Departments of Public Health and Psychiatry,
University of Helsinki, Finland
Department of Epidemiology,
Columbia University, USA
Genetic and environmental influences on
body image, disordered eating,
and intentional weight loss
To be presented, with the permission of the Faculty of Medicine at the
University of Helsinki, for public examination
at the Lapinlahti Hospital
on November 26, 2004, at noon
Kansanterveystieteen laitoksen julkaisuja M184:2004
Prof. Aila Rissanen
Obesity Research Unit, Department of Psychiatry, University of Helsinki
Prof. Jaakko Kaprio,
Department of Public Health, University of Helsinki
and National Institute of Public Health, Finland
Prof. Matti Virkkunen,
Department of Psychiatry, University of Helsinki
Dr. Ritva Prättälä
National Institute of Public Health, Helsinki
Dr. Jennifer R. Harris
The National Institute of Public Health,
Department of Population Health Sciences, Oslo, Norway.
The National Institute on Aging, The National Institutes of Health, USA
Prof. Mauri Marttunen
University of Kuopio
and National Institute of Public Health, Finland
Cover photo © Natalia Baer 2004
ISBN (paperback) 952-10-1366-4
ISBN (pdf) 952-10-2143-8 http://ethesis.helsinki.fi/
Picaset Oy, Helsinki 2004
List of original publications ..............................................................................................................9
2. Literature Review...........................................................................................................................13
Definitions and prevalences............................................................................................................13
Breakfast skipping ...........................................................................................................................14
Body shape and weight concerns.........................................................................................................16
Intentional weight loss.......................................................................................................................17
Weight control behaviours: correlates and demographic determinants ...................................18
Disordered eating and peripubertal problems......................................................................................19
Behavioral and demographic correlates of breakfast skipping ..............................................................19
Weight control behaviours .................................................................................................................20
Familial aggregation and heritability ..............................................................................................23
Genetics of weight change...................................................................................................................23
Genetics of disordered eating..............................................................................................................24
Genetics of food intake......................................................................................................................26
3. Aims of the study ...........................................................................................................................27
4. Subjects and methods...................................................................................................................29
Weight control behaviours: correlates and demographics .........................................................37
Breakfast skipping ...........................................................................................................................37
Body shape and weight concerns.........................................................................................................39
Intentional weight loss.......................................................................................................................40
Genetic and environmental influences on weight control behaviors .......................................43
Breakfast skipping ...........................................................................................................................43
Body shape and weight concerns.........................................................................................................44
Intentional weight loss.......................................................................................................................46
Correlates of weight control behaviors.........................................................................................47
Genetic and environmental influences on weight control behaviours......................................................48
Strengths and limitations ..................................................................................................................50
7. Acknowledgments .........................................................................................................................53
Many previous studies have shown that body size is strongly genetically determined. The rapid
global increase in rates of obesity implies that environmental influences must also be
important. In concert, disordered eating and eating disorders have become increasingly
widespread in Western countries during the second half of the 20th century. However,
relatively little is known about genetic and environmental contributions to individual variation
in weight control. The aim of this study was to explore the genetic and environmental
contributions to breakfast skipping, body weight and shape concerns, and intentional weight
loss attempts in a large population sample of young Finnish twins.
The participants, five birth cohorts of Finnish twins born 1975-79, were followed
longitudinally from the age 16 to age 22-27. At baseline, the health habits of the twins were
assessed using self-report questionnaires that included questions on weight, height,
sociodemographics, various health habits, and breakfast skipping frequency. Simultaneously,
similar information was obtained by a behavioral and medical self-report questionnaire from
the twins’ mothers and fathers. The twins received three follow-up questionnaires at ages 17,
18, and 22-27. The final fourth questionnaire assessed weight, height, intentional weight loss,
eating styles, and many other nutrition-related variables. Body shape and weight concerns were
measured using the Body Dissatisfaction and Drive for Thinness subscales from the Eating
Disorder Inventory. Demographic and behavioral correlates of breakfast skipping, body shape
and weight concerns, and intentional weight loss were analyzed using cross-tabulations and
univariate and multivariable logistic regression models. The family-based sampling was taken
into account in all analyses. Different structural equation modeling strategies were used to
estimate genetic and environmental influences on body image and weight control patterns in
twins; twin-family modeling was used to explore breakfast eating patterns in twins and their
Parental breakfast skipping was the strongest factor associated with adolescent breakfast
skipping. In both adults and adolescents, health-compromising factors, such as smoking,
infrequent exercise, frequent alcohol use, and a high BMI, were significantly associated with
breakfast skipping. Breakfast skipping was associated with low family socioeconomic status in
adults and adolescent boys, but not in girls. Additive genetic effects explained 41% (95%
confidence interval [CI]: 21-63%) of the variance in breakfast skipping patterns in girls and
66% (95% CI: 47-79%) in boys, and common environmental effects 45% (95% CI: 23-62%)
in girls and 14% (95% CI: 5-29%) in boys. Twin-family models confirmed that there are
substantial additive genetic and shared environmental influences in addition to individual-
specific environmental effects and gender differences.
With regards to body shape and weight concerns, various psychosomatic symptoms were
significantly associated with high Body Dissatisfaction (BD) and Drive for Thinness (DT) in
both genders. In women, early onset of puberty, early initiation of sexual activity, multiple sex
partners, and lower educational attainment at age 16-17 were statistically significant risk
factors of BD experienced in young adulthood. In gender-specific univariate twin models,
additive genes accounted for 59% (95% CI: 53-65%) of variance in BD and 51.0% (95% CI:
44-58%) of DT in females, but for none in males.
Individuals engaging in intentional weight loss (IWL) attempted to restrict food intake and
avoid fatty and calorie-rich foods, but simultaneously exhibited disordered and unhealthy
eating patterns. Snacking and eating in the evening were characteristic of individuals with at
least two IWL attempts. Eating in response to visual and emotional cues was very pronounced
in women who had engaged in IWL, but much less so in men. IWL was estimated to have a
heritability of 38% (95% CI: 19-55%) in men and 66% (95% CI: 55-75%) in women. The
overlap in genetic effects on BMI and IWL was 0.38 (95% CI: 0.28-0.47) for men and 0.45
(95% CI: 0.41-0.52) for women, implying that genetic effects that affect BMI are only partially
shared with those affecting IWL.
Weight control related issues have general health implications: breakfast skipping was
associated with health-compromising behaviors. Body shape and weight concerns were
associated with larger body size and multiple psychosomatic symptoms. Intentional weight
loss attempts were also associated with larger body size and disordered eating styles,
particularly restricting and overeating.
Overall, both genes and environment were important for the eating-related phenotypes
studied in Finnish young adult twins. Breakfast skipping in girls and boys, intentional weight
loss in women and men, and body shape and weight concerns in women exhibited moderate
to strong genetic influences. Body shape and weight concerns exhibited substantial and
breakfast skipping modest influences of shared family environment. The genetic influences on
intentional weight loss were only partially shared with those affecting BMI.
List of original publications
I Keski-Rahkonen A, Kaprio J, Rissanen A, Virkkunen M, Rose RJ. Breakfast skipping
and health-compromising behaviors in adolescents and adults. Eur J Clin Nutr 2003;
II Keski-Rahkonen A, Viken RJ, Rissanen A, Kaprio J, Rose RJ. Genetic and
environmental factors in breakfast eating patterns. Behavior Genetics 2004; 34:503-
III Keski-Rahkonen A, Bulik, CM, Neale BM, Rose RJ, Rissanen A, Kaprio J. Body
dissatisfaction and drive for thinness in Finnish young adult twins. International
Journal of Eating Disorders (in press).
IV Keski-Rahkonen A, Neale BM, Bulik CM, Pietiläinen K, Rose RJ, Kaprio J, Rissanen
A. Intentional weight loss in young adults: sex-specific genetic and environmental
A additive genetic effects
ACE a structural equation model that consists of three parameters: additive genetic
effects, common environmental effects, and non-shared environmental effects
ADE a structural equation model that consists of three parameters: additive genetic
effects, dominance effects, and non-shared environmental effects
BD body dissatisfaction
BMI body mass index
C common environmental effects
CI confidence interval
D dominance effects
DT drive for thinness
DZ dizygotic twins
E non-shared environmental effects
EDI Eating Disorder Inventory
GHQ General Health Questionnaire
IWL intentional weight loss
MZ monozygotic twins
OR odds ratio
SE standard error
SES socioeconomic status
“Food is the new sex.”
Anonymous magazine headline, circa 2001
“Fat can be fatal. Obesity is the great new global health scare.... The danger is baffling because
it is paradoxical. For ours is the most diet-conscious era and diet-obsessed culture in the
history of the world. We think thin and we get fat.”
Felipe Fernandez-Armesto, The Guardian, Sept 14, 2002
The current Western popular culture sends powerful and conflicting messages about food and
weight. On one hand, indulgence is encouraged: celebrity chefs, lushly edited coffee-table
cookbooks, magazine extras, and food commercials promote the sinful goodness of sugar and
fat, backed by multibillion marketing budgets and governments grappling with agricultural
overproduction (Schlosser 2002; Knapp 2003; Critser 2003). On the other hand, it is common
knowledge that overindulgence results in overweight and obesity, implying serious health risks
(Pi-Sunyer 1991) and social stigma (Harris 1983; Turnbull et al. 2000).
Over the past millennia, the relative scarcity of highly nutritious food may have given an
evolutionary advantage to individuals with “thrifty genes”, ie, the capability of effective long-
term energy storage (Neel 1962; Pijl 2003). The very recent and dramatic changes in the living
environments of Western countries have turned our societies highly obesogenic. If genes load
the gun and the environment pulls the trigger, the question of obesity poses a double-barreled
challenge. The genetic propensity for obesity is fed by the environment. The result emerges as
a rapidly unfolding obesity epidemic.
In this obesogenic and fat-phobic society, weight control strategies assume paramount
importance in people’s lives. The market for quick fixes to weight problems is ever expanding,
but there are few good long-term weight loss solutions. The discrepancy between ideals and
reality causes much psychological distress (Lautenbacher et al. 1992; Rierdan & Koff 1997).
Even an ascetic and disciplined attitude toward food is not necessarily safe: restrictive eating
attitudes can become severe clinical eating disorders, or encourage disordered eating, such as
overeating and unhealthy weight control methods which may increase the risk of obesity.
It is not exactly easy to grow up in this conflicting cultural climate. Moreover, as traits
predisposing to obesity are the result of genes, environment, and the interaction of genes and
environment, individuals with certain genetic make-ups seem to be more vulnerable to
obesogenic environmental influences than others. Thus there is an increased need to explore
genetic and environmental contributions to various weight and eating related phenotypes.
This study focuses on adolescents and young adults who are trying to come to terms with
different aspects of eating patterns, body image, and weight control. Methods of genetic
epidemiology, particularly twin analyses, as well as those of classical epidemiology are
implemented in studying genetic, environmental, and behavioral correlates of disordered
eating (here exemplified by breakfast skipping), body shape and weight concerns, and
intentional weight loss attempts.
2. Literature Review
The weight regulation mechanisms in humans are highly asymmetrical: various pathways of
appetite control render weight loss difficult, but safeguards against weight gain are few
(Blundell & Gillett 2001; Flier 2004). This asymmetry makes many individuals in our current
environment vulnerable to obesity. The risk of obesity is often addressed by weight control
behaviors. Sometimes, however, weight control behaviors exacerbate rather than solve
problems: dieting attempts may spiral out of control, resulting in clinical eating disorders, or in
The genetic and molecular basis of weight control is very poorly understood.
Interindividual variation in weight control behaviors is considerable, probably because these
behaviors are a result of complex genetic and environmental interplay. Yet a deeper
understanding of these behaviors might give us more effective means of combating the
In the following three sections, three weight control behaviors, disordered eating (as
exemplified by breakfast skipping), body shape and weight concerns, and intentional weight
loss attempts, are defined and quantified. Second, existing research into demographic and
behavioral correlates of these behaviors is reviewed. Third, what is already known about the
genetics of various weight control related behaviors is briefly summarized.
Definitions and prevalences
The term disordered eating emerged in medical and psychological literature in the late 1970s,
coinciding with the introduction of diagnostic criteria for bulimia nervosa (Russell 1979).
Disordered eating was first used to describe dietary chaos and emotional instability experienced
during recovery from anorexia nervosa (Palmer 1979). Soon, the term was used more loosely
to describe young women, who “…diet at some time and lose more than 3 kg in weight; […]
may experience episodes of binge eating and "picking" behavior; […] wish to be thinner
irrespective of their current body weight [,] and abuse laxatives or diuretics in order to achieve
a fashionably slim Figure (Abraham et al. 1983).” Another early study defined disordered
eating as “bingeing, highly restrictive dieting, emotional eating, or purging (Kagan & Squires
Although the concept still lacks uniform definition, it is generally used to describe
disordered eating behaviors that are more broad and benign than eating disorders defined in
ICD-10 and DSM-IV classifications. In contrast to these diagnostic classifications, milder
forms of disordered eating are often not considered illnesses worthy of medical attention,
although they are relatively common among adolescents and young adults in the general
population. In a national survey conducted in the U.S., 13% of girls and 7% of boys reported
disordered eating, defined as self-reported binge-purge cycling (Neumark-Sztainer & Hannan
2000). In a Finnish nationwide school health survey, loosely defined bulimic-type eating
behavior was reported by 16.5% of girls and 12.3% of boys, although discrepancies in how the
study subjects understood the concept bingeing may possibly have inflated these prevalence
estimates (Beglin & Fairburn 1992). The prevalences of these behaviors may be higher still in
some particularly vulnerable subgroups, such as dancers (Abraham 1996), athletes (Johnson
1994), vegetarians (Neumark-Sztainer et al. 1997b), and patients with juvenile-onset diabetes
(Neumark-Sztainer et al. 1996b; Rydall et al. 1997).
Breakfast skipping, a commonly used weight control practice among adolescents (Neumark-
Sztainer & Story 1998) and adults (Levy & Heaton 1993), can also be considered a mild form
of disordered eating – it commonly coexists with body shape and weight concerns and other
symptoms of subclinical eating disorders (Melve & Baerheim 1994; Pastore et al. 1996). If
present in early adolescence, it may constitute a risk factor or an early manifestation of eating
disorders (Fernandez-Aranda et al, manuscript in preparation).
Breakfast skipping has been of interest for many researchers in the recent years, perhaps
because it is a question often addressed in large-scale population surveys. In scientific
literature, there is much more information on breakfast skipping than skipping lunch or
dinner, perhaps because a morning meal is more homogenous across cultures and populations
than the other main meals of the day. Skipping one of the daily main meals also predisposes to
skipping another meal: breakfast and lunch/dinner skipping often coexist (Urho & Hasunen
1999; Sjöberg et al. 2003). Breakfast skipping has been variously defined in different studies:
usually it means the omission of the first morning meal, intended to be consumed at home.
Sometimes specific time points are also used. As eating habits vary greatly in different
countries and cultures, comparisons are often difficult.
Breakfast skipping has become increasingly widespread among children, adolescents and
adults in Western countries during past 30 years (Haines et al. 1996; Nicklas et al. 1998; Siega-
Riz et al. 1998). Lifestyle changes may play a role in this trend. In Finland, breakfast used to be
a substantial, cooked meal (Prättälä & Roos 1999). As majority of women now have full-time
careers, there is less time and opportunities for family meals, and the traditional three-meal
pattern has changed into a pattern of one or two daily meals (Prättälä & Roos 1999).
Breakfasts have become lighter, consisting most often of bread, cheese, coffee or tea, and
contributing typically less than 20% of the daily energy intake (Kleemola et al. 1997). In the
most recent FINRAVINTO survey, breakfast accounted for 16% of the daily energy intake in
adult men and women (Männistö et al. 2003). When adolescents were asked why they skip
breakfast, “lack of time” or “not hungry” were the reasons supplied by the majority of them
(Shaw 1998; Urho & Hasunen 1999). Many Finnish adolescents may skip breakfast simply
because they will have an early school lunch or may purchase a snack whenever it feels
Prevalences of breakfast skipping in adolescents are detailed in Table 1. In studies
conducted in different countries, estimates of breakfast skipping among adolescents have
ranged from 3% to 66% (Terre et al. 1990; Neale & Cardon 1992; al Sudairy & Howard 1992;
Isralowitz & Trostler 1996; Samuelson et al. 1996; Brugman et al. 1998; Höglund et al. 1998;
Shaw 1998; Siega-Riz et al. 1998; Urho & Hasunen 1999; Cavadini et al. 2000; Murata 2000;
O'Dea & Caputi 2001; Abalkhail & Shawky 2002). Some of this great variability is due to age
and gender effects, but methodological and cultural differences are a more likely source of
Table 1. Prevalence of breakfast skipping in adolescents
Age N % breakfast
Study setting Author How was breakfast skipping
14-15 7605 20-32% Göteborg,
Höglund ”less than 3 days a week”
16-18 not reported,
Terre “usually skip breakfast”
15-18 1513 47-66% Israel Isralowitz not detailed
15-18 5243 35% Bogalusa, LA,
Siega-Riz “consumption of food, beverage or
both between 0500 and 1000”,
13-15 876 13% The Netherlands Brugman 24h recall
13 699 12% Australia Shaw questionnaire, not detailed
12-15 800 15% Jeddah City,
Abalkhail questionnaire, not detailed;
16-18 114 20% Riyadh,
al Sudairy questionnaire, not detailed;
15 411 5% Trollhättan and
Samuelson “less than 5 x/week”,
7-day dietary record
16-19 419 3% Lorraine, France Michaud 2-day dietary record
13-16 3248 15% Finland
Urho "24-h recall"
7-14 the boys in
4% Tokyo, Japan Murata not detailed
15-19 the boys in
18% Tokyo, Japan Murata not detailed
6-12 466 8% girls,
New South Wales,
O’Dea “whether on most days they
12-19 660 24% girls,
New South Wales,
O’Dea “whether on most days they
usually [did not] consume
variability. For instance, some studies assess breakfast eating using 24-hour dietary recall;
others use questionnaire items that are much more vaguely defined. All studies are liable to
Considerably fewer studies have assessed breakfast skipping in adult populations. A
Finnish study published almost a quarter-century ago found that 23% of Finnish adult women
and 33% of adult men reported skipping breakfast (Puska & Smolander 1980); in 2003, 17%
of Finnish adult women and 21% of adult men reported breakfast skipping (Helakorpi et al.
2003). In a similar large population study, conducted in 1987-98 in the Netherlands, 34% of
women and 38% of men skipped breakfast. In contrast, in a recent 24-h dietary recall based
population study of American adults (Ma et al. 2003), only 3.6% of participants were classified
as breakfast skippers. However, in that study, 18.9% of the breakfasts were eaten away from
home, usually on the way to work or school, reflecting perhaps a fast food lifestyle more
typical of the U.S. than Northern Europe. In many self-report questionnaire assessments of
breakfast eating, a morning meal eaten away from home is either implicitly or explicitly
defined as a snack, not breakfast (Gatenby 1997).
Body shape and weight concerns
In modern Western societies, high-energy food is abundant, and the necessity for physical
activity is decreasing. Body shape ideals for women and men are gravitating towards those that
are the most difficult to obtain and maintain: thinness, leanness, and muscularity (Gordon
2000; Morgan 2000). The discrepancy between ideals and reality results in body shape and
weight concerns, evident early in childhood (Sands et al. 1997; Schur et al. 2000) and
widespread in adolescents and young adults (Moore 1988; Heatherton et al. 1997; Jaeger et al.
A variety of approaches have been used to measure body shape and weight concerns.
Sometimes assessment by a simple question is deemed sufficient. A more indirect approach is
to contrast actual body weight and shape to ideal body weight and shape; for instance, an array
of silhouettes of varying degrees of muscularity and leanness can be used (Bulik et al. 2001).
Of self-report measures, the Eating Disorder Inventory (EDI) (Garner & Olmsted) Body
dissatisfaction (BD) and Drive for Thinness (DT) subscales (Table 2) are widely used and
The risk of body and weight dissatisfaction increases with increasing body mass,
particularly in the presence of strong thinness idealization, teasing and bullying, or if there is
external pressure to be thin (Stice & Whitenton 2002; van den Berg et al. 2002). Nevertheless,
many people who are not overweight are still dissatisfied with their bodies. In a large
(N=19,841) and representative sample of Canadian adults, 32% of women and 10% of men
who were at normal weight (BMI 20-24) were still trying to reduce their weight (Green et al.
1997). In a national survey conducted in the US, 47% of women currently trying to lose
weight had a body mass index under 25 (Biener & Heaton 1995). Similarly, in a school-based
Table 2. Three subscales of Eating Disorder Inventory 1 (Garner 1991). All items were
rated on a 6-point Likert scale (always, usually, often, sometimes, rarely, never).
Drive for Thinness
I eat sweets and carbohydrates without feeling nervous.
I think about dieting.
I feel extremely guilty after overeating.
I am terrified of gaining weight.
I exaggerate or magnify the importance of weight.
