P1: GYQ/GVM/GCY/GIR/GGT/GOQ P2: GYQ

Clinical Child and Family Psychology Review (CCFP) pp426-ccfp-369382 February 26, 2002 18:8 Style file version Nov. 07, 2000

Clinical Child and Family Psychology Review, Vol. 5, No. 1, March 2002 (°
C 2002)

The Children of Mothers With Eating Disorders

Priti Patel,1 Rebecca Wheatcroft,2 Rebecca J. Park,3,4 and Alan Stein5,6

There is good evidence that children of parents with psychological disorders are themselves
at increased risk of disturbances in their development. Although there has been considerable
research on a variety of disorders such as depression and alcohol, research on the children of
parents with eating disorders has been relatively recent. This paper aims to review the evidence
and covers a number of areas, including genetic factors, pregnancy, the perinatal and postpar-
tum period, infancy, and the early years of life, focusing on feeding and mealtimes, general
parenting functions, and growth. This is followed by a consideration of psychopathology in
the children, parental attitudes to children’s weight and shape, and adolescence. What is clear
is that although there are numerous case reports and case series, the number of systematic
controlled studies is relatively small, and almost nothing has been written about the children
of fathers with eating disorders. What is evident from the available evidence is that children of
mothers with eating disorders are at increased risk of disturbance, but that the risk depends on
a variety of factors, and that difficulties in the children are far from invariable. The paper con-
cludes by summarizing five broad categories of putative mechanisms, based on the evidence
to date, by which eating disturbance in parents can influence child development.
KEY WORDS: eating disorder; intergenerational transmission; feeding; body shape; growth.

Since the 1960s, studies have established that chil- Rutter, 1989). Eating disorders have been the subject
dren of parents with psychological disturbance are at of a great deal of research. However, eating disor-
risk of disturbance themselves (Garmezy & Masten, ders in parents and the implications for their children
1994; Richman, Graham, & Stevenson, 1982; Rutter, have only been a relatively recent research focus. It is
1989; Rutter & Quinton, 1984). Furthermore the dis- the aim of this review to examine the literature to
turbance in children may persist well after remission identify what is known on the subject; in the con-
of parental disorder and this has been established cluding section we suggest possible mechanisms by
particularly in conditions such as postnatal depres- which eating disorders influence parenting and child
sion (Barnes & Stein, 2000; Murray & Cooper, 1996; development.
It is notable that almost all the research to date
in this area has concerned mothers with eating dis-
orders. Very few studies have examined fathers with
1 Tavistock Centre, London.
2 Leopold
eating disorders or included the roles and attitudes of
Muller Department of Child and Family Mental Health,
fathers in exacerbating, ameliorating, or acting inde-
Royal Free and University College Medical School, Royal Free
Campus, London. pendently of the possible influences of a mother who
3 Developmental Psychiatry Section, Department of Psychiatry, has an eating disorder. This is an area that requires
University of Cambridge, Cambridge. research attention. There is much research on family
4 Bedford and Luton Community NHS Trust, Cambridge.
5 Section of Child and Adolescent Psychiatry, Department of Psy-
functioning in adolescents with eating disorders but
it has not generally dealt with the issue of parents
chiatry, Warneford Hospital, University of Oxford, Oxford.
6 Address all correspondence to Alan Stein, Department of Psychi- with eating disorders and their influence on the chil-
atry, Warneford Hospital, University of Oxford, Oxford OX3 7JX, dren, and was judged to be not within the scope of this
United Kingdom; e-mail: [email protected]. review.

1
1096-4037/02/0300-0001/0 °
C 2002 Plenum Publishing Corporation
P1: GYQ/GVM/GCY/GIR/GGT/GOQ P2: GYQ
Clinical Child and Family Psychology Review (CCFP) pp426-ccfp-369382 February 26, 2002 18:8 Style file version Nov. 07, 2000

