fully explained. The youths
and their mothers were interviewed separately, and both interviewers were blind to the responses of the other informant. Additional methodological information is available from previous
reports (15, 16).
Assessment
Offspring psychiatric disorders and childhood temperament. The parent and youth versions of the Diagnostic Interview Schedule for Children (17) were administered in 1983,
19851986, and 19911993 to assess offspring psychiatric disorders. Both the parents and the youths were interviewed because
the use of multiple informants tends to increase the reliability
and validity of psychiatric diagnoses (18, 19). A symptom was
considered present if it was reported by either informant. Previous research has indicated that the reliability and validity of
the Diagnostic Interview Schedule for Children as employed in
the present study are comparable to those of other structured
interviews (20).
Diagnostic Interview Schedule for Children items assessed
the diagnostic criteria for eating disorders, as well as a wide
range of specific eating and weight problems. Computer algorithms were subsequently developed to determine whether individuals met the DSM-IV criteria for eating disorders. A diagnosis of eating disorder not otherwise specified (e.g., anorexia
without amenorrhea, binge eating disorder) was made if clinically significant eating disorder symptoms were present but the
full criteria for anorexia nervosa, binge eating disorder, or bulimia nervosa were not met. Participants were defined as obese
if their weight was 150% of normal body weight and 2 standard deviations above the sample mean. Participants were defined as having low body weight if their weight was 90% of normal body weight and 2 standard deviations below the sample
mean. Neither obesity nor low body weight was considered an
eating disorder unless the individual had symptoms that met
the DSM-IV criteria for an eating disorder. Additional data were
drawn from the interviewers observations of the childs behavior during the interview.
Ten dimensions of difficult childhood temperament were assessed by using the Disorganizing Poverty Interview (15) during
the 1975 maternal interviews: 1) clumsiness-distractibility,
2) nonpersistence-noncompliance, 3) anger, 4) aggression to
peers, 5) problem behavior, 6) temper tantrums, 7) hyperactivity,
8) crying-demanding, 9) fearful withdrawal, and 10) moodiness.
Children with severe problems in one or more of these domains
were identified as having a difficult temperament. Previous reAm J Psychiatry 159:3, March 2002
395
Prevalence in Male
Offspring (N=397)
Analysis
N
102
27
79
38
72
20
59
%
26
7
21
10
19
5
15
N
34
8
24
13
70
5
7
%
9
2
6
3
18
1
2
Odds Ratiob
3.85
3.67
4.01
3.23
1.07
4.30
10.08
95% CI
2.535.85
1.648.18
2.486.49
1.706.17
0.751.55
1.6011.56
4.5422.38
0
8
35
9
8
43
0
2
9
2
2
11
1
2
6
0
1
9
<1
1
2
0
<1
2
4.19
6.52
8.40
5.42
0.8819.86
2.7115.68
1.0567.51
2.6011.28
A subset of 976 randomly sampled families in two upstate New York counties. Mothers were interviewed in 1975, 1983, 19851986, and
19911993. One randomly selected child from each family was interviewed in 1983, 19851986, and 19911993. Interviewers used a variety
of instruments assessing childhood maltreatment, eating problems, environmental risk factors, temperament, maladaptive parental behavior, and parental psychopathology.
b Odds ratios represent comparisons between male and female offspring.
c Of the 41 subjects with eating disorder not otherwise specified, nine had anorexia nervosa without amenorrhea, nine had binge eating disorder, and 23 had other forms of eating disorder not otherwise specified (e.g., self-induced vomiting after eating small amounts of food).
the sample mean and if there was clear evidence of parental neglect (e.g., failure to vaccinate the child).
Results
Descriptive Statistics
The Disorganizing Poverty Interview was used to assess the following childhood adversities in 1975, 1983, and 19851986: death
of a parent, disabling parental accident or illness, living in an unsafe neighborhood, low level of parental education, parental separation or divorce, peer aggression, low family income, school violence, the presence of a crime victim in the household, and
upbringing by a single parent. Family income was transformed to
percentage of the current U.S. poverty levels in 1975, 1983, and
19851986. Poverty was defined as mean income below 100% of
the U.S. poverty levels. Low level of parental education was defined as less than a high school education for one or both parents.
Adversities were considered present if reported at any of the three
assessments. Numerous studies have supported the reliability
and validity of the Disorganizing Poverty Interview (15, 16).