I am preoccupied with the desire to be thinner.
If I gain a pound, I worry that I will keep gaining.
I eat when I am upset.
I stuff myself with food.
I have gone on eating binges where I felt that I could not stop.
I think about bingeing (overeating).
I eat moderately in front of others and stuff myself when they are gone.
I have the thought of trying to vomit in order to lose weight.
I eat or drink in secrecy.
I think that my stomach is too big.
I think that my thighs are too large.
I think that my stomach is just the right size.
I feel satisfied with the shape of my body.
I like the shape of my buttocks.
I think my hips are too big.
I think that my thighs are just the right size.
I think my buttocks are too large.
I think that my hips are just the right size.
survey of Finnish adolescents aged 14-16 (N=60,252) 46% of girls and 34% of boys were not
satisfied with their weight; among normal-weight adolescents dissatisfied with their weight,
81% of girls and 48% of boys thought they were overweight (Mikkilä et al. 2003).
Body dissatisfaction is a risk factor for eating disorders and disordered eating, particularly
bulimic behavior patterns; the risk is probably mediated through dieting and negative affect
(Stice 2001; Stice & Shaw 2003). However, at least in young women, dissatisfaction with body
weight and shape is currently so widespread that it is almost normative. Estimates of rates of
body dissatisfaction vary with measures used. In a Southern Italian study of students from
secondary schools, 58.4% of girls and 19.7% of boys displayed dissatisfaction with regard to
their own body (Dalla Grave et al. 1997). Of Norwegian students of comparable age, 37% of
girls and 20% of boys reported that they felt “a bit or much too fat” (Borresen & Rosenvinge
2003). Body shape and weight ideals are also strongly culture dependent. In the European
context, the most extreme body dissatisfaction is evident in Northern Mediterranean
countries, followed by Northern European countries (Jaeger et al. 2002). Levels of body and
weight dissatisfaction are intermediate to high in women living in cultures currently
undergoing westernization (Abdollahi & Mann 2001; Becker et al. 2002; Tsai et al. 2003;
Sarlio-Lähteenkorva et al. 2003). Women in non-western countries demonstrate rather low
levels of body dissatisfaction (Jaeger et al. 2002). However, there is some evidence from
Finland that although obesity and overweight are becoming increasingly common, the level of
weight dissatisfaction in the general population of adolescents is actually decreasing (Kaltiala-
Heino et al. 2003a).
Intentional weight loss
Body shape and weight concerns are inextricably associated with weight loss attempts. In this
study, we focus on weight loss, because it is less vague and easier to quantify than dieting.
Intentional weight loss is often used to make a distinction to unintentional weight loss, such as
weight loss resulting from malignancies, endocrine abnormalities, or severe depression
(French et al. 1995a; Meltzer & Everhart 1996; Williamson 1997).
In an American population study conducted in 1989, approximately 40% of women and
25% of men were trying to lose weight; the average weight loss goal was -14 kg (Williamson et
al. 1992); more recently, 44% of women and 29% of men reported attempting to lose weight
(Serdula et al. 1999).
Surprisingly little is known about factors that motivate weight loss and means taken to lose
weight. In the Weight Loss Practices Survey (Levy & Heaton 1993), conducted among a
representative sample of adults in 1992 in the U.S., 33% of women and 20% of men were
currently trying to lose weight. Weighing oneself regularly was the most common currently
used means of weight loss (reported by 71% of women and 70% of men, respectively),
followed by walking (58% and 44%), using diet soft drinks (52% and 45%), taking vitamins
and minerals (33% and 26%), counting calories (25% and 17%), and participating in organized
weight-loss programs (13% and 5%). Less healthful practices that also reflect the population
prevalence of disordered eating in adults included skipping meals (21% and 20%), taking diet
pills (14% of women and 7% of men), fasting (6% in both sexes), misuse of laxatives (3% of
women and 1% of men), use of dieting devices (1% in both sexes), and vomiting (1% of
women, 0% of men). In another and more recent large American study (Neumark-Sztainer et
al. 2000), 57% of adult women, 50% of adult men, 44% of adolescent girls, and 37% of
adolescent boys reported current weight control behaviors. Disordered eating behaviors
included meal skipping (in 19% of women and men, 23% of girls, and 14% of boys currently
trying to control their weight), fasting (3% of women and men, 8% of girls, 7% of boys), diet
pills or laxatives (7% of women, 3% of men, 4% of girls, 3% of boys), and vomiting (<1% of
women and men, 7% of girls, 2% of boys). Thus it seems that a substantial minority of
individuals trying to control their weight engage in questionable weight control practices that
fall within the boundaries of disordered eating.
Although many people report that health concerns or health improvement are their
primary motivation for weight loss, particularly if they are overweight, dieting for appearance
reasons in the absence of overweight is very common, as noted above, particularly among
young women (Levy & Heaton 1993; Biener & Heaton 1995). Paradoxically, weight loss
attempts are rarely successful, and “weight cycling” (repeated weight loss and weight gain) may
have harmful consequences (Brownell & Rodin 1994; Weight cycling. National Task Force on
the Prevention and Treatment of Obesity 1994). Women also generally have lower BMI goals
than men, and normal-weight individuals have lower BMI goals than overweight ones
(Anderson et al. 2003).
Weight control behaviours: correlates and demographic determinants
Operating on the hypothesis that weight control behaviors may also have harmful, undesired
consequences, numerous studies have explored the relationship of weight control behaviours
and health-compromising behaviors. There is considerable common interface particularly
between weight control behaviors and psychological and psychiatric problems (Figure 1).
Figure 1. Schematic drawing of associations of weight control behaviors: although
weight control is often considered healthful, some psychiatric and psychological
problems and health-compromising behaviors, particularly substance use, are
sometimes associated with weight control.
Below, a more detailed overview focuses on three core areas of importance for this study,
disordered eating (with particular focus on breakfast skipping), body shape and weight
concerns, and intentional weight loss.
Disordered eating and peripubertal problems
Obesity in childhood and adolescence and subsequent body dissatisfaction are well-known
risk factors of eating disorders and disordered eating (Fairburn et al. 1997; Ackard & Peterson
2001). They are particularly important precursors of bulimic behavior patterns; the risk is
probably mediated through dieting and negative affect (Stice 2001; Stice & Shaw 2003). A
common thread of these problems seems to be their peripubertal manifestation. Adolescents
who experience puberty earlier than their peers seem to be at higher risk for body
dissatisfaction, eating disorders, meal skipping, and many psychological problems (Fairburn et
al. 1997; Kaltiala-Heino et al. 2001; Sjöberg et al. 2003; Kaltiala-Heino et al. 2003b).
Abnormal eating attitudes and disordered eating have been found to cluster with problem
behavior, eg, tobacco, alcohol, substance use, and risk-taking in general (Fisher et al. 1991;
Chandy et al. 1994; Chandy et al. 1995; Neumark-Sztainer et al. 1997c; von Ranson et al.
2002). Suicidal attempts, self-harm, delinquency, school problems, sexual abuse, date rape and
violence, and high-risk sexual behaviours are also associated with disordered eating (Neumark-
Sztainer et al. 1996a; Neumark-Sztainer et al. 1998; Ackard & Peterson 2001; Wonderlich et al.
2001; Ackard & Neumark-Sztainer 2002). In addition to suicidal behaviours, also other
psychosomatic symptoms are more common in individuals with disordered eating attitudes
(Buddeberg-Fischer et al. 1996). In fact, preoccupation with weight and/or dieting concerns in
male and female adolescents is likely to indicate psychological problems (Casper 1998;
Neumark-Sztainer & Hannan 2000). In a population sample of 2525 Australian teenagers, 7%
of girls and 1% of boys fell into a group of extreme dieters: of them, 62% reported high levels
of depression and anxiety; these levels are fully comparable with the psychiatric comorbidity
of clinical eating disorders.
Weight dissatisfaction seems to be more prevalent in girls from low SES families, at least
in a large (N=60,252) Finnish school-based survey (Mikkilä et al. 2003). Similarly, weight
dissatisfaction was associated with more smoking in girls; smoking is perhaps used as a chosen
method of weight control. Interestingly, health behaviours seems to be associated more with
perceived weight/weight satisfaction than with actual body weight (Mikkilä et al. 2003).
Behavioral and demographic correlates of breakfast skipping
Meal skipping shares the timing of peripubertally initiated disordered eating behaviors:
perhaps reflecting differential parental supervision of the morning meal, children in primary
school are usually at a lower risk than teenagers (Brugman et al. 1998; Nicklas et al. 1998;
Murata 2000; O'Dea & Caputi 2001). Similarly than other disordered eating behaviors,
breakfast skipping is associated with health-compromising behaviors, such as smoking
(Höglund et al. 1998; Sjöberg et al. 2003) and alcohol and drug use (Isralowitz & Trostler
1996). Conversely, regular breakfast eating has been associated with a health-conscious
lifestyle and regular exercise (Cavadini et al. 2000).
Probably the most widely researched health risk associated with breakfast skipping is
obesity and related weight control issues. Individuals who skip meals tend to compensate by
snacking (Urho & Hasunen 1999; Sjöberg et al. 2003). Meal-skippers seem to end up with less
healthful food choices, significantly lower intakes of micronutrients, poorer overall nutrient
intake, but higher intakes of sucrose and alcohol compared to the groups with regular meal
intake (Cho et al. 2003; Sjöberg et al. 2003). Given that meal skipping is often associated with
a sedentary lifestyle (Terre et al. 1990; Baumert, Jr. et al. 1998; Boutelle et al. 2002), it has been
hypothesized that meal skipping, particularly breakfast skipping, may be a risk factor of
obesity. Many cross-sectional studies have observed this association (Kaufmann et al. 1975;
Wolfe et al. 1994; Nordlund & Jacobson 1999; Boutelle et al. 2002; Cho et al. 2003), but this
finding has not been confirmed by longitudinal studies (Berkey et al. 2003).
Some experts have recommended breakfast skipping to obese individuals because it may
limit total daily energy intake (Schlundt et al. 1992). Indeed, baseline breakfast eaters lost more
weight if they started skipping breakfast rather than continued to have it. Baseline breakfast
skippers, however, lost more weight if they started to have regular morning meals (Schlundt et
al. 1992). Several other studies have successfully used the introduction of regular breakfasts as
a weight loss intervention, hypothesizing that regular meals limit excessive energy intake from
impulsive snacks (Mattes 2002; Falkenberg & Hellman 2002; Wyatt et al. 2002). Indeed,
regular breakfast eating, rather than breakfast skipping, seems to be characteristic of
individuals who can maintain weight loss over long periods of time (Wyatt et al. 2002).
Despite these findings, meal skipping remains a popular weight control strategy (Levy &
Heaton 1993; Bellisle et al. 1995; Neumark-Sztainer & Story 1998). Some studies suggest that
adolescent girls may be more likely to diet by skipping breakfast or lunch and by decreasing
their meal size; boys, on the other hand, tend to limit snacking and high-energy foods in order
to lose weight (Brugman et al. 1997; Nowak 1998; Shaw 1998). However, boys who were
encouraged to diet by their mothers were at risk for health-compromising eating and dieting
behaviors, particularly skipping meals, fasting, and binge-eating (Fulkerson et al. 2002). In
other studies, adolescent girls who skipped meals were more likely to diet and expressed
greater levels of body shape and weight dissatisfaction (Shaw 1998; Mikkilä et al. 2003). In
Norway, 15-19-year-old girls with subclinical eating disorders skipped breakfast more often
than healthy control subjects (Melve & Baerheim 1994). Meal skipping is often an early
manifestation of an eating disorder, and if present before puberty, a risk factor of anorexia
(Fernandez-Aranda, in preparation).
Irregular meal patterns have traditionally been associated with food insecurity and
deprivation. Following the general patterns of SES and food behavior (Roos et al. 1998; Irala-
Estevez et al. 2000; Robinson et al. 2004), family SES and educational attainment are clearly
linked with breakfast eating patterns. In Western countries, several previous studies have
found and association between breakfast skipping and low socioeconomic status (Pastore et al.
1996; Brugman et al. 1998; Höglund et al. 1998; Nordlund & Jacobson 1999; O'Dea & Caputi
2001; Robinson et al. 2004), although this finding is not consistent (Walker et al. 1982;
Brugman et al. 1998; Höglund et al. 1998; Nordlund & Jacobson 1999; O'Dea & Caputi 2001).
In developing countries and in countries where income inequality is large, poverty has been
the most important reason for children and adolescents to skip breakfast (Melnik et al. 1998;
Tzimis & Kafatos 2000; Gross et al. 2004). Given that multiple studies link cognitive problems
(decreased ability to concentrate and memory problems) to missing breakfast, special school
breakfast programs and other community-level interventions have been implemented in many
low-income settings (Murphy et al. 1998; Pollitt & Mathews 1998; Powell et al. 1998).
However, in Finland and other affluent societies, lack of economic resources in itself does
probably not explain breakfast skipping.
Weight control behaviours
As already discussed above, weight control behaviours are associated with disordered eating
and body shape and weight concerns. Although it would be natural to assume that weight
control behaviours are triggered by overweight and obesity, they behaviors are actually
relatively common among normal-weight individuals, especially women (Strauss 1999; Wardle
& Johnson 2002). Maladaptive weight control strategies are particularly common among low
SES women: the higher risk of obesity and family exposure to these behaviors places women
at increased risk of unhealthy behaviors (Breitkopf & Berenson 2004).
Adolescents' perceptions of direct pressure from their parents to diet has been found to be
a significant predictor of dieting, and perceived parental encouragement of autonomy, and
self-confidence were associated with less dieting behaviour (Huon & Strong 1998). Adolescent
dieters have poorer self esteem than their non-dieting counterparts (Pesa 1999).
Like body shape and weight concerns, weight loss has been linked with depression and
low mood. This probably has a physiological basis: extreme starvation has severe
psychological consequences, such as depression and anxiety (Keys et al. 1950), and
experiments of extreme weight loss have triggered low mood, heightened irritability,
difficulties concentrating, and fatigue (Laessle et al. 1996). In some studies, weight cycling (ie,
repeated weight gain and weight loss) has been associated with psychological problems
(Venditti et al. 1996). Mostly, this link has been refuted, and it may be that the psychological
discomfort is mediated rather by binge eating than dieting or by the individuals perception of
being a weight cycler rather than actual weight loss (Telch & Agras 1994; Bartlett et al. 1996;
Foster et al. 1996; Venditti et al. 1996; Foster et al. 1997; Friedman et al. 1998; Kensinger et al.
1998). Severe binge eating is often associated with severe comorbid psychopathology (Grilo et
al. 2001). However, many recent studies confirm that in nonclinical samples, the rates of
psychopathology do not appear significantly greater in obese than nonobese individuals
(Wadden & Stunkard 1985; Stunkard & Wadden 1992; Fitzgibbon et al. 1993), even though
obesity often carries severe social stigma and is associated with a low degree of life satisfaction
(Rosmond & Björntorp 1998).
Very little is known about behavioural correlates of weight loss on the population level. A
meta-analysis of mostly observational studies of structured weight loss programs in the U.S.
revealed that the results were relatively modest at best: five years after completing the
programs, the average individual maintained a weight loss of about 3 kg (Anderson et al.
2001). In a Finnish study, only 6% of all overweight individuals lost at least 5% of their weight
and maintained this weight loss over nine years (Sarlio-Lähteenkorva et al. 2000).
It seems that only extreme behavioural modifications yield long-term weight maintenance
after weight loss: usually this entails continued attention to body weight, low levels of stress,
consumption of a low-energy and low-fat diet, and continued physical activity (Klem et al.
1997; Shick et al. 1998; Sarlio-Lähteenkorva et al. 2000). It is also possible that long-term
weight-maintainers experience higher levels of anxiety than obese individuals who do not
attempt weight loss (Sarlio-Lähteenkorva & Rissanen 1998). Weight-loss maintainers seem to
use more cognitive and behavioral strategies to control their weight than weight-regainers or
weight-stable controls (McGuire et al. 1999a; Dohm et al. 2001). Factors that seem to predict
weight regain include very recent and very large weight losses, higher levels of depression,
dietary disinhibition, binge eating, and overall stress (McGuire et al. 1999b; Sarlio-
Lähteenkorva et al. 2000).
If the long-term success of weight loss is rather pessimistic, it is also unclear whether
individuals who self-report dieting actually have a lower energy intake and a higher energy
expenditure than their non-dieting counterparts. Although most research in this field has been
conducted in experimental studies, a few population studies, all conducted in the U.S., are
available. In adolescents, moderate weight control methods seem to be associated with more
health-promoting and exercise behaviours than extreme weight control methods (Story et al.
1998). However, in another study, dieting frequency was positively associated with concern
about weight and shape, but not with physical activity; energy intake was lower in dieting high
school students than their non-dieting counterparts, but in preadolescents “dieting” did not
imply any significant differences in energy intake (Field et al. 1993). Dieting was also
associated with unhealthy weight control practices (diet pill, laxative, and vomiting) and
psychological changes (higher BD and DT and more frequent alcohol use) (French et al.
1995b). In adults, the association of exercise and dieting varies from one study to another –
women dieters may be more likely to exercise than their non-dieting counterparts, but often
self-reported dieting makes no difference in energy expenditure (French et al. 1994; Biener &
Heaton 1995). Weight loss or weight maintenance attempts were associated with lower
percent of energy intake from fat and sweets, more frequent consumption of vegetables and
fruits, more frequent self-weighing, and lower tolerance for weight gain prior to taking action
(French & Jeffery 1997).
The great pains and concerns invested in weight control reveal how toxic and obesogenic our
current environment is. The huge difficulty of successful weight loss underscores the
underlying biologically hardwired asymmetry in weight regulation, which may be compounded
by environmental influences, such as socioeconomic disadvantage. Moreover, interindividual
variation in weight control behaviors seems great: the ground is not level for different
individuals in terms of weight loss. The same amount of energy consumed do not mean the
same reasons for different individuals, but the genetic mechanisms responsible for this
discrepancy are largely unclear. These issues will be further explored in the next chapter.
Figure 2. Components of the weight regulation system: The brain integrates long-term
afferent signals from fat (leptin) and pancreatic beta cells (insulin) with short-term,
meal-related afferent signals from the gut, such as inhibitors of feeding (PYY, GLP-1,
and CCK), and stimulator of feeding (ghrelin). Efferent outputs regulate appetite,
energy expenditure, hormonal milieu, energy partitioning, and the status of
reproduction and growth. (Reprinted from Flier, J. S. Obesity wars: molecular progress
confronts an expanding epidemic. Cell 2004:116, 337-350, with permission from
Familial aggregation and heritability
Obesity and eating disorders are likely disorders characterized by etiological heterogeneity.
The risk of obesity or eating disorders is not equal for all individuals: it is probably based on a
complex interplay of individual-specific genetic vulnerability and different environmental
exposures. Identical environmental exposures (eg, the same amount of energy consumed) may
result in dissimilar outcomes because of genetic variation; conversely, identical genes may also
result in dissimilar outcomes in the presence of different environmental influences. Some
currently known pathways of weight regulation are briefly introduced in Figure 2. As
comparatively little is known about the genetics these pathways, defining eating-related
phenotypes and examining them in using twin and family designs is a useful first piece in
understanding the relative contributions of genes and environment to weight control
Family studies are used to assess whether a trait of interest runs in families. Twin and
adoption studies can help to understand the extent to which familial resemblance is caused by
genetic influences and what part of the resemblance is environmental in origin. Understanding
how genes and environments work together will increase our options to correctly target
interventions to modify harmful behavioural patterns, such as obesity and its consequences.
This section attempts to review some of the most essential research findings pertaining to
genetic influences on eating patterns and energy intake, weight loss and weight change, body
shape and weight concerns, and disordered eating.
Studies of familial aggregation of weight control related behaviors have focused on a few
specific themes. There are numerous twin and family studies of BMI: all these studies agree
that BMI is substantially heritable, with possible age and gender specific effects (for a recent
combined analysis from eight twin registries, see Schousboe et al. 2003b; for a review of earlier
twin studies of BMI, see Maes et al. 1997).
Genome-wide linkage scans of BMI and obesity are also starting to emerge and have
identified chromosomal regions possibly containing predisposing genes, although findings
have not been consistent. This is probably due to relatively small sample sizes as well as real
differences between populations (Adeyemo et al. 2003; Saar et al. 2003). Some gene mutations
giving rise to relatively rare forms of obesity are already known (Farooqi et al. 2003; Branson
et al. 2003; Suviolahti et al. 2003; Snyder et al. 2004). A few studies also assess the heritability
other measures of body composition, such as truncal and extremity skinfolds, waist
circumference, hip, and total body fat %, obtaining estimates that are very much in line with
the heritability estimates of BMI (Maes et al. 1997; Rice et al. 1999; Schousboe et al. 2003a).
