Associations Between Childhood Socioeconomic Position and Adulthood Obesity

Abstract

Childhood socioeconomic position (SEP) is inversely associated with cardiovascular disease and all-cause mortality. Obesity in adulthood may be a biologic mechanism. Objectives were to systematically review literature published between 1998 and 2008 that examined associations of childhood SEP with adulthood obesity. Five databases (Cochrane Library, MEDLINE, EMBASE, PsycINFO, Web of Science) were searched for studies from any country, in any language. Forty-eight publications based on 30 studies were identified. In age-adjusted analyses, inverse associations were found between childhood SEP and adulthood obesity in 70% (14 of 20) of studies in females and 27% (4 of 15) in males. In studies of females showing inverse associations between childhood SEP and adulthood obesity, typical effect sizes in age-adjusted analyses for the difference in body mass index between the highest and lowest SEP were 1.0-2.0 kg/m(2); for males, effect sizes were typically 0.2-0.5 kg/m(2). Analyses adjusted for age and adult SEP showed inverse associations in 47% (8 of 17) of studies in females and 14% (2 of 14) of studies in males. When other covariates were additionally adjusted for, inverse associations were found in 4 of 12 studies in females and 2 of 8 studies in males; effect sizes were typically reduced compared with analyses adjusted for age only. In summary, the findings suggest that childhood SEP is inversely related to adulthood obesity in females and not associated in males after adjustment for age. Adulthood SEP and other obesity risk factors may be the mechanisms responsible for the observed associations between childhood SEP and adulthood obesity.

Full text

Epidemiologic Reviews
ªThe Author 2009. Published by the Johns Hopkins Bloomberg School of Public Health.
All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.
Vol. 31, 2009
DOI: 10.1093/epirev/mxp006
Advance Access publication July 31, 2009
Associations Between Childhood Socioeconomic Position and Adulthood
Obesity
Laura C. Senese, Nisha D. Almeida, Anne Kittler Fath, Brendan T. Smith, and Eric B. Loucks
Accepted for publication May 8, 2009.
Childhood socioeconomic position (SEP) is inversely associated with cardiovascular disease and all-cause
mortality. Obesity in adulthood may be a biologic mechanism. Objectives were to systematically review literature
published between 1998 and 2008 that examined associations of childhood SEP with adulthood obesity. Five
databases (Cochrane Library, MEDLINE, EMBASE, PsycINFO, Web of Science) were searched for studies from
any country, in any language. Forty-eight publications based on 30 studies were identified. In age-adjusted
analyses, inverse associations were found between childhood SEP and adulthood obesity in 70% (14 of 20) of
studies in females and 27% (4 of 15) in males. In studies of females showing inverse associations between
childhood SEP and adulthood obesity, typical effect sizes in age-adjusted analyses for the difference in body
mass index between the highest and lowest SEP were 1.0–2.0 kg/m
2
; for males, effect sizes were typically 0.2–0.5
kg/m
2
. Analyses adjusted for age and adult SEP showed inverse associations in 47% (8 of 17) of studies in females
and 14% (2 of 14) of studies in males. When other covariates were additionally adjusted for, inverse associations
were found in 4 of 12 studies in females and 2 of 8 studies in males; effect sizes were typically reduced compared
with analyses adjusted for age only. In summary, the findings suggest that childhood SEP is inversely related to
adulthood obesity in females and not associated in males after adjustment for age. Adulthood SEP and other
obesity risk factors may be the mechanisms responsible for the observed associations between childhood SEP
and adulthood obesity.
adult; child; health status disparities; obesity; review; social class; socioeconomic factors
Abbreviations: CI, confidence interval; SEP, socioeconomic position.
INTRODUCTION
There is public health concern about socioeconomic gra-
dients in health (1). In recent decades, a great amount of
research has investigated childhood socioeconomic position
(SEP) and how it may relate to health outcomes in adulthood.
Childhood SEP can be measured in a number of ways, such
as through parents’ education, parents’ occupation, house-
hold income, and household conditions (2, 3). Systematic
reviews have demonstrated reasonable associations between
childhood SEP and increased risk for coronary heart disease,
stroke, and all-cause mortality (4, 5). The mechanisms that
may be responsible for the association between childhood
SEP and chronic disease have been increasingly sought.
Obesity is one of the prime suspects and, consequently over
the past decade, a multitude of articles have been published
on this association. Given the population health concern
about how socioeconomic disadvantage over the life course
may hasten poor health, the burgeoning research linking
childhood SEP to adulthood obesity, and the current epi-
demic and resulting health effects of obesity in nations
around the world, a systematic review on this topic is
needed.
A review on the relation between childhood SEP and
obesity in adulthood was performed by Parsons et al. (6)
on 12 studies published up to the year 1998. They found
consistent inverse associations between childhood SEP and
adulthood obesity in males (8 of 9 studies) and females (4 of
5 studies). Considering that obesity rates have continued to
increase (7, 8), that a substantial number of articles have
been published on the topic since 1998, and that there is
some indication that socioeconomic gradients in obesity
may be flattening in countries such as the United States
where obesity rates have been rising in all socioeconomic
Correspondence to Dr. Eric B. Loucks, Center for Population Health and Clinical Epidemiology, Department of Community Health, Epidemiology
Section, Brown University, 121 South Main Street, Box G-S121-2, Providence, RI (e-mail: eric.loucks@brown.edu).
21 Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