396
Prevalence of Problem
Prevalence of Problem
Analysis
Total N
Odds Ratiob
95% CI
711
644
48
44
7
7
24
22
6
6
25
27
4.82
5.11
1.8212.73
1.9113.72
586
625
63
68
11
11
196
157
40
35
20
22
2.13
2.35
1.383.29
1.493.69
562
711
707
19
38
36
3
5
5
220
24
75
29
5
12
13
21
16
4.34
4.66
3.55
2.387.92
1.6513.16
1.767.17
711
644
117
120
16
19
24
22
10
9
42
41
3.63
3.02
1.588.36
1.267.24
711
644
120
111
17
17
24
22
9
11
38
50
2.96
4.80
1.266.91
2.0311.35
586
625
620
39
43
43
7
7
7
196
157
162
27
23
23
14
15
14
2.24
2.32
2.22
1.333.77
1.353.99
1.303.81
711
644
21
21
3
3
24
22
4
4
17
18
6.57
6.59
2.0620.92
2.0521.19
Total N
A subset of 976 randomly sampled families in two upstate New York counties. Mothers were interviewed in 1975, 1983, 19851986, and
19911993. One randomly selected child from each family was interviewed in 1983, 19851986, and 19911993. Interviewers used a variety
of instruments assessing childhood maltreatment, eating problems, environmental risk factors, temperament, maladaptive parental behavior, and parental psychopathology.
b Each odds ratio was significant when an alpha level of 0.01 was used, and each remained significant after the effects of offspring age, sex,
difficult childhood temperament, childhood eating problems, parental psychiatric disorders, and co-occurring childhood adversities were
controlled statistically.
FIGURE 1. Association of Maladaptive Maternal and Paternal Behavior With Prevalence of Eating Disorders During
Adolescence or Early Adulthood in 782 Offspring in a Community Sample of Familiesa
13
Eating Disorder Prevalence During
Adolescence or Early Adulthood (%)
ratio=4.58, 95% CI=1.9910.46) were associated with elevated risk for problems with eating or weight during early
adulthood. The index of maladaptive paternal behavior
was associated with eating or weight problems during
early adulthood (r=0.10, df=781, p=0.004). The index of
maladaptive maternal behavior was not associated with
eating or weight problems during early adulthood.
12
11
Paternal behavior
Maternal behavior
10
9
8
7
6
5
4
3
2
1
0
0
1 or 2
397
Prevalence of Problem
Prevalence of Problem
Analysis
Total N
Total N
Odds Ratiob
95% CI
313
28
72
15
21
2.68
1.355.33
368
71
19
17
47
3.72
1.399.98
347
319
286
356
354
344
13
9
7
14
11
12
4
3
2
4
3
3
38
66
99
14
31
41
6
10
12
3
8
7
16
15
12
21
26
17
4.82
6.15
5.50
6.66
10.85
5.70
1.7113.54
2.3915.82
2.0914.40
1.6726.59
3.9729.59
2.1015.43
368
92
25
17
10
59
4.29
1.5911.58
367
288
52
35
14
12
18
97
7
24
39
25
3.85
2.38
1.4310.40
1.324.25
325
16
17
24
5.94
1.7420.29
319
13
24
17
4.71
1.4115.76
A subset of 976 randomly sampled families in two upstate New York counties. Mothers were interviewed in 1975, 1983, 19851986, and
19911993. One randomly selected child from each family was interviewed in 1983, 19851986, and 19911993. Interviewers used a variety
of instruments assessing childhood maltreatment, eating problems, environmental risk factors, temperament, maladaptive parental behavior, and parental psychopathology.
b Each odds ratio was significant when an alpha level of 0.01 was used, and each remained significant after controlling for offspring age, difficult childhood temperament, childhood eating problems, parental psychiatric disorders, and co-occurring childhood adversities.
The index of maladaptive paternal behaviors was significantly correlated with the total number of eating or weight
problems during adolescence or early adulthood among
the male (r=0.12, df=396, p=0.01) and female (r=0.10, df=
384, p=0.04) offspring. The index of maladaptive maternal
behaviors was not significantly correlated with the total
number of eating or weight problems during adolescence
or early adulthood in either subsample. This pattern of
findings was obtained regardless of whether the mother or
the youth provided the data regarding parental behavior.
Maladaptive paternal behaviors remained significantly associated with the total number of offspring problems with
eating or weight after the effects of co-occurring childhood
adversities were controlled statistically (t=2.89, df=777, p=
0.004).