Genetics of weight change
Little is known about genetic factors that influence weight change. Analyses of genetic
influences on weight change, dieting, weight loss, and ultimately permanent weight change are
fraught with difficulties. As noted previously, permanent long-term weight loss is difficult to
achieve: in a recent Finnish study, only 6% of overweight individuals were able to maintain a
5% loss of body weight over a period nine years or more (Sarlio-Lähteenkorva & Rissanen
1998). Dieting is difficult to operationalize reliably (Neumark-Sztainer et al. 1997a), and weight
changes in the general population may have several other plausible explanations, such as
pregnancies and diseases, that may be confounded with intentional weight loss attempts.
However, a few existing studies have attempted to resolve these difficult issues.
Family analyses conducted in the Quebec Family Study showed that BMI change over 12
years had a heritability of 37%, and change in total subcutaneous fat (as measured by
skinfolds) over the same period had a heritability of 16% (Rice et al. 1999). Similarly, extended
pedigrees of the Framingham study population revealed that longitudinal changes in body
weight change had a heritability of 15%, and body weight change up to age 50 a heritability of
22% (no standard errors reported; Golla et al. 2003), and the heritability of difference between
maximum and mean BMI was 23% (95% CI: 9-37%) (Coady et al. 2002). Some susceptibility
areas for genes possibly influencing long-term body weight regulation were identified in the
long arm of Chromosome 8 (Coady et al. 2002; Golla et al. 2003).
A factor analytic solution of the Eating Attitudes Test (EAT) produced a “dieting” factor
with an estimated heritability of 41% (no standard errors were reported) in a volunteer sample
of young female British twins (Rutherford et al. 1993), implying that some genetic influences
on dieting were evident. Some evidence of genetic influences on weight change was found in a
Swedish twin sample (Heitmann et al. 1999). Although the net population weight change over
a period of eleven years was positive and fairly consistent across different zygosity groups, no
information was given on subgroups of weight trajectories. Thus the study population is
probably a genetically heterogenous mixture of individuals who gain weight, lose weight, and
maintain their weight, making understanding the construct of weight change and drawing
conclusions about it very challenging. Furthermore, the study sample size was relatively small
(N=98 MZ and 176 DZ twin pairs).
Another study, conducted in Finland, also attempted to understand genetic influences on
weight change utilizing longitudinal twin models (Korkeila et al. 1995). In this study, the
phenotype was more clearly defined: the type and directionality of weight change were
specified. The analyses suggested that additive genetic influences on BMI remain relatively
stable throughout adulthood, but that the influence of genetic effects on a 6-year weight
change was modest at best. Twins who cohabited in adulthood were more likely to be similar
in regard to weight change than twins who lived separately, suggesting some role of familial
environment (Korkeila et al. 1995). In later studies of the same population, it was found that
genetic factors may modify the effects of physical activity on weight change, in particularly so
that sedentary lifestyles may have an obesity-promoting effect in men with a genetic
predisposition (Heitmann et al. 1997), and that genetic and familial factors may contribute to
major weight gain after weight loss attempts (Korkeila et al. 1999). Education level also plays a
role in weight change (Silventoinen et al. 2004).
Genetics of disordered eating
In contrast to weight change, there is a lot of research on the genetics of disordered eating.
Very early in life, overeating eating style seems to be passed on from parents to children
(Wardle et al. 2001). In adults, bingeing on food is moderately heritable (Virginia Twin
Registry, women: 46%, 95% CI: 33-59%; Norwegian Twin Registry, both women and men:
51%, 95% CI: 43-58%) and self-induced vomiting is highly heritable (72%, 95% CI 55-88%)
(Sullivan et al. 1998; Reichborn-Kjennerud et al. 2003). There is substantial genetic overlap
between vomiting and bingeing (r= 0.74)(Sullivan et al. 1998), but much more modest overlap
between between bingeing and obesity (r=0.34, 95% CI, 0.19-0.50) (Bulik et al. 2003). Overall,
about 60% (95% CI 50-68) of the variance in liability to disordered eating is explained by
genetic factors (Wade et al. 1999). For clinical eating disorders, heritability estimates are similar
or higher (for reviews, see Fairburn et al. 1999; Bulik et al. 2000).
Several relatively recent studies have addressed the familiality and genetics of body shape
and weight concerns (Holland et al. 1988; Rutherford et al. 1993; Wade et al. 1998; Klump et
al. 2000; Wade et al. 2001; Wade et al. 2003; Reichborn-Kjennerud et al. 2004). The search for
candidate genes affecting eating disorders and weight-regulation has stimulated research in this
field, as variability in perceptions of body image may mark endophenotypes (ie, intermediary
phenotypes) of eating disorders (Devlin et al. 2002; Gottesman & Gould 2003). Twin studies
of body shape and weight concerns are detailed in Table 3. These studies have used different
approaches, ranging from standard interviews (EDE: Fairburn & Cooper 1983), and validated
questionnaires (EDI: Garner 1991, BAQ: Ben Tovim & Walker 1991) to assessing body image
by figural silhouettes (Wade et al. 2001) and to using a DSM-IV criterion of disordered body
image (“undue influence of weight on self-evaluation”) in a questionnaire form (Reichborn-
Kjennerud et al. 2004). To give a different cultural perspective, there is also a recent Japanese
twin study of EDI subscales (Kamakura et al. 2003). Regrettably, the authors only report result
from subscales that are not directly related to weight and shape concerns (maturity fears,
ineffectiveness, interpersonal distrust, interoceptive awareness, and perfectionism).
Considering the heterogeneity of the measures, it is not surprising that estimates of
heritability obtained from these studies vary considerably. However, heritability estimates for
the subscales of EDI, the most widely used measure of body and weight dissatisfaction, are
Table 3. A summary of twin studies of body image and body shape attitudes (adapted
from Wade et al. 2003).
1 Given only when CIs or SEs of the parameter estimates were reported in the original article.
2 Volunteer twin sample
3 EDE: Eating disorder examination (Fairburn & Cooper 1983)
4 Population-based twin sample
5 FRS: Figure rating scale (Stunkard et al. 1983)
6 BAQ: Body attitudes Questionnaire (Ben Tovim & Walker 1991)
7This study sample contains 3443 males and 4602 females; males and females were entered in the
models both independently (male and female parameters estimates freely) and by constraining males
and females equal. The model fitted the data better using the latter approach, thus parameter
estimates reported here are equal for males and females.
Study N Measure Additive
(95% CI) 1
(95% CI) 1
Holland et al,
45 EDI, global score 98 0 2
Rutherford et al,
246 EDI: BD
Wade et al,
174 EDE3: Weight concern
EDE3: Shape concern
Klump et al,
EDI: BD at 11 y
EDI: DT at 11 y
EDI: BD at 17 y
EDI: DT at 17 y
Wade et al,
5325 FRS5: current body size
FRS5: desired body size
Wade et al,
1788 BAQ6: Feeling fat
BAQ6: Body disparagement
BAQ6: Salience of Weight/Shape
BAQ6: Lower Body Fatness
80457 „Undue influence of weight on
0 31 (24-38) 69 (68-76)
very consistent in late adolescence and young adulthood: in a volunteer sample of 492 adult
female twins from the United Kingdom (Rutherford et al. 1993), and in the 17-year-old twins
from a population study conducted in Minnesota (Klump et al. 2000), EDI subscale scores
were moderately heritable. Unfortunately, their relatively small sample sizes, and the absence
of confidence intervals limit conclusions that can be drawn (Neale & Miller, 1997). With the
exception of the Norwegian study, none of these studies addresses body image in men.
Genetics of food intake
There are far fewer studies of genetics of food intake, perhaps because the phenotype is much
more complex (Keller et al. 2002; Keller et al. 2003). The largest one of these studies focused
on twins over 50 y using a food frequency questionnaire (van den Bree et al. 1999). Two
distinct eating patterns were identified: a preference for food items high in fat, salt, and sugar,
and a preference for healthful eating habits. Both of these patterns were highly
environmentally modulated, with heritabilities between 15% and 38% for the first pattern, and
33-40% for the latter pattern; clear gender differences were also evident. A Swedish twin study
focusing on twins aged 25-59 y found some evidence of genetic influences on the frequency
of intake of certain types of food, such as flour and grain products and fruit, but the majority
of food intake types seemed to be largely environmentally modulated (Heitmann et al. 1999).
In a small volunteer twin sample (110 MZ and 102 DZ twins, both male and female same-sex
pairs, and 53 pairs of opposite-sex DZ twins), food intake as recorded in food diaries was
examined: 65% of the variance in daily energy intake could be attributed to heritable factors;
the heritability of average meal size was also 65% and that of meal frequency 44% (standard
errors were not estimated) (de Castro 1993). Daily energy intake was largely genetically
independent from body size, and individual meal size was independent from daily intake. In
another study from the same sample, the heritability of the daily timing of the meal was 37%
(morning meal timing h2=24%, afternoon meal timing h2=18%, and evening meal timing
h2=22%) Intake during those meals was estimated to be 47% heritable for morning meals,
64% heritable for afternoon meals, and 58% heritable for evening meals: however, at this age
(30-50 years), no shared environmental effects were discernible (de Castro 2001). Another
American study, of Minnesota twins reared apart (ranging in age from 18 to 77 y, with a mean
of 42 years), found that 32% (95% CI: 10-51%) of the variance in daily energy consumed was
attributable to additive genetic effects. The heritability of meal frequency was 33% (95% CI:
11-52%) and that of snack frequency was 30% (95% CI: 9-49%); moderate spousal
correlations were also evident (0.30 for total energy consumed, 0.09 for snacking frequency,
and 0.40 for meal frequency) (Hur et al. 1998).
Genes have an important role on weight regulation; however, the specific mechanisms are
largely unknown. Although there has been considerable improvement in understanding the
genetics of BMI, obesity, eating disorders, and disordered eating, much less is known about
the genetics of other weight control behaviours, such as eating behaviors and weight loss.
Thus, further research is clearly warranted.
3. Aims of the study
The genetic and molecular basis of weight control behaviors is relatively poorly understood.
The complex interplay of genes and environment, great interindividual variation, and various
measurement issues make this task even more daunting. Twin and family studies can be used
to examine genetic and environmental contributions to weight control related phenotypes and
to level the ground for further exploration of genes in this area. In this study, three
phenotypes, breakfast skipping, body shape and weight concerns, and intentional weight loss,
The specific aims were:
1. To describe demographic and behavioural correlates of breakfast skipping, attitudes
about body shape and weight, and intentional weight loss (I).
2. To investigate familial patterns of breakfast skipping (II).
3. To estimate genetic and environmental components in the liability in breakfast
skipping patterns (II), body shape and weight concerns (III), and intentional weight
4. To assess to what degree the same genetic effects influence both BMI and IWL (IV).
4. Subjects and methods
To address the aims of this study, the unique twin-family data from the population-based
FinnTwin16 were used. FinnTwin16 comprises five consecutive nationwide birth cohorts of
twins born between 1975 and 1979 (Rose et al. 1999; Kaprio et al. 2002). Local ethics
committees in Finland and in the U.S. approved the collection of these data. At baseline, a
questionnaire was mailed to twins born in 1975 through 1979 within 2 months of their
sixteenth birthday. The questionnaire assessed personality, social relationships, general health,
health habits, and nutritional matters, including breakfast skipping. Three follow-up
assessments were conducted by postal questionnaire at the age of 17 y, 18.5 y, and as young
adults. In the fourth wave of data collection, each birth cohort of twins were contacted
semiannually between autumn 2000 (1975 cohort) and autumn 2002 (1979 cohort), and the
mean age was 24.4 y, range 22-27 yrs. Body dissatisfaction, drive for thinness, and intentional
weight loss attempts were assessed in the fourth wave of data collection. Response rates were
high (>80%) across all occasions.
A flowchart depicting different waves of questionnaires and the number of participants in
each wave relevant for these studies is presented in Figure 3. A total of 3065 families were
contacted. The parents of the twins received their own questionnaires which were mailed at
the same time as the baseline twin questionnaires. Of the 6130 twins in these families, 5561
(91%) returned the baseline questionnaire. Individual response rates were 93% for girls, 88%
for boys, 84% for fathers and 87% for mothers at baseline.
Information on mothers’ and fathers’ occupation, SES, lifestyle and health habits
(particularly breakfast skipping and weight were used (I, II, III); other variables of interest are
detailed below in the Measures section). At the time of the assessment, the age of the mothers
ranged from 32.2 to 62 (mean 44.3, SD 4.9) years, and fathers, respectively, from 33.6 to 69.8
(mean 46.5, SD 5.7) years.
At the time when the data for studies III and IV were analyzed, 4667 twins (84% of those
taking part in the first wave and 76% of the entire twin sample), 2545 females and 2122 males,
had returned their young adult questionnaires. Information on IWL was available from 4662
subjects, BD from 4545 and DT from 4596 subjects. After this, the response rate has further
increased still, but we could not include the late responders in these analyses.
Zygosity was determined by standard items included in the baseline questionnaire (Sarna et
al. 1978; Sarna & Kaprio 1980) and was in a few cases supplemented with photographs,
fingerprints and DNA-marker studies. The twin pairs were classified as monozygotic (MZ),
dizygotic (DZ), or unknown zygosity.
Subjects and methods
Figure 3. Flowchart depicting participation in the FinnTwin16 study
The frequency of breakfast skipping was assessed by the following question: "How often do
you eat breakfast (for example, sandwiches, milk, hot cereal, other similar food) before going
to school or going to work?" The three alternative responses were "every morning", "a few
times a week", "about once a week or less often".
Body shape and weight concerns
The young adult questionnaire included three subscales of the Eating Disorder Inventory-1
(Garner 1991): drive for thinness (DT), body dissatisfaction (BD), and bulimia. A Finnish
version of this instrument has been translated and validated (Charpentier et al., unpublished
manuscript). The EDI responses were scored 1 to 6 to ensure a more normal distribution. The
DT subscale has 7 items with a coefficient alpha in our sample of 0.87 for females and 0.75
Subjects and methods
for males. The BD subscale has 8 items with alphas of 0.92 in females and 0.86 in males. For
the correlational analysis and twin modelling of male data, DT and BD scores were
dichotomised into high vs. low groups, using 75% percentiles as cut-off points (DT, females:
24, males: 14; BD, females: 34, males: 19), because male BD and DT had extremely skewed
Intentional weight loss
IWL was assessed in our sample using the following question: “How many times during your
life have you intentionally lost over 5kg of weight?” The responders in the category ”never”
formed the no-IWL group. Individuals responding “once” formed the 1-IWL group and those
responding “2-4 times” or “5 times or more” formed the 2-IWL group. The last two
categories were combined because of a very small number of male respondents in either
category. For genetic analyses, IWL was dichotomised to “never” or “once or more” because
of the skewed distribution of responses.
In this study, self-reported weight and height at 16, 17, and 22-27 were used to calculate
respective BMIs. At 22-27, the twins also self-measured their waist circumference using a tape
measure that was mailed with the questionnaire, together with detailed instructions including a
body drawing indicating the site of measurement. The validity of self-reported weight, height,
and waist were examined in a subsample of 212 young adult twins. Trained personnel
measured the twins’ height, weight, and waist circumference using a stadiometer, calibrated
beam balance, and a tape measure. The agreement between these standardized measurements
and self-report was excellent (correlations of 0.96 for height, 0.94 for weight, 0.88 for waist;
Silventoinen et al. 2003); the median time interval between self-report and the standardized
measurement was 356 days. In addition to their current weight, the 22-27-year old twins also
reported their ideal and maximum weights at adult height.
Dietary restraint / disinhibition
At 22-27y, dietary restraint was assessed with the question: “Which of the following best
describes you?” “It’s easy for me to eat about the amount I need to.” (normal eating, reference
category); “I quite often eat more than I actually need.” (overeating); “I often try to restrict my
eating.”; (restrictive eating); “At times, I’m on a strict diet, at others I overeat.” (alternating
restrictive eating / overeating).
To assess eating styles of the twins, a short 12-item questionnaire was devised: 5 items
assessed snacking / grazing styles, 3 health-conscious eating, 2 emotional eating, 1 external
eating, and 1 night eating. Exploratory factor analysis was carried out, and the factors largely
corresponded to the a priori groupings of items. Individual items will be referred to
throughout this paper, but groupings are based on the factor structure. The factor pattern was
similar in both genders with the exception night eating, which emerged in women only, was
very rare and had a very distinct response pattern, possibly reflecting an underlying eating
disorder. Thus, night eating is not included in eating styles explored in this paper.
Subjects and methods
We measured education level at ages 16, 17, and 22-27. Educational attainment was self-
reported. For some analyses, the variable was dichotomised as mandatory school only vs.
higher education (vocational school, high school, polytechnic school, or university). The
education level in adolescence was taken from the 17 y questionnaire (when 93.0% of twins
had finished mandatory education) instead of the 16 y questionnaire (when only 57.5% had
finished mandatory education). Education level at 17 is more informative of academic success
in mid-adolescence than education level at 16, because further education is voluntary after 16,
and the choice of educational paths reflects academic performance.
Other health-related behaviors
The following other variables obtained from the questionnaires at the age of 16 were used in
our analyses: smoking status, alcohol use; use of coffee, tea, caffeinated soda, and cocoa; types
of milk and bread spread used; frequency of physical exercise, self-perceived physical
condition and health, and age of puberty onset (menarche for females, voice break for males).
From the questionnaires filled out at age 17, behavioral disinhibition, experience seeking,
and susceptibility to boredom scores as measured by the Sensation Seeking Scale (Zuckerman
1979) were used.
Family socioeconomic status (SES) was determined by the occupation of the father or the
mother, whichever ranked higher. If only one parent had responded, his or her occupation
determined the family’s SES. If both parents’ occupations were unknown, the family was
excluded from the analyses of the effect of SES on breakfast skipping. The parental
occupations were divided in seven categories using the Statistics Finland 1989 classification of
SES (Sosioekonomisen aseman luokitus 1989); we contrasted the two highest socioeconomic
categories (upper-level white-collar workers; independent entrepreneurs and farmers) with the
five lower categories (lower-level white-collar workers, blue-collar workers, students,
pensioners, and those of unclassified or unknown occupation) to create “higher SES” and
“lower SES” categories.
Other parental variables
From parental questionnaires, we obtained data on father’s and mother’s breakfast skipping
and other variables we deemed possibly important for breakfast skipping. The most central
variables were parental BMI, smoking, alcohol use, frequency of physical exercise, and highest
level of parental education. Smoking and alcohol use variables were dichotomized, primarily to
facilitate the assessment of possible interactions. Other information used in the analyses
included coffee, and tea use; types of milk and bread spread used; use of vitamin and trace
mineral supplements, and of natural and herbal drugs; unemployment in the family; working
in shifts, amount of sleep, and feeling tired in the morning.
Subjects and methods
General statistical methods
We investigated differences between outcome variables and correlates using cross-tabulations
and the Pearson chi-squared test of independence, corrected for clustered sampling within the
twin pair (Rao and Scott correction, svytab procedure in Stata 8.0) (Rao & Scott 1984). To
account for the clustered sampling structure, continuous variables in manuscript III were
analyzed using the svymeans procedure in Stata 8.0. Odds ratios obtained from univariate,
multiple, and polytomous logistic regression models were also used in these analyses (Hosmer
& Lemeshow 2000), again correcting for clustered sampling. Twin correlations were assessed
using the polychoric correlations function in SAS.
The basic principles of twin studies
Twin studies are a natural experiment that allow estimation of genetic and environmental
contributions to variation on traits of interest. The basic premise of twin studies is that MZ
twins are the product of a single fertilized oocyte, being thus genetically identical. DZ twins
are the result of two oocytes and share, on average, 50% of their segregating genes. In
quantitative genetics, genetic and environmental factors are assumed to contribute to the
phenotype of a given trait. Assuming that MZ and DZ twin environmental variances are
equivalent (ie, that both twin types have similar environments with respect to the phenotype
being studied), the relative contribution of genetic and environmental influences to trait
variation can then be resolved using variance components models. These models specify
genetic and environmental sources of phenotypic covariance in MZ and DZ twin pairs and
yield estimates of the degree to which trait variation is explained by the genetic and
The understanding of genetic and environmental components can be further refined:
genetic contribution can be additive (A), which refer to the additive effect of alleles at a locus
and summed over all loci, or non-additive (D, dominant), which refers to intra-locus allelic
interactions, summed over all loci. Conversely, the environmental contribution is usually
partitioned into shared (C, common) environmental effects and non-shared (E, unique)
environmental effects. Common twin or family environment denotes environmental factors
that make members of the same family alike; unique environmental influences make members
of the same family different from each other. Measurement error is usually estimated in
unique environmental effects. C and D effects cannot be simultaneously modeled from twins
reared together (Falconer & Mackay 1996): for this reason, classical twin studies estimate
variances components using ACE or ADE models and their submodels, but not ADCE
Several software packages have been developed to perform maximum likelihood
estimation or related function based optimizations of twin and twin-family data. Currently,
Mx, Mplus and Lisrel are the most widely used programs for these purposes.