2 Patel, Wheatcroft, Park, and Stein

EATING DISORDERS: PREVALENCE binge eating without the inappropriate compensatory

AND DIAGNOSTIC CRITERIA behaviors of BN. The disorder commonly occurs with
obesity (Fairburn, Cowen, & Harrison, 1999).
Eating disorders are a common source of psy- The potential influence of maternal eating disor-
chiatric morbidity in women of childbearing age. ders on child development will be considered broadly
Epidemiological studies estimate the prevalence of chronologically as follows: beginning before concep-
anorexia nervosa and bulimia nervosa to be 1–2% in tion with genetic influences, then considering preg-
women aged 16–40 years (Fairburn & Beglin, 1990). nancy, the postpartum period, and infancy and early
This rate increases to 4% when the full spectrum of childhood where we focus specifically on topics such
the disorder is considered (Hoek, 1993). The onset of as feeding and growth and general parenting. We then
eating disorders is usually in adolescence and there move onto later childhood, to parental concerns about
is some evidence that it is on the increase (Treasure, their children’s weight and shape, the possible de-
Troop, & Ward, 1996), affecting women from wider velopment of psychopathology in the children, and
socioeconomic and ethnic backgrounds than previ- adolescence.
ously considered.
The main criteria for a DSM-IV diagnosis of
GENETIC FACTORS
anorexia nervosa (AN) are the following: a body
weight 15% below expected standard (Body Mass
It has long been recognized that eating disorders
Index (BMI) < 17.5); an intense fear of weight gain;
aggregate in families (Strober, Morrell, Burroughs,
distorted body image and amenorrhoea in women.
Salkin, & Jacobs, 1985). Family studies report a
Onset is usually in adolescence and generally be-
higher risk for eating disorders in first-degree relatives
gins with efforts at dieting that then become ex-
than in relatives of controls (Strober, 1991; Strober,
treme. Bulimia nervosa (BN) was first described by
Freeman, Lampert, Diamond, & Kaye, 2000; Strober
Russell (1979) as an “ominous variant” of anorexia
et al., 1985). Two of the largest and most systematic
nervosa. The central feature of this disorder is recur-
of these studies (Lilenfeld et al., 1998; Strober et al.,
rent episodes of uncontrolled binge eating alongside
2000) found a 7–12-fold increase in the prevalence of
extreme compensatory measures to control weight
AN and BN in relatives of individuals with eating dis-
gain, such as vomiting, purging, and exercise. These
order compared with controls, providing strong sup-
individuals also have a morbid fear of fatness and eval-
port for familial transmission of both types of eating
uate themselves by body shape and weight. Women
disorder. Cross-transmission of AN and BN in fam-
with BN are usually of normal weight and have regu-
ilies suggests a common or shared familial diathesis,
lar menstrual cycles. Comorbid depressive symptoms
which extends to milder eating disorder phenotypes
are common and the disorder tends to run a chronic
(Strober et al., 2000).
course (Herzog, Keller, Sacks, Yeh, & Lavori, 1992).
As well as considerable psychiatric morbidity,
mortality rates in eating disorders are raised com- Twin Studies
pared with the general population. Reported mortal-
ity rates from long-term follow up studies of AN and As first-degree relatives share both genes and
BN are in the region of 0–5% (Sullivan, 1995; Keel, environment, family studies are unable to differen-
Heatherton, Harnden, & Honig, 1997). In addition, tiate genetic from environmental effects. To establish
DSM-IV recognises a third category termed “eating this differentiation, twin studies have been conducted,
disorder not otherwise specified,” which includes dis- which compare similarity for a trait or disorder be-
orders of eating that do not meet criteria for AN or tween dizygotic (DZ-nonidentical) and monozygotic
BN. This category is important because this group of (MZ-identical) twins; MZ twins share around twice
conditions is more common than AN or BN and the as much genetic material as do DZ twins. Early re-
individuals who fall into this category still experience ports (Holland, Hall, Murray, Russell, & Crisp, 1984;
major disturbance and domination of their lives by Holland, Sicotte, & Treasure, 1988) found that rates
extreme concerns about body shape and weight as of AN were much higher in MZ than in DZ twins (56
well as a significant disruption of their eating patterns. vs. 5%) and suggested that up to 80% of the variance
“Binge eating disorder” appears in the appendix of in liability to AN may be accounted for by genetic
DSM-IV as a category of eating disorder requiring factors. In contrast, analyses on a subset of twins with
further study. The main features of this disorder are BN found negligible genetic involvement (Treasure