The indices of maladaptive maternal behaviors, types of
childhood maltreatment, and other childhood adversities
and socioeconomic variables were not independently associated with offspring risk for eating disorders after the effects of co-occurring childhood adversities were controlled
statistically. However, the index of types of childhood maltreatment was independently associated with strict dieting
(odds ratio=1.60, 95% CI=1.162.20), recurrent fluctuations
in weight (odds ratio=1.58, 95% CI=1.152.16), and vomiting (odds ratio=1.89, 95% CI=1.083.29), and the index of
other adversities was independently associated with obesity during adolescence or early adulthood (odds ratio=
1.23, 95% CI=1.061.43).
398
Discussion
The present findings advance our understanding of the
association between childhood adversities and risk for
eating disorders in several respects. First, our findings indicate that a wide range of childhood adversities tend to
be associated with elevated risk for problems with eating
or weight during adolescence or early adulthood after the
effects of childhood eating problems, difficult childhood
temperament, parental psychopathology, and co-occurring childhood adversities are controlled statistically. Further, the findings suggest that there may be unique associations between specific childhood adversities and specific
Am J Psychiatry 159:3, March 2002
problems with eating or weight and that there may be different patterns of association between adversities and
problems with eating or weight among males and females
in the general population.
The present findings suggest that maladaptive paternal
behavior may play a more important role than maladaptive maternal behavior in the development of eating disorders in offspring. Most of the theoretical literature in this
area has focused on the mother-child relationship (3741).
However, our findings are consistent with previous research suggesting that low paternal affection (8), care (9),
and empathy (12) and high paternal control (8), unfriendliness (14), overprotectiveness (11), and seductiveness (10)
are associated with the development of eating disorders in
offspring. Although the fathers were not interviewed in the
present study, the findings are not likely to be attributable
to reporting bias on the part of the informants. First, the
overall rate of maladaptive paternal behavior was not
higher than that of maladaptive maternal behavior. Second, the same pattern of findings was obtained with data
for paternal behavior that were obtained during the maternal and offspring interviews. Third, maladaptive paternal behavior was not more strongly associated with offspring risk for other psychiatric disorders than was
maladaptive maternal behavior (34). Our findings also
suggest that low paternal identification may partially mediate the association between maladaptive paternal behavior and eating disorders in offspring. It will be of interest for future research to fu rther investiga te the
mechanisms that underlie this association.
The present findings are consistent with previous crosssectional research suggesting that childhood maltreatment (13) and maladaptive parental behavior (813, 42)
may contribute to the development of eating disorders
and that many types of childhood adversities may be associated with risk for problems with eating or weight (47,
43). Because the present findings are based on prospective
longitudinal data, they provide more compelling support
for these hypotheses. Moreover, our findings support the
hypothesis that the causes of eating disorders tend to be
heterogeneous and multifactorial (44). However, because
previous research has provided inconsistent findings
about the nature of the association between childhood adversities and the development of weight problems, it will
be important for future research to investigate this association more extensively.
The limitations of the present study merit consideration. There were not enough cases to permit analyses of
the relationships between childhood adversities and specific eating disorders. Therefore, we have reported associations between childhood adversities and several different
types of eating and weight problems. To have enough statistical power to conduct separate analyses with the female and male subsamples, data on eating and weight
problems during adolescence and early adulthood were
pooled, and there was some overlap in the periods during
Am J Psychiatry 159:3, March 2002
which some of the risk factors and outcomes were assessed. To address this concern, we have reported associations between childhood adversities and problems with
eating or weight during early adulthood. Despite these
limitations, the present findings provide a detailed, systematic, and methodologically rigorous contribution to
the literature.
Received Feb. 6, 2001; revisions received May 18 and July 26, 2001;
accepted Aug. 15, 2001. From Columbia University and the New York
State Psychiatric Institute; and Mount Sinai School of Medicine, New
York. Address reprint requests to Dr. Johnson, Box 60, New York State
Psychiatric Institute, 1051 Riverside Dr., New York, NY, 10032;
jjohnso@pi.cpmc.columbia.edu (e-mail).