In this study, various twin modelling strategies ranging from the simple and well known to
novel and experimental were used. Classical univariate twin models on same-sex twin pairs
were the centerpiece of study III. More sophisticated sex-limitation models that included
information on opposite-sex twin pairs were used to explore the genetic and environmental
influences on breakfast skipping in study II. Bivariate twin models were used both in study III
to check the genetic overlap in BD and DT, and in study IV to explore shared genetic and
environmental influences between BMI and IWL. Finally, twin-family models, ie, models that
Subjects and methods
include information from the mothers and fathers to the information obtained from the twins,
were used in study II to examine the results obtained from sex-limitation models. Specific
details of these modelling strategies are briefly discussed below.
Classic univariate twin models
Classic twin models assess whether a model that consists of additive genetic influences,
shared environmental influences, and unique environmental influences (ACE) fits the data
better than simpler models that only specify either additive genes and unique environment
(AE), shared and unique environment (CE), or unique environment alone (E). Alternatively,
ADE and its submodels can be assessed.
In study III, based on twin correlations (Pearson correlations for females, tetrachoric
correlations for males, calculated using SAS, version 8.0), ACE and its submodels were
deemed most probable explanatory models for body image related variables BD in both
genders and DT in men. Pairwise DT correlations in women showed possible dominance
effects: both ACE and ADE and submodels were fitted. Model fitting started with the most
parsimonious models, AE and CE, and then advanced to ACE/ADE models. However, in
this setting the power to detect D effects is relatively low (Falconer & Mackay 1996).
For the classic ACE twin models used in study III, we used the raw data maximum
likelihood estimation option in Mx (Neale et al. 2002) that enabled inclusion of unmatched
twin pairs (about 10% of our sample). When non-normality was apparent in the distributions,
dichotomization was used before model fitting, if simple attempts of transformation (such as
log transformation, square and cubic root transformation) did not render a close
approximation of the normal distribution. Such dichotomization was necessary for male EDI
subscales, where 75% percentile points were used as cut-off points. In females, we used DT
and BD scores as a continuous variable, from which variance-covariance matrices were
calculated and fitted using Mx. The distributions and variances for EDI subscale scores in
each gender were so different that sex-limitation modeling approaches would have proven
futile. Instead, we modeled each gender separately, implementing information obtained from
same-sex twins only.
fem al e
fem a le
1.0 (M Z) or 0.5 (DZ) 1.0
Common ef fects se x-limitat ion model
General sex-limitation model
Figure 4. Se
Subjects and methods
Univariate sex-limitation models
In study II, we tested two different models of sex effects on breakfast skipping, first
specifying a general sex-limitation model (see Figure 4, previous page) that estimates the
magnitude of a sex effect on A, C, and E separately for males and females. Second, we fitted a
more restrictive model, the common-effects model that allows the magnitudes of sex effects
to vary. The general sex-limitation model allows that sex differences may be caused by
different genes (or environmental factors) in males and females; the common-effects model
assumes that the same genes and environmental factors influence the breakfast skipping of
both boys and girls, but that the magnitude of the influences differs for sexes. Data from
same-sex and opposite-sex twin pairs were arranged in 3 x 3 contingency tables; sex-specific
threshold values were used. Variances were standardized by constraining the values in males
or females to 1. The sex-specific threshold values were 0.6, 1.2 for males and 0.5, 1.1 for
Bivariate twin models
Bivariate twin models estimate genetic and environmental contributions to the phenotypic
variance in two traits and the covariance between those two traits. However, the model does
not determine causality. There are several different ways to model bivariate relationships:
Cholesky decompositions were used in this study to assess the relationship of body image
related subscales BD and DT (study III) and the relationship of BMI and intentional weight
loss (study IV).
In study II, twin-family models of breakfast skipping were used to confirm and expand the
results obtained from sex-limitation twin models. Twin-family models were fitted to data from
polychoric correlation matrices and asymptotic covariance matrices based on data from all
four family members using weighted least squares estimation. Twin-family models try to
explain parent-child behavioral resemblance by specifying biometric expectations to explain
the variance-covariance relationship in the data. This model is then fit to the variance-
covariance matrix to test whether the model explains the data. First, we tested a model of social
homogamy (Eaves et al. 1989). A path diagram detailing a social homogamy model is presented
in Figure 5. Social homogamy refers to spouse selection that is based on environmental
similarities. The spousal resemblance is solely due to this common environment (C) (Reynolds
et al. 1996). This might for example be true for people marrying mainly because they belong to
the same religious community. According to this model, parents influence their offspring in
two ways: through genes and through parental common environment (which brought the
parents together in the first place). Parental unique environment (E) is assumed not to
transmit to the offspring.
Second, we tested a phenotypic assortment model. A path diagram detailing a phenotypic
assortment model is presented in Figure 6. Phenotypic assortment refers to the selection of
spouses based on observable characteristics (phenotypes) (Neale & Cardon 1992; Neale et al.
1994; Reynolds et al. 1996), that is, traits that can be influenced by either genes, family
environment or unique environmental experiences. In contrast to the simplistic social
homogamy model, spouse selection is not based on any particular component of the
phenotype (A, C,or E) alone. Cultural transmission to the offspring originates from the
parental phenotype (being thus affected by A, C, and E), and influences the common
Subjects and methods
environment of the offspring. The model assumes that interdependencies of the model
parameters are constant across generations. The phenotypic assortment model also implies
that because parents pass both genetic and environmental factors to their offspring, these
factors (A and C) are somewhat correlated in the offspring. In the classical twin model, this
correlation is assumed to be zero; in the phenotypic assortment model, the magnitude of this
correlation can be estimated.
Model fitting and indicators of fit
In these procedures, hierarchically nested models were fitted. Models are usually compared to
each other using the principle of parsimony. Models with fewer parameters are preferable if
they do not provide a substantially worse fit. To assess how well a model fits the data, several
fit statistics can be used. We used the –2 log-likelihood statistic: the likelihood-ratio of
alternative hierarchically nested models was calculated by the difference in their χ2 values. The
difference of two such χ2 goodness-of-fit statistics is itself distributed as a χ2 statistic with
degrees of freedom equal to the difference in degrees of freedom of the two models being
compared (Neale & Cardon 1992). If the difference in the χ2 values of two models was not
statistically significant, the principle of parsimony was applied: the model with fewer
parameters was preferred. Akaike’s information criterion (AIC, calculated as χ2- 2df) was used
as an additional indicator of fit: a model with a lower AIC value is the more parsimonious
(Neale & Cardon 1992).
Social homogamy model
Phenotypic assortment model
Weight control behaviours: correlates and demographics
The first aim of this study was to
characterize the phenotypes breakfast
skipping, body shape and weight
concerns, and intentional weight loss,
and to assess some of their demographic
determinants and behavioural correlates.
Below, each of these three domains is
As detailed in study I, breakfast skipping was relatively common: 15.7% of girls, 13.0% of
boys, 19.0% of mothers and 26.1% of fathers ate breakfast only once a week or less often (see
Figure 7). Parental breakfast skipping patterns were the statistically most significant factor
associated with adolescent breakfast skipping (Table 4). Low family socioeconomic status was
associated with breakfast skipping in adolescent boys, but not in girls (Table 4).
Table 4. Correlates of breakfast skipping in adolescents: the influence of parental
factors. In sex-adjusted polytomous logistic regression models, adolescents who skip
breakfast were contrasted with adolescents who always eat breakfast
Breakfast skippers (vs. eaters)
Sex-adjusted odds ratios
(95% Confidence Interval1)
Breakfast a few
times a week
Breakfast once a
week or less often
every morning (71.0%) 1.0 (reference) 1.0 (reference)
a few times a week, 10 (10.0%) 1.74 (1.30-2.34) 2.24 (1.65-3.05)
once a week or less (19.0%) 2.26 (1.82-2.80) 3.57 (2.84-4.49)
every morning (59.8%) 1.0 (reference) 1.0 (reference)
a few times a week (14.1%) 1.95 (1.50-2.54) 1.79 (1.30-2.46)
once a week or less (26.1%) 2.06 (1.66-2.54) 2.93 (2.32-3.70)
upper-level employee (22.6%) 1.0 (reference) 1.0 (reference)
self-employed, including self-employed farmers (18.8%) 1.03 (0.78-1.36) 1.43 (1.05-1.95)
lower-level employee (21.7%) 0.91 (0.70-1.18) 1.43 (1.06-1.93)
manual worker, student, retired, or of unclassified
1.07 (0.85-1.36) 1.55 (1.18-2.03)
68.4 73.5 71.0 59.8
16.0 13.4 10.0
15.7 13.0 19.0 26.1
breakfast once a
week or less often
breakfast a few
times a week
Figure 7. Frequency of breakfast skipping in
girls, boys, women, and men of FinnTwin16
In study I, also a second group of correlates, health-compromising behaviors, ie, smoking,
infrequent exercise, frequent alcohol use, behavioral disinhibition, and high body mass index
(BMI), also had clear cross-sectional associations with breakfast skipping in adolescents (Table
5). Female gender (OR 1.3, 95% CI: 1.1-1.5) and low education level at 17 were also
significantly associated with breakfast skipping in adolescents.
Table 5. Correlates of breakfast skipping: individual-level factors in adolescents.
Adolescents who skip breakfast vs. always eat breakfast: odds ratios of correlated
individual-level behaviors and other characteristics, based on sex-adjusted
polytomous logistic regression models
Breakfast skippers (vs. eaters)
Sex-adjusted odds ratios
(95% Confidence Interval1)
Breakfast a few
times a week
Breakfast once a
week or less often
BMI, kg/m2, continuous (N=5403)2 1.06 (1.02-1.09) 1.06 (1.02-1.09)
never (48.7%) 1.0 (reference) 1.0 (reference)
past smoker (20.4%) 1.43 (1.15-1.77) 1.56 (1.23-1.96)
less than once a week (9.4%) 1.52 (1.16-2.00) 1.28 (0.94-1.76)
once a week or more but less than daily
1.60 (1.07-2.37) 1.47 (0.97-2.24)
daily (17.4%) 2.28 (1.83-2.85) 4.17 (3.34-5.21)
behavioral disinhibition at 17 y, continuous (N=5334) 3 1.07 (1.02-1.12) 1.15 (1.10-1.20)
age of puberty onset (years), continuous (N=5279) 4 1.01 (0.94-1.08) 0.88 (0.82-0.95)
never (23.5%) 1.0 (reference) 1.0 (reference)
once a year or less (9.3%) 0.60 (0.40-0.90) 1.33 (0.94-1.87)
a few times a year (14.4%) 1.09 (0.82-1.46) 1.55 (1.15-2.09)
once in a few months (13.7%) 1.22 (0.92-1.62) 1.44 (1.05-1.96)
once a month (11.7%) 1.31 (0.98-1.75) 1.47 (1.07-2.02)
a few times a month (17.8%) 1.67 (1.29-2.15) 2.13 (1.61-2.81)
weekly or more (9.7%) 2.03 (1.51-2.73) 2.89 (2.11-3.96)
senior high or polytechnic school (36.1%) 1.0 (reference) 1.0 (reference)
vocational school (19.2%) 1.41 (1.13-1.77) 2.00 (1.56-2.57)
junior high school (43.0%) 1.16 (0.96-1.41) 1.79 (1.44-2.22)
education at 16 y
not in school (1.7%) 2.92 (1.62-5.25) 4.91 (2.79-8.64)
senior high school, vocational college or
polytechnic school (62.1%)
1.0 (reference) 1.0 (reference)
vocational school (30.6%) 1.52 (1.26-1.83) 1.97 (1.62-2.41)
junior high school (3.4%) 1.60 (0.98-2.61) 3.49 (2.25-5.43)
education at 17 y
not in school (4.0%) 2.23 (1.49-3.34) 3.43 (2.36-4.97)
daily (15.8%) 1.0 (reference) 1.0 (reference)
4-5 times a week (15.2%) 1.36 (1.01-1.84) 0.95 (0.67-1.35)
2-3 times a week (28.0%) 1.28 (.98-1.69) 1.32 (0.99-1.77)
about once a week (19.1%) 1.60 (1.20-2.12) 1.36 (1.00-1.85)
once or twice a month (10.4%) 1.60 (1.15-2.22) 2.33 (1.66-3.28)
less than once a month (6.6%) 1.72 (1.17-2.52) 4.03 (2.84-5.72)
never (5.0%) 1.39 (0.90-2.15) 3.75 (2.57-5.47)
1 Confidence intervals adjusted for intraclass correlation
2 For 118 BMI values missing at 16 y, BMI values at 17 y were substituted
3 For 180 disinhibition scores missing at 17 y, disinhibition scores at 18 y were substituted
4 Menarche for females, voice break for males
In adults examined in study I, the findings were similar for health-compromising factors,
although slightly different measures were used: smoking, infrequent exercise, higher BMI, and
more frequent alcohol use were associated with breakfast skipping (Table 6). In the
sociodemographic domain, adult males, not females, were more likely to skip breakfast. The
tendency to have breakfast regularly increased with age, but low education was associated with
Body shape and weight concerns
The general characteristics of FinnTwin 16 twins at the 4th wave of assessment (study III) are
detailed in Table 7. In average, the level of body shape and weight dissatisfaction, as measured
by BD and DT, was higher in women than in men. The association of BD and education level
in adult women reflected a greater prevalence of overweight (12.7% vs. 7.1%, p<0.00001 in
women; 16.9% vs. 11.8%, p=0.0018 in men) among individuals with a low level of education.
As detailed in Table 8, larger body size and shape (higher BMI, larger waist circumference,
both currently and at the age 16), disordered eating patterns (such as restrictive eating,
overeating, and alternating restricting-binging), and psychosomatic symptoms (such as poor
self-perceived health, nervousness, sleeping difficulties), depressive mood, dissatisfaction with
parents, and feelings of loneliness were associated with BD and DT in both genders in study
III. Also, sexual maturation related items were associated with BD, but not with DT: early
pubertal onset, early initiation of sexual activity, and multiple sex partners increased BD and
late puberty emerged as a protective factor from BD.
Table 6. Correlates of breakfast eating in adults: polytomous logistic regression
models adjusted for age and sex.
Breakfast skippers vs. eaters
Adjusted odds ratios
(95% Confidence Interval1)
Breakfast a few
times a week
Breakfast once a
week or less
upper-level employee (19.3%) 1.0 (reference) 1.0 (reference)
self-employed, including farmers (16.1%) 1.18 (0.84-1.67) 1.69 (1.29-2.21)
lower-level employee (34.0%) 1.56 (1.16-2.09) 1.75 (1.37-2.22)
manual worker, student, retired (30.6%) 2.16 (1.63-2.96) 2.66 (2.11-3.35)
never (40.8%) 1.0 (reference) 1.0 (reference)
past smoker (33.7%) 1.11 (0.89-1.39) 1.35 (1.11-1.62)
current smoker (25.6%) 2.06 (1.63-2.61) 3.53 (2.91-4.28)
less than twice a month (41.0%) 1.0 (reference) 1.0 (reference)
3-8 times a month (41.5%) 1.19 (0.97-1.47) 1.12 (0.94-1.33)
N=4625 over 8 times a month (17.5%) 0.84 (0.63-1.12) 1.19 (0.96-1.48)
university degree (4.5%) 1.0 (reference) 1.0 (reference)
senior high school graduation (17.0%) 1.03 (0.59-1.78) 1.29 (0.80-2.07)
vocational school (21.2%) 1.63 (0.96-2.75) 1.99 (1.27-3.12)
mandatory education only (57.3%) 1.99 (1.21-3.29) 2.54 (1.64-3.92)
BMI, kg/m2, continuous (N=4608) 1.03 (1.01-1.06) 1.03 (1.01-1.05)
exercise ≥6 times a month (41.4%) 1.0 (reference) 1.0 (reference)
exercise 3-5 times a month (30.1%) 1.14 (0.90-1.45) 1.16 (0.97-1.38)
exercise 1-2 times a month (24.6%) 1.18 (0.93-1.49) 1.69 (1.40-2.03)
exercise less than once a month (3.9%) 0.89 (0.50-1.56) 2.83 (2.00-4.00)
1 Confidence intervals adjusted for clustered sampling.
Intentional weight loss
In study IV, IWL was much more common among women than men (Table 9). Obesity
(BMI>30) was equally common in women (3.7%) and men (4.1%) (F=0.39, p=0.53), but IWL
was significantly more common among obese women (81.9%) than obese men (69.0%)
(F=6.1, p=0.015). Associations of IWL and SES or education were not specifically explored in
mean BMIs of individuals with mandatory education only (23.4 kg/m2) or vocation education
this study, but we did find a relationship of BMI and SES that was gender-independent: the
(23.5 kg/m2) were significantly (F=13.7, p<0.00001) higher than those who had completed
high school or polytechnic (22.7 kg/m2), or university (22.1 kg/m2). Women and men who
had lost at least 5kg of weight at least once had significantly higher past, current, maximum,
and ideal BMIs and larger waist circumferences than individuals who did not engage in IWL
(Table 9). Both women and men who had lost at least 5 kg at least twice had significantly
Table 7. General descriptive statistics of FinnTwin16 twins
MZ (N) 35.0 (868) 26.3 (540)
Same-sex DZ 30.8 (765) 34.4 (705)
Opposite-sex DZ 31.9 (793) 34.5 (717)
Unknown 2.3 (57) 4.3 (89)
Ongoing education at 17
Senior high or polytechnic 71.5 53.7
Vocational school 22.0 38.4
Junior high school 2.5 3.7
Not in school 4.0 3.9
Educational attainment at 22-27
University degree 7.9 4.7
Polytechnic degree 14.8 8.4
Senior high school graduation 41.6 38.3
Vocational college graduation 9.2 7.2
Vocational school or training degree 21.6 35.4
Junior high school graduation only 4.8 6.0
Mean (SD) 22.2 (3.5) 23.9 (3.1)
Median BD score2 26 14
Median DT score3 18 11
Intentional weight loss
Never lost >5kg weight 58.2 75.7
Lost >5kg at once 24.2 14.0
Lost >5kg at least twice 17.6 10.2
1 Percentages calculated from the proportion of twins still participating at age 22-27 y
2 The scores range from 8 to 48 when the instrument is scored from 1 to 6. a higher score denoting
greater body dissatisfaction.
3 The scores range from 7 to 42 when the instrument is scored from 1 to 6, a higher score denoting
greater drive for thinness
Table 8. Correlates of high body dissatisfaction and high drive for thinness: odds
ratios (with 95% CIs) from univariate female and male logistic regression models. Eg,
compared to women with low body dissatisfaction, the risk of overeating is 7.9 times
higher in women reporting high body dissatisfaction.
High body dissatisfaction High drive for thinness
women men women men
Correlate OR 95% CI OR 95% CI OR 95% CI OR 95% CI
Body size and shape
Current BMI (OR per kg/m2) 1.4 (1.3-1.4) 1.3 (1.2-1.3) 1.2 (1.1-1.2) 1.4 (1.3-1.4)
BMI at 16 (OR per kg/m2) 1.3 (1.3-1.4) 1.3 (1.2-1.3) 1.2 (1.2-1.3) 1.4 (1.3-1.5)
Current waist (OR per cm) 1.1 (1.1-1.1) 1.1 (1.1-1.1) 1.0 (1.0-1.1) 1.1 (1.1-1.1)
Overweight at 16 3.9 (3.0-5.2) 4.3 (3.1-5.8) 2.7 (2.1-3.5) 6.0 (4.3-8.2)
Overeating 7.9 (6.1-10.1) 5.5 (4.3-7.1) 5.7 (4.4-7.4) 5.8 (4.5-7.4)
Restrictive eating 8.1 (6.0-10.9) 16.6 (9.2-
Alternating overeating and
Intentional weight loss of ≥5kg at
4.0 (3.2-4.9) 4.5 (3.6-5.6) 4.0 (3.3-4.9) 6.3 (5.0-7.9)
Self-reported ideal BMI (OR per
1.3 (1.3-1.4) 1.2 (1.2-1.3) 1.1 (1.1-1.2) 1.2 (1.2-1.3)
Poor self-perceived physical
3.1 (2.5-3.8) 3.0 (2.4-3.7) 1.7 (1.4-2.1) 1.9 (1.5-2.3)
Puberty and sexuality
Early puberty 1.6 (1.2-2.2) 1.4 (1.0-1.9) 1.3 (1.0-1.7) 1.5 (1.1-2.1)
Late puberty 0.5 (0.4-0.6) 0.8 (0.6-1.1) 0.8 (0.6-1.0) 0.8 (0.6-1.0)
Number of sex partners (OR per
1.1 (1.1-1.2) 1.0 (0.9-1.1) 1.1 (1.0-1.2) 1.0 (1.0-1.1)
Age at first sexual intercourse 0.9 (0.9-0.9) 1.0 (0.9-1.0) 1.0 (0.9-1.0) 1.0 (1.0-1.1)
Current unhappiness (vs. current
2.0 (1.5-2.6) 2.6 (2.0-3.4) 2.6 (2.0-3.4) 1.9 (1.4-2.5)
Current self-perceived poor health 2.1 (1.7-2.7) 2.5 (1.9-3.3) 1.8 (1.5-2.3) 1.7 (1.3-2.3)
Self-perceived poor health at 16 2.3 (1.8-2.9) 1.8 (1.3-2.4) 1.9 (1.5-2.4) 1.4 (1.0-1.9)
Current frequent stomach pain (vs.
monthly or less)
1.8 (1.4-2.3) 1.7 (1.1-2.6) 1.9 (1.5-2.4) 1.5 (1.0-2.2)
Frequent stomach pain at 16 (vs.
monthly or less)
1.4 (1.0-2.0) 0.7 (0.4-1.3) 1.8 (1.3-2.5) 0.9 (0.5-1.6)
Current headaches (vs. monthly or
1.8 (1.5-2.2) 1.6 (1.2-2.2) 1.6 (1.3-1.9) 1.3 (0.9-1.8)
Frequent headaches at 16 (vs.
monthly or less)
1.5 (1.2-1.8) 1.4 (1.0-1.8) 1.7 (1.3-2.1) 1.4 (1.0-2.0)
Current nervousness 1.7 (1.4-2.1) 1.8 (1.4-2.2) 2.4 (1.9-3.0) 1.8 (1.4-2.2)
Nervousness at 16 1.3 (1.0-1.6) 1.3 (1.0-1.7) 1.8 (1.4-2.2) 1.3 (1.0-1.7)
Current depressive mood 1.6 (1.3-2.0) 2.0 (1.6-2.6) 2.0 (1.6-2.4) 1.8 (1.4-2.3)
Current sleeping difficulties 2.0 (1.6-2.4) 2.0 (1.6-2.5) 2.1 (1.7-2.5) 1.5 (1.2-1.9)
Sleeping difficulties at 16 1.5 (1.2-1.9) 1.6 (1.2-2.1) 1.3 (1.0-1.6) 1.2 (0.9-1.6)
Current dissatisfaction with partner 1.2 (1.0-1.5) 1.4 (1.1-1.7) 1.4 (1.1-1.7) 1.2 (1.0-1.5)
Current dissatisfaction with
1.4 (1.0-1.8) 2.0 (1.8-2.7) 1.4 (1.1-1.9) 2.0 (1.5-2.8)
Current dissatisfaction with father 1.5 (1.3-2.0) 1.8 (1.4-2.3) 1.4 (1.2-1.8) 1.5 (1.2-2.0)
Current frequent feelings of
1.4 (1.2-1.7) 1.7 (1.6-2.4) 1.6 (1.3-2.0) 1.4 (1.1-1.7)
1 Confidence intervals adjusted for clustered sampling.
Table 9. Means (95% confidence intervals, adjusted for clustered sampling) of weight-
related variables in individuals who have never engaged in intentional weight loss
(no-IWL), have lost ≥5kg of weight intentionally once (1-IWL), or at least twice (2-IWL).
no-IWL 1-IWL 2-IWL no-IWL 1-IWL 2-IWL
N 1479 615 448 1605 297 218
% 58.2 24.2 17.6 75.7 14.0 10.3
Age (years) 24.3
BMI at 16-17y
Waist (cm) 72.8
higher current, maximum, and ideal BMIs than their counterparts who had engaged in IWL
only once (Table 9). As detailed in Table 10, IWL was significantly associated with restricting,
overeating, and alternating restricting/overeating (IV). Snacking and eating in the evening
were characteristic of individuals with at least 2 IWL attempts. As expected, avoiding fatty
foods and calories was significantly more pronounced in individuals who had engaged in IWL
than in the no-IWL group.
Table 10. Gender-specific odds ratios (95% confidence intervals, adjusted for
clustered sampling) of eating patterns in individuals who have never lost weight
intentionally (no-IWL, reference group) vs. individuals who have lost weight
intentionally once (1-IWL) or at least twice (2-IWL).
Restrictive eating / overeating
Restrictive eating 1.7; 1.4-2.2 3.6; 2.7-4.8 2.5; 1.9-3.4 3.3; 2.3-4.6
Frequent overeating 3.2; 2.3-4.3 8.3; 6.0-11.6 9.5; 5.2-17.5 10.4; 5.4-20.2
Alternating overeating and restricting 6.7; 3.9-11.6 25.4; 14.8-43.6 7.3; 2.1-25.6 34.6; 12.4-96.6
Frequent snacking between meals 1.0; 0.8-1.3 1.7; 1.3-2.1 1.2; 0.9-1.6 1.5; 1.1-2.1
Frequent snacks replace meals 1.4; 1.0-1.8 2.0; 1.5-2.7 1.5; 1.0-2.4 2.5; 1.6-3.8
Highest food consumption in the
1.2; 1.0-1.5 1.9; 1.6-2.4 1.2; 0.9-1.5 1.3; 1.0-1.8
Grazing throughout the evening 0.9; 0.7-1.2 1.4; 1.1-1.8 0.7; 0.5-1.1 1.6; 1.1-2.4
Eating while watching TV 1.0; 0.8-1.2 1.3; 1.0-1.6 1.2; 0.9-1.6 1.2; 0.9-1.6
Maintaining healthy eating patterns 1.4; 1.1-1.8 0.9; 0.7-1.2 1.2; 0.9-1.5 1.0; 0.7-1.3
Avoiding fatty foods 1.6; 1.3-2.0 1.6; 1.3-2.0 1.7; 1.3-2.2 2.0; 1.5-2.7
Avoiding calories 1.9; 1.6-2.3 2.0; 1.7-2.6 2.4; 1.7-3.3 2.9; 2.0-4.1
Psychological aspects of eating
Visual cues (seeing food or food ads)
0.9; 0.5-1.6 2.3; 1.5-3.7 1.0; 0.4-2.6 1.9; 0.8-4.5
Food used as a reward 1.3; 1.0-1.7 2.1; 1.6-2.8 1.4; 1.0-2.0 1.7; 1.1-2.5
Comfort eating 1.9; 1.3-2.7 3.5; 2.5-5.1 1.2; 0.6-2.4 2.1; 1.2-3.9
Genetic and environmental influences on weight control behaviors
These studies (II-IV) also aimed to estimate genetic and environmental liability in breakfast
skipping, body shape and weight concerns, and intentional weight loss, to investigate familial
patterns of breakfast skipping (II), and to assess overlap of genetic influences on BMI and
IWL (IV). Below, each of these three domains is assessed separately.
Estimates of additive genetic, common environmental and unique environmental influences
were obtained for breakfast skipping using twin and twin-family models in study II. In the sex-
limitation twin models based on information based on twin data only, the sex-specific
environmental and genetic components could be removed (Table 11). In the resulting best-
fitting common-effects model, variation in breakfast skipping frequency was explained by
genetic and environmental factors in boys and girls. Additive genetic effects explained 41%
(95% CI: 21-63%) of the variance in breakfast skipping in girls and 66% (95% CI: 47-79%) in
boys, and common environmental effects 45% (95% CI: 23-62%) in girls and 14% (95% CI:
5-29%) in boys. The gender differences apparent in the common-effects sex-limitation model
were significant: constraining male and female ACE estimates to be equal caused a significant
deterioration in model fit (∆χ2=7.25, ∆df=2, p=0.03), implying that common environmental
effects in breakfast skipping are more important for females than males. In summary, additive
genetic effects account for a larger proportion of variability in breakfast skipping in males than
in females, but both sources of variation are needed to account for the data.
Table 11. Sex-limitation models of breakfast skipping patterns
1 Either male-specific additive genetic (AM) or male-specific common environmental component (CM)
2 Akaike Information Criterion
3 Best-fitting model by the Akaike Information Criterion
components of variance estimates (95% CI) goodness-of-fit tests
model additive genetic
df P AIC2
A C E M
♀ ♂ ♀ ♂ ♀ ♂
- 35.03 32 0.326 -28.973
33.94 31 0.328 -28.06
33.94 31 0.328 -28.06
Family models were also used in study II to check the validity of modeling results
explained above, and to explore assortative mating effects. The overall breakfast eating
correlation between parents of the twins was fairly strong, 0.39 (95% CI: 0.34-0.45). The
overall mother-daughter breakfast eating correlation was 0.30 (95% CI: 0.25-0.36), mother-son
correlation 0.35 (95% CI: 0.29-0.41), father-daughter correlation 0.27 (95% CI: 0.21-0.33), and
father-son correlation 0.29 (95% CI: 0.23-0.35). Although mothers appear slightly more like
their offspring than fathers do, the difference was not statistically significant.
Of twin-family models, we first fit a social homogamy model to the parent and offspring
data from the five zygosity groups. Taking the results of the sex-limitation models (Figure 4, p.
31) as a starting point, our baseline model assumed common effects sex limitation in the ACE
estimates, but allowed for differing cultural transmission for fathers-sons, fathers-daughters,
mothers-sons and mothers-daughters. This model fit the data very well, χ2=22.33, df=21,
p=0.38. Relative to the sex-limitation model based on the twins only (Table 11), estimates of
additive genetic effects were lower for both females (0.31) and males (0.47), and estimates of
common environmental effects were higher for both females (0.54) and males (0.27). Spousal
assortment based on the shared environment was estimated at the upper boundary of 1.0
(95% CI= 0.81-1.0). The individual cultural transmission parameters were small and
nonsignificant, and the four parameters could be constrained to an equal estimate (0.16;
95%CI=0.08-0.24) without a significant decrease in fit (∆χ2=0.24, ∆df=3). As in the twin-
based sex limitation models, addition of sex-specific genetic (∆χ2=2.65, ∆df=1) or shared
environmental (∆χ2=2.66, ∆df=1) parameters did not significantly improve fit, but
constraining genetic and environmental estimates to be equal for males and females did result
in a significant decrease of fit (∆χ2=6.24, ∆df=2): thus on the family model level as well, the
common-effects model provided the best fit.
Second, we fit a phenotypic assortment model to the data to explore whether the
assumptions of phenotypic assortment (as described on p. 35-6) provide our data a better fit
than the assumption that spouse selection is based on environmental influences shared by the
spouses. Our baseline model again assumed common effects sex limitation in the ACE
estimates (see Figure 4, p. 34), but allowed for differing cultural transmission for fathers-sons,
fathers-daughters, mothers-sons and mothers-daughters. This model fit the data very well
(χ2=15.64, df=21, p=0.79). Addition of sex-specific genetic (∆χ2=3.35, df=1) or shared
environmental (∆χ2=3.35, df=1) parameters did not significantly improve fit. As in the social
homogamy model, the four cultural transmission parameters could be set equal without a
significant decrease in fit (∆χ2=5.16, df=3, p=0.16). This single cultural transmission variable
(-0.16; 95% CI=-0.37-0.01) did not quite reach significance. The assortative mating parameter
was estimated at 0.40 (95% CI=0.34-0.47). Additive genetic variance components were 0.72
(95% CI=0.46-0.98) for women and 0.63 (95%CI=0.38-0.89) for men. Common environment
estimates were 0.29 (95%CI=0.20-0.41) for women and 0.25 (95% CI=0.16-0.36) for men.
The gene-environment correlation induced by cultural transmission was -0.19 (95% CI=-0.47-
Body shape and weight concerns
Study III assessed genetic and environmental contributions to body shape and weight
concerns. As detailed in Figure 8 that presents the results of sex-specific univariate twin
models of BD and DT, AE models provided the best fit for female body shape and weight
concerns: additive genes accounted for 59.4% (95% CI: 53.2-64.7%) of variance in BD and
51.0% (95% CI: 43.7-57.5%) of DT in females. In men, the situation was entirely different:
CE models provided the best fit. Thus, according to our models, DT and BD were purely
Figure 8. Comparisons of gender-specific univariate twin models of
body dissatisfaction and drive for thinness
† The p value associated with ∆χ2 (change in model fit compared to the saturated model).
* Best fitting model.
1 For females, the C estimate in ACE models is very small, and good model fit is obtained by
omitting the C estimate altogether: AE models give female data the best fit. For males, the A
estimate in ACE models is very small, and omission of A improves model fit: CE models give
male data the best fit.
2 For females, ADE model fits the data better than ACE according to the –2 Log Likelihood
statistic, but AE is the most parsimonious model. For males, the A estimate in the ACE model is
very small, and it can be removed altogether without a significant decrease in model fit. CE is
the most parsimonious male model.
goodness-of-fit tests heritability common
Body dissatisfaction1 -2LL df ∆χ2 p
2 c2 e
ACE (saturated) 11749 1606 - - 59.4 (42.8-64.7) 0.0 (0.0-14.3) 40.7 (35.3-46.8)
CE 11786 1607 37.1 <0.001
- 44.2 (38.4-49.7) 55.8 (50.3-61.6)
AE* 11749 1607 0 - 59.4 (53.2-64.7) - 40.7 (35.3-46.8)
ACE (saturated) 2133 1222 - - 7.3 (0.0-15.1) 80.0 (73.1-85.7) 12.7 (10.6-15.5)
CE* 2137 1223
3.6 0.057 - 85.3 (83.2-87.0) 14.7 (13.0-16.8)
AE 2243 1223
109.5 <0.001 88.0 (85.6-89.9) - 12.0 (10.1-14.4)
Drive for thinness2 -2LL df ∆χ2 p
2 c2 e
ACE (saturated) 10840 1589 - - 50.1 (37.3-57.5) 0.0 (0.0-10.9) 49.0 (42.5-56.3)
CE 10868 1590
- 36.6 (30.1-42.7) 63.4 (57.3-69.8)
AE* 10840 1590 0 -
51.0 (43.7-57.5) - 49.0 (42.5-56.3)
ADE (saturated) 10838 1589 - -
27.0 (0.0-56.1) 25.2 (0.0-57.6) 47.8 (41.3-55.1)
DE 10840 1590
1.7 0.66 - 52.9 (45.8-59.1) 47.1 (40.9-54.2)
AE* 10840 1590 1.5 0.63
51.0 (43.7-57.5) - 49.0 (42.5-56.3)
ACE (saturated) 2084 1207 - - 1.2 (0.0-9.1) 84.9 (78.9-87.9) 13.0 (10.9-15.5)
CE* 2084 1208 0.3 0.58 - 86.4 (84.4-88.0) 13.6 (12.0-15.6)
AE 2218 1208 134.4 <0.001
87.7 (85.2-89.7) - 12.3 (10.3-14.8)
environmentally modulated in men only.
When the genetic and environmental overlap of DT and BD was assessed using bivariate
Cholesky decompositions in study III, the genetic effects of DT correlated with the genetic
effects of BD (rg = 0.80, 95% CI: 0.74-0.85) in women, and unique environmental effects of
BD and DT had a correlation of 0.59 (95% CI: 0.53-0.65). DT and BD in males were
influenced solely by environmental factors in our models, and genetic and environmental
correlations of BD and DT in males were not estimated.
Intentional weight loss
In bivariate twin models (Figure 9) tested in study IV, IWL was estimated to have a heritability
of 38% (95% CI: 19-55%) in men and 66% (95% CI: 55-75%) in women. Heritability
estimates of BMI were similar in both genders. The genetic covariance of BMI and IWL was
0.38 (95% CI: 0.28-0.47) for men and 0.45 (95% CI: 0.41-0.52) for women, implying that
although there is some overlap in genetic factors that influence BMI and IWL, there are also
substantial unique genetic influences on each trait. In men, environmental effects influencing
BMI and IWL were shared, albeit to a very modest degree (rg =0.11, 95% CI: 0.04-0.19). In
women, the overlap in common environmental effects influencing BMI and IWL was very
small (rc = 0.02, 95% CI: -0.02-0.07).
Figure 9. Variance estimates from gender-specific models of body mass index (BMI)
and intentional weight loss (IWL)
Correlates of weight control behaviors
In this study, weight control behaviors were associated with the study participants’ general
psychological and physical health. Breakfast skipping was significantly associated with health-
compromising behaviors, such as smoking, infrequent exercise, frequent alcohol use, and high
BMI in both adults and adolescents. Thus this study confirms similar earlier findings from
adolescents (Isralowitz & Trostler 1996; Höglund et al. 1998; Cavadini et al. 2000; Sjöberg et
al. 2003) and expands them to adult populations, which have not been studied this extensively
previously. In our study, the tendency to eat breakfast regularly increased with age in adults.
Body shape and weight concerns were associated with larger body size and multiple
psychosomatic symptoms and with dissatisfaction with close relationships, such as
relationship with one’s parents, and feelings of loneliness. To my knowledge, these
associations have not been reported previously in a population sample of young adults. Body
shape and weight concerns were also clearly associated with disordered eating patterns,
particularly restrictive eating, overeating, and alternating restricting-binging, confirming many
similar previous observations (Grigg et al. 1996; Ackard & Peterson 2001).
Breakfast skipping and body shape and weight concerns are examples of health-related
behaviors that become widespread peripubertally (Sjöberg et al. 2003). Late pubertal onset
emerged as a protective factor from body dissatisfaction in women: this finding has recently
been confirmed by another population study (Slof et al. 2003). Unfortunately, we do not know
at which age body and weight dissatisfaction emerged in our sample, although generally they
manifest relatively early, before puberty (Schur et al. 2000; Borresen & Rosenvinge 2003).
Because our study design was incompletely longitudinal (body image, intentional weight loss,
and nutritional variables were only assessed at the fourth wave, although information on BMI
and breakfast skipping were available at baseline and the fourth wave), longitudinal evaluations
of all variables of interest would be important in establishing the time sequence of these
events and determining the direction of causality.
Intentional weight loss attempts were also associated with larger body size and disordered
eating styles. Individuals who engaged in intentional weight loss attempted to restrict food
intake and avoid fatty and calorie-rich foods, but also reported overeating, snacking, and
eating in the evening. These conflicting eating styles often manifest as weight cycling;
unfortunately, the amounts of weight lost and regained in each cycle were not assessed in our
sample even when the number of intentional weight loss attempts was recorded.
Women in particular reported eating in response to visual and emotional cues. This trend
is worrying, but fits well with earlier reports of factors that make weight loss maintenance
difficult, particularly stress and anxiety, responsiveness to dietary lapses, and difficulties in
implementing behavioural weight control strategies (Sarlio-Lähteenkorva & Rissanen 1998;
McGuire et al. 1999a; Sarlio-Lähteenkorva et al. 2000; Dohm et al. 2001).
Our demographic analyses showed female gender was associated with breakfast skipping
in adolescents, but not adults, and also with a higher risk of body and weight dissatisfaction
and intentional weight loss attempts, confirming that gender differences in food intake and
selection usually appear in adolescence (Rolls et al. 1991). In average, men consume more
energy than women do, and there are clear gender differences in eating styles, which may in
addition to clear physiological factors indicate that women have been socialized to eat in a
more feminine manner and experience more food-related conflict than men do (Rolls et al.
The parents of low SES families were clearly more likely to skip breakfast than parents of
high SES families; however, low family SES was only reflected by boys’ breakfast skipping and
had no influence on girls. In young adults, no association was found between DT and
education level. BD and IWL were more common among those with low education, but this
relationship was largely mediated by BMI, which has a clear inverse association with education
in the Finnish adult population (Lahti-Koski et al. 2000). Perhaps body shape and weight
concerns are pervasive in the society, but individuals with a low education level have a greater
risk of unhealthy lifestyles and also a greater tolerance of consequent obesity and weight gain.
Genetic and environmental influences on weight control behaviours
Substantial genetic effects were evident on most weight control behaviors assessed in this
study. However, surprisingly, less than half of the genetic influences affecting weight loss
attempts were shared with those affecting BMI. Thus, there seem to be considerable genetic
influences on eating related phenotypes that are distinct from body size. In women, genetic
contributions to weight loss and body shape and weight concerns were relatively large. In
young men, body shape and weight concerns were entirely environmentally modulated, and
weight loss was substantially less heritable in men than women. In adolescence, examining the
phenotype breakfast eating, genetic influences were relatively strong on both girls and boys,
particularly when family modeling was implemented. As the direction of the gender difference
changed when different modeling strategies were implemented, the clear direction of the
gender differences remains unclear.
To our knowledge, the analyses of genetic covariance between weight loss and BMI have
not been accomplished previously, although several earlier reports have found small to modest
genetic influences on weight change and weight gain in twins and families (Korkeila et al.
1995; Heitmann et al. 1997; Korkeila et al. 1999; Coady et al. 2002; Golla et al. 2003). On the
other hand, our study replicates and confirms the heritability estimates of body shape and
weight concerns in young adult women, as measured EDI subscales BD and DT and reported
by previous authors (Rutherford et al. 1993; Klump et al. 2000), demonstrating that when the
measure is clearly defined and well-validated, the gender, age group, and cultural context is
broadly similar, heritability estimates can be very consistent indeed. But given our larger
samples, we could estimate more precise 95% CIs. Also, given that the populations studied
might have varying frequencies of genes that predispose to BD and DT, the consistency of
these findings is remarkable.