References
1. Fairburn CG, Welch SL, Doll HA, Davies BA, OConnor ME: Risk
factors for bulimia nervosa: a community-based case-control
study. Arch Gen Psychiatry 1997; 54:509517
2. De Groot J, Rodin GM: The relationship between eating disorders and childhood trauma. Psychiatr Annals 1999; 29:225
229
3. Connors ME, Morse W: Sexual abuse and eating disorders: a review. Int J Eat Disord 1993; 13:111
4. Horesh N, Apter A, Lepkifker E: Life events and severe anorexia
nervosa in adolescence. Acta Psychiatr Scand 1995; 91:59
5. Schmidt U, Humfress H, Treasure J: The role of general family
environment and sexual and physical abuse in the origins of
eating disorders. Eur Eating Disorders Rev 1997; 5:184207
6. Welch SL, Doll HA, Fairburn CG: Life events and the onset of bulimia nervosa: a controlled study. Psychol Med 1997; 27:515
522
7. Troop NA, Treasure JL: Setting the scene for eating disorders, II:
childhood helplessness and mastery. Psychol Med 1997; 27:
531538
8. Wonderlich S, Ukestad L, Perzacki R: Perceptions of nonshared
childhood environment in bulimia nervosa. J Am Acad Child
Adolesc Psychiatry 1994; 33:740747
9. Kendler KS, Walters EE, Neale MC, Kessler RC, Heath AC, Eaves
LJ: 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. Arch Gen Psychiatry 1995; 52:374383
10. Rorty M, Yager J, Rossotto E, Buckwalter G: Parental intrusiveness recalled by women with a history of bulimia nervosa and
comparison women. Int J Eat Disord 2000; 28:202208
11. Calam R, Waller G, Slade PD, Newton T: Eating disorders and
perceived relationships with parents. Int J Eat Disord 1990; 9:
479485
12. Steiger H, Van der Feen J, Goldstein C, Leichner P: Defense
styles and parental bonding in eating-disordered women. Int J
Eat Disord 1989; 8:131140
13. Horesh N, Apter A, Ishai J, Danziger Y, Miculincer M, Stein D,
Lepkifker E, Minouni M: Abnormal psychosocial situations and
eating disorders in adolescence. J Am Acad Child Adolesc Psychiatry 1996; 35:921927
14. Wonderlich S, Klein MH, Council JR: Relationship of social perceptions and self-concept in bulimia nervosa. J Consult Clin
Psychol 1996; 64:12311237
15. Kogan LS, Smith J, Jenkins S: Ecological validity of indicator
data as predictors of survey findings. J Soc Service Res 1977; 1:
117132
16. Cohen P, Cohen J: Life Values and Adolescent Mental Health.
Mahwah, NJ, Lawrence Erlbaum Associates, 1996
399
400
31. Brook JS, Whiteman M, Gordon AS, Brook DW: Paternal determinants of female adolescents marijuana use. Dev Psychol
1984; 20:10321043
32. Wagner BM, Cohen P: Adolescent sibling differences in suicidal
symptoms: the role of parent-child relationships. J Abnorm
Child Psychol 1994; 22:321337
33. Johnson JG, Cohen P, Smailes E, Skodol A, Brown J, Oldham J:
Childhood verbal abuse and risk for personality disorders during adolescence and early adulthood. Compr Psychiatry 2001:
42:1623
34. Johnson JG, Cohen P, Kasen S, Smailes E, Brook J: The association of maladaptive parental behavior with psychiatric disorder among parents and their offspring. Arch Gen Psychiatry
2001; 58:453460
35. Johnson JG, Smailes EM, Cohen P, Brown J, Bernstein DP: Associations between four types of childhood neglect and personality disorder symptoms during adolescence and early adulthood: findings of a community-based longitudinal study. J
Personal Disord 2000; 14:171187
36. Straus MA, Kinard EM, Williams LM: The Neglect Scale.
Durham, University of New Hampshire, Family Research Laboratory, 1995
37. Likierman M: On rejection: adolescent girls and anorexia. J
Child Psychotherapy 1997; 23:6180
38. Humphrey LL, Stern S: Object relations and the family system
in bulimia: a theoretical integration. J Marital Fam Ther 1988;
14:337350
39. Marsden P: Demeter and Persephone: fear of cannibalistic engulfment in bulimia. Br J Psychother 1997; 13:318326
40. Hirsch M: The body as a transitional object. Psychother Psychosom 1994; 62:7881
41. Strober M, Humphrey LL: Familial contributions to the etiology
and course of anorexia nervosa and bulimia. J Consult Clin Psychol 1987; 55:654659
42. Chatoor I, Egan J, Getson P, Menvielle E: Mother-infant interactions in infantile anorexia nervosa. J Am Acad Child Adolesc
Psychiatry 1988; 27:535540
43. Wamala SP, Wolk A, Orth-Gomer K: Determinants of obesity in
relation to socioeconomic status among middle-aged Swedish
women. Prev Med 1997; 26:734744
44. Rosen DS, Neumark Sztainer D: Reviews of options for primary
prevention of eating disturbances among adolescents. J Adolesc Health 1998; 23:354363