Perhaps more remarkably, we were able to find substantial shared family environment
influences on body shape and weight concerns in men, and smaller but still significant family
environmental influences on breakfast skipping in adolescents. The relative paucity of shared
environmental influences on BMI-related phenotypes has amazed researchers, because many
of the lifestyle changes that have rendered our environment more obesogenic operate on the
level of families (eg, cars, televisions, and physical effort saving household appliances are
usually shared by the members of a family), leaving them to conclude: “Evidently, what the
family has on table must be less important than what individuals take up from the table or
leave behind” (Hewitt 1997). Another explanation may be that twin studies often have limited
power to detect shared environmental effects, and that adhering to the principle of parsimony
often leads to the loss of small shared environmental influences (Hopper 2000; Sullivan &
Eaves 2002). Also, shared environmental influences are probably most relevant and
detectable in childhood, and begin to decrease during puberty when adolescents become more
independent, particularly in Finnish society. For Finnish twin pairs and matched classmate
controls aged 12 years, significant familial and local community effects were observed for a
range of phenotypes (Rose et al. 2003).
In our study, breakfast skipping patterns in 16-year-old twins showed strong genetic and
shared family environmental influences; individual-specific environment was much less
important. Perhaps a part of the explanation lies in the nature of breakfast: it is a meal that
even relatively young children can usually prepare unsupervised, although maintaining
breakfast foods in the house usually requires parental contribution. This is illustrated by an
American school-based study (Terre et al. 1990): 38% of 11-year-old children reported that
they prepared their breakfast themselves, 45% had it prepared by their mothers, 7% by
fathers, and 3% by siblings or grandparents; the rest either skipped breakfast (8%) or had it at
fast-food restaurants (1%). At age 16-18 in the same study, similarly 38% of the adolescents
prepared the breakfast themselves, but the proportion of breakfasts prepared by parents had
declined to 34% and breakfast prepared by other family members to 1%; breakfast was
skipped by 21% and eaten at fast-food restaurants by 5%. It would be of interest to assess
whether the changing levels of parental contribution are reflected in breakfast eating and other
meal patterns at earlier ages.
Our study on familial breakfast skipping patterns is one of the first to investigate sex-
limitation using twin-family models. Beyond the parameter estimates for genetic and
environmental effects described above, twin-family models were able to assess some
additional features, such as cultural transmission, assortative mating, and the gene-family
environment covariance. In our data, twin-family models broadly supported the results
obtained using classical twin models: in the phenotypic assortment model, the covariance
between genes and shared environment was small and negative, meaning that additive genetic
factors and common environmental factors were passed on relatively independently in this
sample, and that the classical twin model assumption that A-C correlation is zero was a valid
approximation. Although both the social homogamy models and the phenotypic assortment
models fit the data well, the overall fit was better and the parameter estimates were more
reasonable in the phenotypic assortment model. The somewhat unrealistic results of the social
homogamy models were not surprising. In the presence of strong spousal correlations these
models expect larger shared family environmental effects than are typically seen in behavior
genetic research. However, our data cannot explain whether spousal correlations are the result
of cohabitation or assortative mating.
Cultural transmission in breakfast eating was of borderline importance, not quite
statistically significant, and negative in direction, implying that the breakfast eating habits of
parents (or correlated behaviors) have little direct impact on the breakfast eating of children,
or that children tend to behave in opposition to parental example. Another likely possibility is
that the cultural transmission estimate is an artifact of the model's assumption that the same
genetic influences on breakfast eating are being expressed at the age of the children and the
age of the parents. Even though we found no significant effect of parental breakfast eating on
child behavior, we found significant effects of the family environment on adolescent breakfast
eating for both boys and girls, implying that other types of parental effects may be present. It
is possible, for instance, that the parents who encourage regular breakfasts for their children
do not follow their own advice, so that children are responding to something other than direct
It is also possible that the crucial environmental influences on this behavior change from
generation to generation; the environmental factors influencing the adolescents (even those
created by the parents) may differ from those that influence the same behavior in the parents.
The adolescents participating in this study were born in the 1970s and assessed in early 1990s;
their parents represent the post-WWII baby boomer generation. Although breakfast foods
have not changed as significantly during the latter part of the 20th century than during the
earlier part, when Finnish breakfast was still a warm meal, there is ample evidence from many
Western countries that eating patterns have become more irregular, convenience foods more
common, and shared family meals infrequent as a probable result from women entering the
workforce (Samuelson 2000; Briefel & Johnson 2004). The current cultural atmosphere also
prefers individualism over collectivism. The body shape ideals have certainly also changed
during this period (Barber 1998), perhaps increasing pressure to diet and giving the omission
of breakfast new contexts.
Although our twin-family models demonstrated that parental influences have relatively
little role on their children’s breakfast patterns, this influence was still much stronger than any
other influences we were able to detect (eg, sociodemographic factors, or the twin’s BMI).
Breakfast skipping likely constitutes a marker of decreased health-conscious attitudes.
Individuals who skip breakfast may have a higher risk of disordered eating and unhealthy
weight control practices. Thus, in school health settings, screening for breakfast and other
meal habits may offer a useful way to approach other potentially health-compromising
behaviors that are likely to cluster with breakfast skipping.
Strengths and limitations
The strength of this study is that we were able to study several novel phenotypes in a large
population-based sample. Its excellent population coverage and high response rate helped to
minimize non-response bias. The participants of our study were younger and more
homogenous in age than those of many previous studies; thus age x genotype bias was
In contrast to gender, SES was not as clearly associated with eating and weight related
behaviors. Breakfast skipping was associated with low family SES in adults and adolescent
boys, but not in girls. In body shape and weight concerns, the role of family SES and personal
educational attainment was even more limited. Probably weight-conscious attitudes are so
common in all levels of Finnish society that education level and SES make little difference.
However, in eating patterns and BMI these influences are still clear, and their role in weight
loss and maintenance in young adults should be further studied.
The greatest limitation in our analysis of eating and weight related behaviours was that for
many variables of interest, we were limited to a cross-sectional design. This study cannot
reveal whether a causal link exists between breakfast skipping and health-compromising
factors, eating styles and intentional weight loss, or psychosomatic symptoms and body shape
and weight concerns. Nor can it elucidate the direction of causality. Further studies of
prospective samples starting at an early age are needed to resolve this issue; we hope that our
undertakings can serve as a baseline assessment to prospective evaluations of these traits in
Some further caveats apply to interpreting our results, particularly those pertaining to
weight loss and body image. One must be mindful of the gender differences in prevalences
and distributions of those traits. To some degree, we failed to take all possible gender
differences into account in our analyses. Sources of body shape and weight concerns are likely
different in males and females. The measure that we used, the EDI, focuses on core areas of
female body and weight dissatisfaction. Thus BD and DT subscales as measured in EDI are
clearly not ideal measures of male body shape related attitudes, because domains of core
importance such as muscularity and stature are completely ignored (Cohane & Pope, Jr. 2001;
McCabe et al. 2002). We addressed this concern half-way through our raw data collection by
adding questions about dissatisfaction with musculature and stature. Preliminary analyses from
these items reflect that these male-specific body shape concerns also have clear associations
with depressive symptoms and various health-compromising behaviours (Raevuori et al.
The importance of gender differences was apparent in our genetic analyses. In none of the
phenotypes studied here was it reasonable to constrain males and females to be equal. In
breakfast skipping, a gender-limitation approach was used. In body shape and weight
concerns, the very pronounced gender differences necessitated different modelling strategies
for each gender (continuous measures for women, dichotomous for men). Because of this,
results from male and female models may not be directly comparable, and loss of power due
to dichotomization was reflected in the wide confidence intervals and the relative instability of
male parameter estimates. Concerning weight loss, it is not clear whether women and men
understood, mentally recorded, and finally reported in their questionnaires weight loss
attempts similarly, although our key question was phrased in a gender-neutral way. In addition
to different genetic influences, weight-loss behaviours also probably harbour very distinct
socially learned gender roles: although many more men than women were overweight in our
sample, and the prevalence of obesity was equal across genders, women were much more
likely to engage in experiencing body shape and weight dissatisfaction and to attempt weight
The generalizability of results from twin studies to the general population is often
questioned. Some caveats apply to interpreting BMI-related measures derived from a twin
population. MZ twins are, typically, smaller at birth than DZ twins. In an earlier analysis of
this longitudinal sample, the size difference seemed to persist until the end of puberty in males
(Pietiläinen et al. 1999). However, at 22-27, the differences in BMI means between male MZ
and DZ twins were no longer statistically significant (Schousboe et al, 2003). Other studies of
adult Finnish twins have also shown that the BMIs of twins and non-twins are comparable
(Rissanen et al. 1988; Korkeila et al. 1991). The prevalences of breakfast eating in our twins
and their parents were similar to those obtained from three large non-twin populations from
Sweden, the U.S., and Finland (Puska & Smolander 1980; Höglund et al. 1998; Siega-Riz et al.
1998). Also, our measures of body shape and weight concerns, EDI subscale means were in
line with age- and gender-specific norms from the Finnish non-twin EDI validation sample
(Charpentier et al, unpublished manuscript). All this suggests that conclusions derived from
the twin population can be extended to the entire population. However, heritability estimates
are always specific to the environment, cohort, and culture studied (Kendler et al. 2000;
Kendler 2001). Our results likely apply to adolescent and young adults in Western cultures,
but cannot necessarily be extended to other times and places: if cultural influences and
environment are different from those addressed in this study, heritability estimates are
expected to vary accordingly.
Eating and body image related issues have general health implications: breakfast skipping was
associated with health-compromising behaviors. Dissatisfaction with body shape and weight
were associated with larger body size and multiple psychosomatic symptoms. Intentional
weight loss attempts were also associated with larger body size and potentially disordered
eating styles, particularly restricting and overeating.
Young adults, particularly women, who are dissatisfied with their appearances, are often
overlooked by health care services. Yet these types of complaints are very common in the
context of school health services. Perhaps in this population, brief routine screenings for
symptoms of anxiety and depression and substance use might serve as a warranted early
intervention. As health-compromising behaviors exhibit moderate clustering, multifaceted
health promotion efforts addressing also other health-compromising factors, such as tobacco
and alcohol use, need to be considered. General education programs about healthful eating
patterns and healthy weight control methods should be implemented in schools. Because
long-term solutions to overweight and obesity are still difficult to come by, preventive efforts
should receive increased attention and funding. Currently, prevention is the best solution to
Overall, both genes and environment were important for the eating-related phenotypes
studied in Finnish young adult twins. Breakfast skipping in girls and boys, intentional weight
loss in women and men, and body shape and weight concerns in women exhibited moderate
to substantial genetic influences. Body shape and weight concerns exhibited substantial and
breakfast skipping modest influences of shared family environment. The genetic influences on
intentional weight loss were only partially shared with those affecting BMI. This means that
for all of the phenotypes studied, there is clearly room for intervention in modifying our living
environment to a healthier, less obesogenic direction. This task is daunting and likely to
involve all aspects of society from zoning planners and legislators to health educators and
school and workplace cafeteria staff. Parents of children and teenagers have a particular role in
this effort. However, it is a task well worth undertaking.
Figure 10. Thank you for helping me to complete this journey.
This project has been a truly collaborative effort. I’d like to express my heartfelt thanks to
♥ Aila, who inspired me to embark on this journey, and then patiently and lovingly
mentored, mothered, and nurtured this research, always opening new horizons.
♥ Jaakko: thank you for giving us the opportunity to study eating disorders and
disordered eating in Finnish twins. You have been helping along at every single step of
this project. Particular thanks for quiet instilled wisdom and lessons in zen.
♥ Matti, for being so encouraging in getting me started on my project.
♥ Dick, for excellent commentary and for making twin studies in Finland possible.
♥ Cindy and Pat; for being such great role models, and for providing good insights, good
company, and for sharing your wonderful family with me; Brendan, Natalie, and Emily
for comic relief during the snowy winter in Virginia, and later in Chapel Hill.
♥ Ben, for great Mx tutoring and getting back to me by e-mail in various emergencies.
♥ Rick, for your patience and persistence with the complexities of twin and family
♥ Jennifer and Ritva, who kindly reviewed this thesis, coming up with many helpful
suggestions: thanks for your time and your flexibility.
♥ various funding agencies for their generous support, particularly the U.S. National
Institutes of Health (NIMH & NIAAA), Yrjö Jahnsson Foundation, Jalmari and
Rauha Ahokas Foundation, Helsingin Sanomat Foundation, Helsinki University
Central Hospital, Finnish Cultural Foundation, Biomedicum Foundation, and
Doctoral Programs in Public Health.
♥ the European Union 5th Framework program, for interdisciplinary fun, introduction to
phenotypes and genotypes, great food; for helping to appreciate the beauty of
European cities and defining impulsive and compulsive shopping: Andreas, Fernando,
Janet, David, Marija, Kate, Jo, Frances, Anke, and others.
♥ the Columbia University PET program, particularly Sharon Schwarz, Bruce Link, and
Hans Wijbrand Hoek, for providing the opportunity, space, and funding for the final
crush, and for making my life richer with all the cultural diversity New York City can
offer. And to Deidre, Larry, Beverly, and others - thanks listening to my rants and the
opportunity to share our goals.
♥ Kirsi and Elina, for great company and shared moments – without you this project
would have been so boring.
♥ Anu R and Hanna-Reetta for the joy of discovering new thing together.
♥ Erjastiina, Leila, Salla, Synnöve, Timo, Sami, Anu K, Jutta, Seija, Anna-Maarit, Hanna-
Mari and everybody else involved in EDNEURO and EDSCREEN.
♥ Kristian Wahlbeck, Ranan Rimón, Jari Tiihonen, Heikki Vartiainen, Björn Appelberg,
and Hasse Karlsson at the Department of Psychiatry for their encouragement.
♥ my supervisors and colleagues at the Eating Disorder Unit in Helsinki – particular
thanks to Pirkko Nilsson and Riikka Viljanen for staying in touch, never mind the
♥ Kauko and Totte, for brightening my days with your great company, and for your
♥ Pia, Aku, and Eila, for always responding to my bizarre requests with a helpful smile.
♥ the 5th floor coffee break crew at the Helsinki University Dept. of Public Health, for
interdisciplinary talk and for making the start of the day such fun.
♥ Milla, Laura, Timi, Jari, Susanna, Jaana, Cedi, Sanna, Elina, Manne, Pauliina, Seija,
Pekka, Danilo, and Soila for your lasting friendship.
♥ Aija and Maija for your very positive contributions to our family environment
♥ Olli, for always promptly helping with web programming and graphics.
♥ Antti, for shared family environment, computer support, and tours of greater Boston.
♥ mother and father, for genes, environment, inspiration, talks, practical advice, and your
♥ my dear husband Pekka for your unconditional love and backing – and for your
readiness to embark on adventures, be they small, mid-sized, or big.
New York City, November 2004
Abalkhail, B. and Shawky, S. (2002). Prevalence of daily breakfast intake, iron deficiency anaemia and awareness of being
anaemic among Saudi school students. Int. J. Food Sci. Nutr. 53, 519-528.
Abdollahi, P. and Mann, T. (2001). Eating disorder symptoms and body image concerns in Iran: comparisons between Iranian
women in Iran and in America. Int. J. Eat. Disord. 30, 259-268.
Abraham, S. (1996). Eating and weight controlling behaviours of young ballet dancers. Psychopathology 29, 218-222.
Abraham, S. F., Mira, M., Beumont, P. J., Sowerbutts, T. D., and Llewellyn-Jones, D. (1983). Eating behaviours among young
women. Med. J. Aust. 2, 225-228.
Ackard, D. M. and Neumark-Sztainer, D. (2002). Date violence and date rape among adolescents: associations with
disordered eating behaviors and psychological health. Child Abuse Negl. 26, 455-473.
Ackard, D. M. and Peterson, C. B. (2001). Association between puberty and disordered eating, body image, and other
psychological variables. Int. J. Eat. Disord. 29, 187-194.
Adeyemo, A., Luke, A., Cooper, R., Wu, X., Tayo, B., Zhu, X., Rotimi, C., Bouzekri, N., and Ward, R. (2003). A genome-wide
scan for body mass index among Nigerian families. Obes. Res. 11, 266-273.
al Sudairy, A. and Howard, K. (1992). Dietary habits of technical and vocational students in Riyadh, Saudi Arabia--I. Meal
skipping. J. R. Soc. Health 112, 217-218.
Anderson, D. A., Lundgren, J. D., Shapiro, J. R., and Paulosky, C. A. (2003). Weight goals in a college-age population. Obes.
Res. 11, 274-278.
Anderson, J. W., Konz, E. C., Frederich, R. C., and Wood, C. L. (2001). Long-term weight-loss maintenance: a meta-analysis
of US studies. Am. J. Clin. Nutr. 74, 579-584.
Barber, N. (1998). Secular changes in standards of bodily attractiveness in American women: different masculine and feminine
ideals. J. Psychol. 132, 87-94.
Bartlett, S. J., Wadden, T. A., and Vogt, R. A. (1996). Psychosocial consequences of weight cycling. J. Consult Clin. Psychol.
Baumert, P. W., Jr., Henderson, J. M., and Thompson, N. J. (1998). Health risk behaviors of adolescent participants in
organized sports. J. Adolesc. Health 22, 460-465.
Becker, A. E., Burwell, R. A., Gilman, S. E., Herzog, D. B., and Hamburg, P. (2002). Eating behaviours and attitudes
following prolonged exposure to television among ethnic Fijian adolescent girls. Br. J. Psychiatry 180, 509-514.
Beglin, S. J. and Fairburn, C. G. (1992). What is meant by the term "binge"? Am. J. Psychiatry 149, 123-124.
Bellisle, F., Monneuse, M. O., Steptoe, A., and Wardle, J. (1995). Weight concerns and eating patterns: a survey of university
students in Europe. Int. J. Obes. Relat Metab Disord. 19, 723-730.
Ben Tovim, D. I. and Walker, M. K. (1991). The development of the Ben-Tovim Walker Body Attitudes Questionnaire
(BAQ), a new measure of women's attitudes towards their own bodies. Psychol. Med. 21, 775-784.
Berkey, C. S., Rockett, H. R., Gillman, M. W., Field, A. E., and Colditz, G. A. (2003). Longitudinal study of skipping breakfast
and weight change in adolescents. Int. J. Obes. Relat Metab Disord. 27, 1258-1266.
Biener, L. and Heaton, A. (1995). Women dieters of normal weight: their motives, goals, and risks. Am. J. Public Health 85,
Blundell, J. E. and Gillett, A. (2001). Control of food intake in the obese. Obes. Res. 9 Suppl 4, 263S-270S.
Borresen, R. and Rosenvinge, J. H. (2003). Body dissatisfaction and dieting in 4,952 Norwegian children aged 11-15 years: less
evidence for gender and age differences. Eat. Weight. Disord. 8, 238-241.
Boutelle, K., Neumark-Sztainer, D., Story, M., and Resnick, M. (2002). Weight control behaviors among obese, overweight,
and nonoverweight adolescents. J. Pediatr. Psychol. 27, 531-540.
Branson, R., Potoczna, N., Kral, J. G., Lentes, K. U., Hoehe, M. R., and Horber, F. F. (2003). Binge eating as a major
phenotype of melanocortin 4 receptor gene mutations. N. Engl. J. Med. %20;348, 1096-1103.
Breitkopf, C. R. and Berenson, A. B. (2004). Correlates of weight loss behaviors among low-income African-American,
Caucasian, and Latina women. Obstet. Gynecol. 103, 231-239.
Briefel, R. R. and Johnson, C. L. (2004). Secular trends in dietary intake in the United States. Annu. Rev. Nutr. 24:401-431.,
Brownell, K. D. and Rodin, J. (1994). Medical, metabolic, and psychological effects of weight cycling. Arch. Intern. Med. 154,
Brugman, E., Meulmeester, J. F., Spee-van der Wekke, A., Beuker, R. J., Zaadstra, B. M., Radder, J. J., and Verloove-
Vanhorick, P. S. (1997). Dieting, weight and health in adolescents in The Netherlands. Int. J. Obes. Relat Metab
Disord. 21, 54-60.
Brugman, E., Meulmeester, J.F., Spee-van der Wekke, A., & Verloove-Vanhorick, S.P. (1998). Breakfast-skipping in children
and young adolescents in The Netherlands. Eur.J.Public Health 8, 325-328.
Buddeberg-Fischer, B., Bernet, R., Schmid, J., and Buddeberg, C. (1996). Relationship between disturbed eating behavior and
other psychosomatic symptoms in adolescents. Psychother. Psychosom. 65, 319-326.
Bulik, C. M., Sullivan, P. F., and Kendler, K. S. (2003). Genetic and environmental contributions to obesity and binge eating.
Int. J. Eat. Disord. 33, 293-298.
Bulik, C. M., Sullivan, P. F., Wade, T. D., and Kendler, K. S. (2000). Twin studies of eating disorders: a review. Int. J. Eat.
Disord. 27, 1-20.
Bulik, C. M., Wade, T. D., Heath, A. C., Martin, N. G., Stunkard, A. J., and Eaves, L. J. (2001). Relating body mass index to
figural stimuli: population-based normative data for Caucasians. Int. J. Obes. Relat Metab Disord. 25, 1517-1524.
Casper, R. C. (1998). Depression and eating disorders. Depress. Anxiety. 8 Suppl 1, 96-104.
Cavadini, C., Decarli, B., Grin, J., Narring, F., and Michaud, P. A. (2000). Food habits and sport activity during adolescence:
differences between athletic and non-athletic teenagers in Switzerland. Eur. J. Clin. Nutr. 54 Suppl 1, S16-S20.
Chandy, J. M., Harris, L., Blum, R. W., and Resnick, M. D. (1994). Disordered eating among adolescents whose parents
misuse alcohol: protective and risk factors. Int. J. Addict. 29, 505-516.
Chandy, J. M., Harris, L., Blum, R. W., and Resnick, M. D. (1995). Female adolescents of alcohol misusers: disordered eating
features. Int. J. Eat. Disord. 17, 283-289.
Cho, S., Dietrich, M., Brown, C. J., Clark, C. A., and Block, G. (2003). The effect of breakfast type on total daily energy intake
and body mass index: results from the Third National Health and Nutrition Examination Survey (NHANES III). J.
Am. Coll. Nutr. 22, 296-302.
Coady, S. A., Jaquish, C. E., Fabsitz, R. R., Larson, M. G., Cupples, L. A., and Myers, R. H. (2002). Genetic variability of adult
body mass index: a longitudinal assessment in framingham families. Obes. Res. 10, 675-681.
Cohane, G. H. and Pope, H. G., Jr. (2001). Body image in boys: a review of the literature. Int. J. Eat. Disord. 29, 373-379.
Critser, G. (2003). Fat land: how Americans became the fattest people in the world. Houghton Mifflin Company, Boston.
Dalla Grave, R., De Luca, L., and Oliosi, M. (1997). Eating attitudes and prevalence of eating disorders: a survey in secondary
schools in Lecce, southern Italy. Eat. Weight. Disord. 2, 34-37.
de Castro, J. M. (1993). Genetic influences on daily intake and meal patterns of humans. Physiol Behav. 53, 777-782.
de Castro, J. M. (2001). Heritability of diurnal changes in food intake in free-living humans. Nutrition 17, 713-720.
Devlin, B., Bacanu, S. A., Klump, K. L., Bulik, C. M., Fichter, M. M., Halmi, K. A., Kaplan, A. S., Strober, M., Treasure, J.,
Woodside, D. B., Berrettini, W. H., and Kaye, W. H. (2002). Linkage analysis of anorexia nervosa incorporating
behavioral covariates. Hum. Mol. Genet. 11, 689-696.
Dohm, F. A., Beattie, J. A., Aibel, C., and Striegel-Moore, R. H. (2001). Factors differentiating women and men who
successfully maintain weight loss from women and men who do not. J. Clin. Psychol. 57, 105-117.
Eaves, L. J., Fulker, D. W., and Heath, A. C. (1989). The effects of social homogamy and cultural inheritance on the
covariances of twins and their parents: a LISREL model. Behav. Genet. 19, 113-122.
Fairburn, C. G. and Cooper, P. J. (1983). The Eating Disorder Examination (12th Edition). In C. G. Fairburn & G. T. Wilson
(Eds.), Binge eating: Nature, assessment, and treatment. Guilford Press, New York.
Fairburn, C. G., Cowen, P. J., and Harrison, P. J. (1999). Twin studies and the etiology of eating disorders. Int. J. Eat. Disord.
Fairburn, C. G., Welch, S. L., Doll, H. A., Davies, B. A., and O'Connor, M. E. (1997). Risk factors for bulimia nervosa. A
community-based case-control study. Arch. Gen. Psychiatry 54, 509-517.
Falconer, D. S. and Mackay, T. F. C. (1996). Introduction to quantitative genetics. Prentice Hall, Harlow.
Falkenberg, M. and Hellman, M. (2002). [Porridge breakfast--a prescription for weight loss]. Lakartidningen 99, 3631-3632.
Farooqi, I. S., Keogh, J. M., Yeo, G. S., Lank, E. J., Cheetham, T., and O'Rahilly, S. (2003). Clinical spectrum of obesity and
mutations in the melanocortin 4 receptor gene. N. Engl. J. Med. 348, 1085-1095.
Field, A. E., Wolf, A. M., Herzog, D. B., Cheung, L., and Colditz, G. A. (1993). The relationship of caloric intake to frequency
of dieting among preadolescent and adolescent girls. J. Am. Acad. Child Adolesc. Psychiatry 32, 1246-1252.
Fisher, M., Schneider, M., Pegler, C., and Napolitano, B. (1991). Eating attitudes, health-risk behaviors, self-esteem, and
anxiety among adolescent females in a suburban high school. J. Adolesc. Health 12, 377-384.
Fitzgibbon, M. L., Stolley, M. R., and Kirschenbaum, D. S. (1993). Obese people who seek treatment have different
characteristics than those who do not seek treatment. Health Psychol. 12, 342-345.
Flier, J. S. (2004). Obesity wars: molecular progress confronts an expanding epidemic. Cell 116, 337-350.
Foster, G. D., Sarwer, D. B., and Wadden, T. A. (1997). Psychological effects of weight cycling in obese persons: a review and
research agenda. Obes. Res. 5, 474-488.
Foster, G. D., Wadden, T. A., Kendall, P. C., Stunkard, A. J., and Vogt, R. A. (1996). Psychological effects of weight loss and
regain: a prospective evaluation. J. Consult Clin. Psychol. 64, 752-757.
French, S. A. and Jeffery, R. W. (1997). Current dieting, weight loss history, and weight suppression: behavioral correlates of
three dimensions of dieting. Addict. Behav. 22, 31-44.
French, S. A., Jeffery, R. W., Folsom, A. R., Williamson, D. F., and Byers, T. (1995a). Relation of weight variability and
intentionality of weight loss to disease history and health-related variables in a population-based sample of women
aged 55-69 years. Am. J. Epidemiol. 142, 1306-1314.
French, S. A., Jeffery, R. W., and Wing, R. R. (1994). Food intake and physical activity: a comparison of three measures of
dieting. Addict. Behav. 19, 401-409.
French, S. A., Perry, C. L., Leon, G. R., and Fulkerson, J. A. (1995b). Changes in psychological variables and health behaviors
by dieting status over a three-year period in a cohort of adolescent females. J. Adolesc. Health 16, 438-447.
Friedman, M. A., Schwartz, M. B., and Brownell, K. D. (1998). Differential relation of psychological functioning with the
history and experience of weight cycling. J. Consult Clin. Psychol. 66, 646-650.
Fulkerson, J. A., McGuire, M. T., Neumark-Sztainer, D., Story, M., French, S. A., and Perry, C. L. (2002). Weight-related
attitudes and behaviors of adolescent boys and girls who are encouraged to diet by their mothers. Int. J. Obes. Relat
Metab Disord. 26, 1579-1587.
Garner, D.M. (1991). Eating disorder inventory 2, professional manual. Odessa, FL: Psychological Assessment Resources,
Garner, D. M. and Olmsted, M. P. The eating disorder inventory manual. Psychological Assessment Resources, Odessa, FL.
Gatenby, S. J. (1997). Eating frequency: methodological and dietary aspects. Br. J. Nutr. 77 Suppl 1, S7-20.
Golla, A., Strauch, K., Dietter, J., and Baur, M. P. (2003). Quantitative trait linkage analysis of longitudinal change in body
weight. BMC. Genet. 4 Suppl 1:S7., S7.
Gordon, R. A. (2000). Eating disorders. Anatomy of a social epidemic. Second edition. Blackwell Publishers, Oxford.
Gottesman, I. I. and Gould, T. D. (2003). The endophenotype concept in psychiatry: etymology and strategic intentions. Am.
J. Psychiatry 160, 636-645.
Green, K. L., Cameron, R., Polivy, J., Cooper, K., Liu, L., Leiter, L., and Heatherton, T. (1997). Weight dissatisfaction and
weight loss attempts among Canadian adults. Canadian Heart Health Surveys Research Group. CMAJ. 157 Suppl 1,
Grigg, M., Bowman, J., and Redman, S. (1996). Disordered eating and unhealthy weight reduction practices among adolescent
females. Prev. Med. 25, 748-756.
Grilo, C. M., Masheb, R. M., and Wilson, G. T. (2001). Subtyping binge eating disorder. J. Consult Clin. Psychol. 69, 1066-
Gross, S. M., Bronner, Y., Welch, C., Dewberry-Moore, N., and Paige, D. M. (2004). Breakfast and lunch meal skipping
patterns among fourth-grade children from selected public schools in urban, suburban, and rural maryland. J. Am.
Diet. Assoc. 104, 420-423.
Haines, P. S., Guilkey, D. K., and Popkin, B. M. (1996). Trends in breakfast consumption of US adults between 1965 and
1991. J. Am. Diet. Assoc. 96, 464-470.
Harris, M. B. (1983). Eating habits, restraint, knowledge and attitudes toward obesity. Int. J. Obes. 7, 271-286.
Heatherton, T. F., Mahamedi, F., Striepe, M., Field, A. E., and Keel, P. (1997). A 10-year longitudinal study of body weight,
dieting, and eating disorder symptoms. J. Abnorm. Psychol. 106, 117-125.
Heitmann, B. L., Harris, J. R., Lissner, L., and Pedersen, N. L. (1999). Genetic effects on weight change and food intake in
Swedish adult twins. Am. J. Clin. Nutr. 69, 597-602.
Heitmann, B. L., Kaprio, J., Harris, J. R., Rissanen, A., Korkeila, M., and Koskenvuo, M. (1997). Are genetic determinants of
weight gain modified by leisure-time physical activity? A prospective study of Finnish twins. Am. J. Clin. Nutr. 66,
Helakorpi, S., Patja, K., Prättälä, R., Aro, A. R., and Uutela, A. (2003). Health behaviour and health among Finnish adult
population, spring 2003. Publications of the National Public Health Institute, Helsinki.
Hewitt, J. K. (1997). The genetics of obesity: what have genetic studies told us about the environment. Behav. Genet. 27, 353-
Höglund, D., Samuelson, G., and Mark, A. (1998). Food habits in Swedish adolescents in relation to socioeconomic
conditions. Eur. J. Clin. Nutr. 52, 784-789.
Holland, A.J., Sicotte, N., & Treasure, J. (1988). Anorexia nervosa: evidence for a genetic basis. J.Psychosom.Res. 32, 561-
Hopper, J. L. (2000). Why 'common environmental effects' are so uncommon in the literature. In: Spector, T.D., Snieder, H.,
and MacGregor, A.J. (eds): Advances in twin and sib-pair analysis. London.
Hosmer, D. W. and Lemeshow, S. (2000). Applied logistic regression. Wiley, New York.
Huon, G. F. and Strong, K. G. (1998). The initiation and the maintenance of dieting: structural models for large-scale
longitudinal investigations. Int. J. Eat. Disord. 23, 361-369.
Hur, Y. M., Bouchard, T. J., Jr., and Eckert, E. (1998). Genetic and environmental influences on self-reported diet: a reared-
apart twin study. Physiol Behav. 64, 629-636.
Irala-Estevez, J. D., Groth, M., Johansson, L., Oltersdorf, U., Prattala, R., and Martinez-Gonzalez, M. A. (2000). A systematic
review of socio-economic differences in food habits in Europe: consumption of fruit and vegetables. Eur. J. Clin.
Nutr. 54, 706-714.
Isralowitz, R. E. and Trostler, N. (1996). Substance use: toward an understanding of its relation to nutrition-related attitudes
and behavior among Israeli high school youth. J. Adolesc. Health 19, 184-189.
Jaeger, B., Ruggiero, G. M., Edlund, B., Gomez-Perretta, C., Lang, F., Mohammadkhani, P., Sahleen-Veasey, C., Schomer, H.,
and Lamprecht, F. (2002). Body Dissatisfaction and Its Interrelations with Other Risk Factors for Bulimia nervosa
in 12 Countries. Psychother. Psychosom. 71, 54-61.
Johnson, M. D. (1994). Disordered eating in active and athletic women. Clin. Sports Med. 13, 355-369.
Kagan, D. M. and Squires, R. L. (1984). Eating disorders among adolescents: patterns and prevalence. Adolescence 19, 15-29.
Kaltiala-Heino, R., Kautiainen, S., Virtanen, S. M., Rimpelä, A., and Rimpelä, M. (2003a). Has the adolescents' weight concern
increased over 20 years? Eur. J. Public Health 13, 4-10.
Kaltiala-Heino, R., Marttunen, M., Rantanen, P., and Rimpelä, M. (2003b). Early puberty is associated with mental health
problems in middle adolescence. Soc. Sci. Med. 57, 1055-1064.
Kaltiala-Heino, R., Rimpelä, M., Rissanen, A., and Rantanen, P. (2001). Early puberty and early sexual activity are associated
with bulimic-type eating pathology in middle adolescence. J. Adolesc. Health 28, 346-352.
Kamakura, T., Ando, J., Ono, Y., and Maekawa, H. (2003). A twin study of genetic and environmental influences on
psychological traits of eating disorders in a Japanese female sample. Twin. Res. 6, 292-296.
Kaprio, J., Pulkkinen, L., and Rose, R. J. (2002). Genetic and environmental factors in health-related behaviors: studies on
Finnish twins and twin families. Twin. Res. 5, 366-371.
Kaufmann, N. A., Poznanski, R., and Guggenheim, K. (1975). Eating habits and opinions of teen-agers on nutrition and
obesity. J. Am. Diet. Assoc. 66, 264-268.
Keller, K. L., Pietrobelli, A., and Faith, M. S. (2003). Genetics of food intake and body composition: lessons from twin
studies. Acta Diabetol. 40 Suppl 1, S95-100.
Keller, K. L., Pietrobelli, A., Must, S., and Faith, M. S. (2002). Genetics of eating and its relation to obesity. Curr. Atheroscler.
Rep. 4, 176-182.
Kendler, K. S. (2001). Twin studies of psychiatric illness: an update. Arch. Gen. Psychiatry 58, 1005-1014.
Kendler, K. S., Thornton, L. M., and Pedersen, N. L. (2000). Tobacco consumption in Swedish twins reared apart and reared
together. Arch. Gen. Psychiatry 57, 886-892.
Kensinger, G. J., Murtaugh, M. A., Reichmann, S. K., and Tangney, C. C. (1998). Psychological symptoms are greater among
weight cycling women with severe binge eating behavior. J. Am. Diet. Assoc. 98, 863-868.
Keys, A., Brozek, J., Henschel, A., Mickelsen, O., and Taylor, H. L. (1950). The biology of human starvation. University of
Minnesota Press, Minneapolis, MN.
Kleemola, P., Roos, E., and Prättälä, R. (1997). Päivän tärkein ateria - aterioiden koostumus vuonna 1992 [The most
important meal of the day - composition of meals in 1992. In Finnish.]. Suom. Lääkäril. 52, 1027-1032.
Klem, M. L., Wing, R. R., McGuire, M. T., Seagle, H. M., and Hill, J. O. (1997). A descriptive study of individuals successful
at long-term maintenance of substantial weight loss. Am. J. Clin. Nutr. 66, 239-246.
Klump, K. L., McGue, M., and Iacono, W. G. (2000). Age differences in genetic and environmental influences on eating
attitudes and behaviors in preadolescent and adolescent female twins. J. Abnorm. Psychol. 109, 239-251.
Knapp, C. (2003). Appetites: why women want. Counterpoint Press, New York.
Korkeila, M., Kaprio, J., Rissanen, A., and Koskenvuo, M. (1991). Effects of gender and age on the heritability of body mass
index. Int. J. Obes. 15, 647-654.
Korkeila, M., Kaprio, J., Rissanen, A., and Koskenvuo, M. (1995). Consistency and change of body mass index and weight. A
study on 5967 adult Finnish twin pairs. Int. J. Obes. Relat Metab Disord. 19, 310-317.
Korkeila, M., Rissanen, A., Kaprio, J., Sorensen, T. I., and Koskenvuo, M. (1999). Weight-loss attempts and risk of major
weight gain: a prospective study in Finnish adults. Am. J. Clin. Nutr. 70, 965-975.
Laessle, R. G., Platte, P., Schweiger, U., and Pirke, K. M. (1996). Biological and psychological correlates of intermittent dieting
behavior in young women. A model for bulimia nervosa. Physiol Behav. 60, 1-5.
Lahti-Koski, M., Vartiainen, E., Mannisto, S., and Pietinen, P. (2000). Age, education and occupation as determinants of
trends in body mass index in Finland from 1982 to 1997. Int. J. Obes. Relat Metab Disord. 24, 1669-1676.
Lautenbacher, S., Thomas, A., Roscher, S., Strian, F., Pirke, K. M., and Krieg, J. C. (1992). Body size perception and body
satisfaction in restrained and unrestrained eaters. Behav. Res. Ther. 30, 243-250.
Levy, A. S. and Heaton, A. W. (1993). Weight control practices of U.S. adults trying to lose weight. Ann. Intern. Med. 119,
Ma, Y., Bertone, E. R., Stanek, E. J., III, Reed, G. W., Hebert, J. R., Cohen, N. L., Merriam, P. A., and Ockene, I. S. (2003).
Association between eating patterns and obesity in a free-living US adult population. Am. J. Epidemiol. 158, 85-92.
Maes, H. H., Neale, M. C., and Eaves, L. J. (1997). Genetic and environmental factors in relative body weight and human
adiposity. Behav. Genet. 27, 325-351.
Männistö, S., Ovaskainen, M.-L., and Valsta, L. ed. (2003). Finravinto 2002 -tutkimus. The National FINDIET 2002 Study.
Mattes, R. D. (2002). Ready-to-eat cereal used as a meal replacement promotes weight loss in humans. J. Am. Coll. Nutr. 21,
McCabe, M. P., Ricciardelli, L. A., and Finemore, J. (2002). The role of puberty, media and popularity with peers on strategies
to increase weight, decrease weight and increase muscle tone among adolescent boys and girls. J. Psychosom. Res.
McGuire, M. T., Wing, R. R., Klem, M. L., and Hill, J. O. (1999a). Behavioral strategies of individuals who have maintained
long-term weight losses. Obes. Res. 7, 334-341.
McGuire, M. T., Wing, R. R., Klem, M. L., Lang, W., and Hill, J. O. (1999b). What predicts weight regain in a group of
successful weight losers? J. Consult Clin. Psychol. 67, 177-185.
Melnik, T. A., Rhoades, S. J., Wales, K. R., Cowell, C., and Wolfe, W. S. (1998). Food consumption patterns of elementary
schoolchildren in New York City. J. Am. Diet. Assoc. 98, 159-164.
Meltzer, A. A. and Everhart, J. E. (1996). Correlations with self-reported weight loss in overweight U.S. adults. Obes. Res. 4,
Melve, K. K. and Baerheim, A. (1994). Signs of subclinical eating disorders in teenage girls. Scand. J. Prim. Health Care 12,
Mikkilä, V., Lahti-Koski, M., Pietinen, P., Virtanen, S. M., and Rimpelä, M. (2003). Associates of obesity and weight
dissatisfaction among Finnish adolescents. Public Health Nutr. 6, 49-56.
Moore, D. C. (1988). Body image and eating behavior in adolescent girls. Am. J. Dis. Child 142, 1114-1118.
Morgan, D. R. ed. (2000). Eating disorders, body image & the media. British Medical Association, London.
Murata, M. (2000). Secular trends in growth and changes in eating patterns of Japanese children. Am. J. Clin. Nutr. 72, 1379S-
Murphy, J. M., Pagano, M. E., Nachmani, J., Sperling, P., Kane, S., and Kleinman, R. E. (1998). The relationship of school
breakfast to psychosocial and academic functioning: cross-sectional and longitudinal observations in an inner- city
school sample. Arch. Pediatr. Adolesc. Med. 152, 899-907.
Neale, M. C., Boker, S. M., Xie, G., and Maes, H. H. (2002). Mx: Statistical modeling, 6th edition. Department of Psychiatry,
Virginia Commonwealth University, Richmond, VA.
Neale, M. C. and Cardon, L. C. (1992). Methodology for genetic studies of twins and families. Kluwer Academic Publishers,
Neale, M. C., Walters, E. E., Eaves, L. J., Maes, H. H., and Kendler, K. S. (1994). Multivariate genetic analysis of twin-family
data on fears: Mx models. Behav. Genet. 24, 119-139.
Neel, J. V. (1962). Diabetes mellitus: a "thrifty" genotype rendered detrimental by "progress"? Am. J. Hum. Genet. 14:353-62.,
Neumark-Sztainer, D. and Hannan, P. J. (2000). Weight-related behaviors among adolescent girls and boys: results from a
national survey. Arch. Pediatr. Adolesc. Med. 154, 569-577.
Neumark-Sztainer, D., Jeffery, R. W., and French, S. A. (1997a). Self-reported dieting: how should we ask? What does it
mean? Associations between dieting and reported energy intake. Int. J. Eat. Disord. 22, 437-449.
Neumark-Sztainer, D., Rock, C. L., Thornquist, M. D., Cheskin, L. J., Neuhouser, M. L., and Barnett, M. J. (2000). Weight-
control behaviors among adults and adolescents: associations with dietary intake. Prev. Med. 30, 381-391.
Neumark-Sztainer, D. and Story, M. (1998). Dieting and binge eating among adolescents: what do they really mean? J. Am.
Diet. Assoc. 98, 446-450.
Neumark-Sztainer, D., Story, M., Dixon, L. B., and Murray, D. M. (1998). Adolescents engaging in unhealthy weight control
behaviors: are they at risk for other health-compromising behaviors? Am. J. Public Health 88, 952-955.
Neumark-Sztainer, D., Story, M., and French, S. A. (1996a). Covariations of unhealthy weight loss behaviors and other high-
risk behaviors among adolescents. Arch. Pediatr. Adolesc. Med. 150, 304-308.
Neumark-Sztainer, D., Story, M., Resnick, M. D., and Blum, R. W. (1997b). Adolescent vegetarians. A behavioral profile of a
school-based population in Minnesota. Arch. Pediatr. Adolesc. Med. 151, 833-838.
Neumark-Sztainer, D., Story, M., Toporoff, E., Cassuto, N., Resnick, M. D., and Blum, R. W. (1996b). Psychosocial
predictors of binge eating and purging behaviors among adolescents with and without diabetes mellitus. J. Adolesc.
Health 19, 289-296.
Neumark-Sztainer, D., Story, M., Toporoff, E., Himes, J. H., Resnick, M. D., and Blum, R. W. (1997c). Covariations of eating
behaviors with other health-related behaviors among adolescents. J. Adolesc. Health 20, 450-458.
Nicklas, T. A., O'Neil, C. E., and Berenson, G. S. (1998). Nutrient contribution of breakfast, secular trends, and the role of
ready-to-eat cereals: a review of data from the Bogalusa Heart Study. Am. J. Clin. Nutr. 67, 757S-763S.
Nordlund, G., & Jacobson, T. (1999). Gymnasieelevernas matvanor. Gymnasieelevernas matvanor relaterat till hur de mår
och känner sig, deras fysiska aktivitet, deras skolprestationer ock sociala bakgrund. [High school students' food
habits in relation to their well-being, physical activity, school performance and social background; Swedish]. .
Umeå: Umeå universitetet.
Nowak, M. (1998). The weight-conscious adolescent: body image, food intake, and weight-related behavior. J. Adolesc.
Health 23, 389-398.
O'Dea, J. A. and Caputi, P. (2001). Association between socioeconomic status, weight, age and gender, and the body image
and weight control practices of 6- to 19-year-old children and adolescents. Health Educ. Res. 16, 521-532.
Palmer, R. L. (1979). The dietary chaos syndrome: a useful new term? Br. J. Med. Psychol. 52, 187-190.
Pastore, D. R., Fisher, M., and Friedman, S. B. (1996). Abnormalities in weight status, eating attitudes, and eating behaviors
among urban high school students: correlations with self-esteem and anxiety. J. Adolesc. Health 18, 312-319.
Pesa, J. (1999). Psychosocial factors associated with dieting behaviors among female adolescents. J. Sch Health 69, 196-201.
Pi-Sunyer, F. X. (1991). Health implications of obesity. Am. J. Clin. Nutr. 53, 1595S-1603S.
Pietiläinen, K. H., Kaprio, J., Rissanen, A., Winter, T., Rimpelä, A., Viken, R. J., and Rose, R. J. (1999). Distribution and
heritability of BMI in Finnish adolescents aged 16y and 17y: a study of 4884 twins and 2509 singletons. Int. J. Obes.
Relat Metab Disord. 23, 107-115.
Pijl, H. (2003). Reduced dopaminergic tone in hypothalamic neural circuits: expression of a "thrifty" genotype underlying the
metabolic syndrome? Eur. J. Pharmacol. 480, 125-131.
Pollitt, E. and Mathews, R. (1998). Breakfast and cognition: an integrative summary. Am. J. Clin. Nutr. 67, 804S-813S.
Powell, C. A., Walker, S. P., Chang, S. M., and Grantham-McGregor, S. M. (1998). Nutrition and education: a randomized
trial of the effects of breakfast in rural primary school children. Am. J. Clin. Nutr. 68, 873-879.
Prättälä, R. and Roos, E. (1999). From regional ethnographies to interdisciplinary approaches-research on meals in Finland.
Appetite 32, 66-72.
Puska, P. and Smolander, A. (1980). Suomalaisen aikuisväestön terveyskäyttäytyminen, kevät 1980. [The health habits of
Finnish adults, spring 1980. In Finnish.]. Kansanterveystieteen laitoksen julkaisuja B1.
Raevuori, A., Keski-Rahkonen, A., Kaprio, J., and Rissanen, A. (2004). Muscle dissatisfaction in a population sample of young
men. Int. J. Eat. Disord. 35, 430.
Rao J.N.K. and Scott A.J. (1984). On chi-squared tests for multiway contingency tables with cell proportions estimated from
survey data. Annals of Statistics 12, 46-60.
Reichborn-Kjennerud, T., Bulik, C. M., Kendler, K. S., Roysamb, E., Maes, H., Tambs, K., and Harris, J. R. (2003). Gender
differences in binge-eating: a population-based twin study. Acta Psychiatr. Scand. 108, 196-202.
Reichborn-Kjennerud, T., Bulik, C. M., Kendler, K. S., Roysamb, E., Tambs, K., Torgersen, S., and Harris, J. R. (2004).
Undue influence of weight on self-evaluation: a population-based twin study of gender differences. Int. J. Eat.
Disord. 35, 123-132.
Reynolds, C. A., Baker, L. A., and Pedersen, N. L. (1996). Models of spouse similarity: applications to fluid ability measured in
twins and their spouses. Behav. Genet. 26, 73-88.
Rice, T., Perusse, L., Bouchard, C., and Rao, D. C. (1999). Familial aggregation of body mass index and subcutaneous fat
measures in the longitudinal Quebec family study. Genet. Epidemiol. 16, 316-334.
Rierdan, J. and Koff, E. (1997). Weight, weight-related aspects of body image, and depression in early adolescent girls.
Adolescence 32, 615-624.
Rissanen, A., Heliovaara, M., and Aromaa, A. (1988). Overweight and anthropometric changes in adulthood: a prospective
study of 17,000 Finns. Int. J. Obes. 12, 391-401.
Robinson, S. M., Crozier, S. R., Borland, S. E., Hammond, J., Barker, D. J., and Inskip, H. M. (2004). Impact of educational
attainment on the quality of young women's diets. Eur. J. Clin. Nutr. ..
Rolls, B. J., Fedoroff, I. C., and Guthrie, J. F. (1991). Gender differences in eating behavior and body weight regulation.
Health Psychol. 10, 133-142.
Roos, E., Lahelma, E., Virtanen, M., Prattala, R., and Pietinen, P. (1998). Gender, socioeconomic status and family status as
determinants of food behaviour. Soc. Sci. Med. 46, 1519-1529.
Rose, R. J., Kaprio, J., Winter, T., Koskenvuo, M., and Viken, R. J. (1999). Familial and socioregional environmental effects
on abstinence from alcohol at age sixteen. J. Stud. Alcohol Suppl 13, 63-74.
Rose, R. J., Viken, R. J., Dick, D. M., Bates, J. E., Pulkkinen, L., and Kaprio, J. (2003). It does take a village: nonfamilial
environments and children's behavior. Psychol. Sci. 14, 273-277.
Rosmond, R. and Björntorp, P. (1998). Psychiatric ill-health of women and its relationship to obesity and body fat
distribution. Obes. Res. 6, 338-345.
Russell, G. (1979). Bulimia nervosa: an ominous variant of anorexia nervosa. Psychol. Med. 9, 429-448.
Rutherford, J., McGuffin, P., Katz, R. J., and Murray, R. M. (1993). Genetic influences on eating attitudes in a normal female
twin population. Psychol. Med. 23, 425-436.
Rydall, A. C., Rodin, G. M., Olmsted, M. P., Devenyi, R. G., and Daneman, D. (1997). Disordered eating behavior and
microvascular complications in young women with insulin-dependent diabetes mellitus. N. Engl. J. Med. 336, 1849-
Saar, K., Geller, F., Ruschendorf, F., Reis, A., Friedel, S., Schauble, N., Nurnberg, P., Siegfried, W., Goldschmidt, H. P.,
Schafer, H., Ziegler, A., Remschmidt, H., Hinney, A., and Hebebrand, J. (2003). Genome scan for childhood and
adolescent obesity in German families. Pediatrics 111, 321-327.
Samuelson, G. (2000). Dietary habits and nutritional status in adolescents over Europe. An overview of current studies in the
Nordic countries. Eur. J. Clin. Nutr. 54 Suppl 1:S21-8., S21-S28.
Samuelson, G., Bratteby, L. E., Enghardt, H., and Hedgren, M. (1996). Food habits and energy and nutrient intake in Swedish
adolescents approaching the year 2000 [published erratum appears in Acta Paediatr Suppl 1996 Nov;85(11):1392].
Acta Paediatr. Suppl 415, 1-19.
Sands, R., Tricker, J., Sherman, C., Armatas, C., and Maschette, W. (1997). Disordered eating patterns, body image, self-
esteem, and physical activity in preadolescent school children. Int. J. Eat. Disord. 21, 159-166.
Sarlio-Lähteenkorva, S., Parna, K., Palosuo, H., Zhuravleva, I., and Mussalo-Rauhamaa, H. (2003). Weight satisfaction and
self-esteem among teenagers in Helsinki, Moscow and Tallinn. Eat. Weight. Disord. 8, 289-295.
Sarlio-Lähteenkorva, S. and Rissanen, A. (1998). Weight loss maintenance: determinants of long-term success. Eat. Weight.
Disord. 3, 131-135.
Sarlio-Lähteenkorva, S., Rissanen, A., and Kaprio, J. (2000). A descriptive study of weight loss maintenance: 6 and 15 year
follow-up of initially overweight adults. Int. J. Obes. Relat Metab Disord. 24, 116-125.
Sarna, S. and Kaprio, J. (1980). Use of multiple logistic analysis in twin zygosity diagnosis. Hum. Hered. 30, 71-80.
Sarna, S., Kaprio, J., Sistonen, P., and Koskenvuo, M. (1978). Diagnosis of twin zygosity by mailed questionnaire. Hum.
Hered. 28, 241-254.
Schlosser, E. (2002). Fast food nation: the dark side of the all-American meal. HarperCollins, New York.
Schlundt, D. G., Hill, J. O., Sbrocco, T., Pope-Cordle, J., and Sharp, T. (1992). The role of breakfast in the treatment of
obesity: a randomized clinical trial. Am. J. Clin. Nutr. 55, 645-651.
Schousboe, K., Visscher, P. M., Erbas, B., Kyvik, K. O., Hopper, J. L., Henriksen, J. E., Heitmann, B. L., and Sorensen, T. I.
(2003a). Twin study of genetic and environmental influences on adult body size, shape, and composition. Int. J.
Obes. Relat Metab Disord.
Schousboe, K., Willemsen, G., Kyvik, K. O., Mortensen, J., Boomsma, D. I., Cornes, B., Davis, C. J., Fagnani, C., Hjelmborg,
J., Kaprio, J., De Lange, M., Luciano, M., Martin, N. G., Pedersen, N., Pietiläinen, K. H., Rissanen, A., Saarni, S.,
Sorensen, T. I., Van Baal, G. C. M., and Harris, J. (2003b). Sex differences in heritability of BMI: a comparative
study of results from twin studies in eight countries. Twin Research 6, 409-421.
Schur, E. A., Sanders, M., and Steiner, H. (2000). Body dissatisfaction and dieting in young children. Int. J. Eat. Disord. 27,
Serdula, M. K., Mokdad, A. H., Williamson, D. F., Galuska, D. A., Mendlein, J. M., and Heath, G. W. (1999). Prevalence of
attempting weight loss and strategies for controlling weight. JAMA 282, 1353-1358.
Shaw, M. E. (1998). Adolescent breakfast skipping: an Australian study. Adolescence 33, 851-861.
Shick, S. M., Wing, R. R., Klem, M. L., McGuire, M. T., Hill, J. O., and Seagle, H. (1998). Persons successful at long-term
weight loss and maintenance continue to consume a low-energy, low-fat diet. J. Am. Diet. Assoc. 98, 408-413.
Siega-Riz, A. M., Popkin, B. M., and Carson, T. (1998). Trends in breakfast consumption for children in the United States
from 1965-1991. Am. J. Clin. Nutr. 67, 748S-756S.
Silventoinen, K., Sarlio-Lähteenkorva, S., Koskenvuo, M., Lahelma, E., and Kaprio, J. (2004). Effect of environmental and
genetic factors on education-associated disparities in weight and weight gain: a study of Finnish adult twins. Am. J.
Clin. Nutr. 80, 815-822.
Sjöberg, A., Hallberg, L., Höglund, D., and Hulthen, L. (2003). Meal pattern, food choice, nutrient intake and lifestyle factors
in The Göteborg Adolescence Study. Eur. J. Clin. Nutr. 57, 1569-1578.
Slof, R., Mazzeo, S., and Bulik, C. M. (2003). Characteristics of women with persistent thinness. Obes. Res. 11, 971-977.
Snyder, E. E., Walts, B., Perusse, L., Chagnon, Y. C., Weisnagel, S. J., Rankinen, T., and Bouchard, C. (2004). The human
obesity gene map: the 2003 update. Obes. Res. 12, 369-439.
Sosioekonomisen aseman luokitus, 1989. [The classification of socioeconomic status, 1989; in Finnish.]. (1990). Käsikirjoja
17. Helsinki: Tilastokeskus.
Stice, E. (2001). A prospective test of the dual-pathway model of bulimic pathology: mediating effects of dieting and negative
affect. J. Abnorm. Psychol. 110, 124-135.
Stice, E. and Shaw, H. (2003). Prospective relations of body image, eating, and affective disturbances to smoking onset in
adolescent girls: how Virginia slims. J. Consult Clin. Psychol. 71, 129-135.
Stice, E. and Whitenton, K. (2002). Risk factors for body dissatisfaction in adolescent girls: a longitudinal investigation. Dev.
Psychol. 38, 669-678.
Story, M., Neumark-Sztainer, D., Sherwood, N., Stang, J., and Murray, D. (1998). Dieting status and its relationship to eating
and physical activity behaviors in a representative sample of US adolescents. J. Am. Diet. Assoc. 98, 1127-35, 1255.
Strauss, R. S. (1999). Self-reported weight status and dieting in a cross-sectional sample of young adolescents: National Health
and Nutrition Examination Survey III. Arch. Pediatr. Adolesc. Med. 153, 741-747.
Stunkard, A. J., Sorensen, T., and Schulsinger, F. (1983). Use of the Danish adoption register for the study of obesity and
thinness. In S.S. Kety, L. P. Rowland, R. L. Sidman, & S. W. Mathysse (Eds.), The genetics of neurological and
psychiatric disorders. Raven Press, New York.
Stunkard, A. J. and Wadden, T. A. (1992). Psychological aspects of severe obesity. Am. J. Clin. Nutr. 55, 524S-532S.
Sullivan, P. F., Bulik, C. M., and Kendler, K. S. (1998). Genetic epidemiology of binging and vomiting. Br. J. Psychiatry 173,
Sullivan, P. F. and Eaves, L. J. (2002). Evaluation of analyses of univariate discrete twin data. Behav. Genet. 32, 221-227.
Suviolahti, E., Oksanen, L. J., Öhman, M., Cantor, R. M., Ridderstråle, M., Tuomi, T., Kaprio, J., Rissanen, A., Mustajoki, P.,
Jousilahti, P., Vartiainen, E., Silander, K., Kilpikari, R., Salomaa, V., Groop, L., Kontula, K., Peltonen, L., and
Pajukanta, P. (2003). The SLC6A14 gene shows evidence of association with obesity. J. Clin. Invest 112, 1762-1772.
Telch, C. F. and Agras, W. S. (1994). Obesity, binge eating and psychopathology: are they related? Int. J. Eat. Disord. 15, 53-
Terre, L., Drabman, R. S., and Meydrech, E. F. (1990). Relationships among children's health-related behaviors: a multivariate,
developmental perspective. Prev. Med. 19, 134-146.
Tsai, G., Curbow, B., and Heinberg, L. (2003). Sociocultural and developmental influences on body dissatisfaction and
disordered eating attitudes and behaviors of Asian women. J. Nerv. Ment. Dis. 191, 309-318.
Turnbull, J. D., Heaslip, S., and McLeod, H. A. (2000). Pre-school children's attitudes to fat and normal male and female
stimulus Figures. Int. J. Obes. Relat Metab Disord. 24, 1705-1706.
Tzimis, L. and Kafatos, A. (2000). Drug utilization and nutrition patterns among children from indigent and emigrant families
in Crete, Greece. Public Health 114, 393-397.
Urho, U.-M. and Hasunen, K. (1999). Yläasteen kouluruokailu 1998 [School meals in junior high school in 1998; in Finnish].
Sosiaali- ja terveysministeriön selvityksiä 5.
van den Berg, P., Wertheim, E. H., Thompson, J. K., and Paxton, S. J. (2002). Development of body image, eating
disturbance, and general psychological functioning in adolescent females: a replication using covariance structure
modeling in an Australian sample. Int. J. Eat. Disord. 32, 46-51.
van den Bree, M. B., Eaves, L. J., and Dwyer, J. T. (1999). Genetic and environmental influences on eating patterns of twins
aged >/=50 y. Am. J. Clin. Nutr. 70, 456-465.
Venditti, E. M., Wing, R. R., Jakicic, J. M., Butler, B. A., and Marcus, M. D. (1996). Weight cycling, psychological health, and
binge eating in obese women. J. Consult Clin. Psychol. 64, 400-405.
von Ranson, K. M., Iacono, W. G., and McGue, M. (2002). Disordered eating and substance use in an epidemiological
sample: I. Associations within individuals. Int. J. Eat. Disord. 31, 389-403.
Wadden, T. A. and Stunkard, A. J. (1985). Social and psychological consequences of obesity. Ann. Intern. Med. 103, 1062-
Wade, T., Martin, N. G., Neale, M. C., Tiggemann, M., Treloar, S. A., Bucholz, K. K., Madden, P. A., and Heath, A. C.
(1999). The structure of genetic and environmental risk factors for three measures of disordered eating. Psychol.
Med. 29, 925-934.
Wade, T., Martin, N. G., and Tiggemann, M. (1998). Genetic and environmental risk factors for the weight and shape
concerns characteristic of bulimia nervosa. Psychol. Med. 28, 761-771.
Wade, T. D., Bulik, C. M., Heath, A. C., Martin, N. G., and Eaves, L. J. (2001). The influence of genetic and environmental
factors in estimations of current body size, desired body size, and body dissatisfaction. Twin. Res. 4, 260-265.
Wade, T. D., Wilkinson, J., and Ben Tovim, D. (2003). The genetic epidemiology of body attitudes, the attitudinal component
of body image in women. Psychol. Med. 33, 1395-1405.
Walker, A., Walker, B., Jones, J., & Ncongwane, J. (1982). Breakfast habits of adolescents in four South African populations.
Am.J.Clin.Nutr. 36, 650-656.
Wardle, J., Guthrie, C., Sanderson, S., Birch, L., and Plomin, R. (2001). Food and activity preferences in children of lean and
obese parents. Int. J. Obes. Relat Metab Disord. 25, 971-977.
Wardle, J. and Johnson, F. (2002). Weight and dieting: examining levels of weight concern in British adults. Int. J. Obes. Relat
Metab Disord. 26, 1144-1149.
Weight cycling. (1994). National Task Force on the Prevention and Treatment of Obesity. JAMA 272, 1196-1202.
Williamson, D. F. (1997). Intentional weight loss: patterns in the general population and its association with morbidity and
mortality. Int. J. Obes. Relat Metab Disord. 21 Suppl 1, S14-S19.
Williamson, D. F., Serdula, M. K., Anda, R. F., Levy, A., and Byers, T. (1992). Weight loss attempts in adults: goals, duration,
and rate of weight loss. Am. J. Public Health 82, 1251-1257.
Wolfe, W. S., Campbell, C. C., Frongillo, E. A., Jr., Haas, J. D., and Melnik, T. A. (1994). Overweight schoolchildren in New
York State: prevalence and characteristics. Am. J. Public Health 84, 807-813.
Wonderlich, S., Crosby, R., Mitchell, J., Thompson, K., Redlin, J., Demuth, G., and Smyth, J. (2001). Pathways mediating
sexual abuse and eating disturbance in children. Int. J. Eat. Disord. 29, 270-279.
Wyatt, H. R., Grunwald, G. K., Mosca, C. L., Klem, M. L., Wing, R. R., and Hill, J. O. (2002). Long-term weight loss and
breakfast in subjects in the National Weight Control Registry. Obes. Res. 10, 78-82.
Zuckerman, M. (1979). Sensation seeking: beyond the optimal level of arousal. Lawrence Erlbaum Associates, Publishers,