groups (9), it appeared useful to update this review. Further-
more, we explored differences in the associations between
childhood SEP and adulthood obesity according to several
factors that may influence the observed associations, includ-
ing gender, measures of obesity (self-report vs. directly
measured), race/ethnicity, country, age, measure of SEP,
and birth year. There were a number of reasons for selecting
these factors. For example, it has been found that the asso-
ciation between adulthood SEP and obesity differs by gen-
der, with robust inverse gradients found in females and not
in males (10). These patterns have not been reviewed re-
cently in studies focused on how childhood SEP is related to
adulthood obesity. A recent systematic review demonstrated
that positive associations between adulthood SEP and obe-
sity typically exist in developed countries, while in devel-
oping countries, the associations are generally negative (10).
However, the existence of similar trends in the relation be-
tween childhood SEP and adult obesity has not been re-
cently reviewed. SEP measured in childhood showed
stronger associations with disease outcomes compared with
studies that used adulthood recall of childhood SEP, proba-
bly because of reductions in measurement error (11). Con-
sequently, associations may be estimated more accurately
when childhood SEP is measured during childhood rather
than retrospectively recalled during adulthood. However,
this has not yet been reviewed in the context of childhood
SEP and adulthood obesity. Finally, birth year is an approx-
imation of exposure to environments that have become more
obesogenic over recent decades. We summarized associa-
tions between childhood SEP and adulthood obesity through
a systematic review of articles published from January 1998
through September 2008.
METHODS
Search strategy
We performed a comprehensive review of literature ex-
amining the relation between childhood SEP (<19 years of
age) and adulthood obesity (19 years of age). Specific
inclusion criteria were as follows: exposures examining
some aspect of childhood SEP measured in subjects <19
years of age, outcomes capturing some aspect of obesity
measured in subjects 19 years of age, adjustment
for age, and sample size >600 participants. Searches in-
cluded observational and interventional studies, all races/
ethnicities, all geographic locations, and published in any
language between January 1998 and September 2008. Meet-
ing summaries and thesis abstracts were excluded from the
review. Childhood SEP was measured either directly during
childhood or recalled during adulthood; for example, adult
participants were asked to recall the occupation of each
parent when the participants were children. The study fo-
cused on individual-level SEP and excluded studies that
assessed only area-level SEP. This was done in an effort
to keep the review focused on a manageable number of stud-
ies relating specifically to individual-level SEP. In an effort to
ensure that every included study was adequately powered to
detect clinically meaningful differences, we performed
power analyses (a¼0.05, 1 b¼0.80, effect size ¼
1.0-kg/m
2
body mass index) and found that a sample size of
570 was required using a standard deviation of 4.25 kg/m
2
,
and that a sample size of 638 was required using a stan-
dard deviation of 4.50 kg/m
2
; representative standard devi-
ations were obtained from studies included in this review
(12, 13), using the Power procedure in SAS, version 9.2,
software (SAS Institute, Inc., Cary, North Carolina). Exten-
sive searches of 5 databases (MEDLINE, EMBASE, Psyc-
INFO, Cochrane Library, and Web of Science) were
performed with the assistance of a professional librarian.
Exposures in childhood were identified by using the search
term ‘‘socioeconomic status’’ and synonyms (e.g., social
class, socioeconomic position, education, occupation, in-
come, housing conditions, overcrowding, and caste). Out-
comes in adulthood included measures of obesity and
synonyms, such as body mass index, waist circumference,
hip circumference, waist/hip ratio, skinfold thickness, and
bioelectrical impedance. The complete search strategy is
provided in Web Tables 1–4. (This information is described
in 4 supplementary tables posted on the Journal’s website
(http://aje.oxfordjournals.org/).) The tables of contents of
journals that published the most on this topic, identified by
using the Scopus database (American Journal of Clinical
Nutrition,International Journal of Obesity,European Jour-
nal of Clinical Nutrition,Journal of Nutrition, and Social
Science and Medicine), were searched manually for addi-
tional publications. Six experts were contacted about addi-
tional published or unpublished work, of whom 4 responded
with feedback. The bibliographies of all manuscripts meeting
inclusion criteria were scanned for any additional pertinent
publications.
The 6,609 papers identified were initially assessed by 1
investigator (L. C. S.). If the title or abstract suggested that
associations between childhood SEP and adult obesity may
have been examined, the full-text article was retrieved.
Upon retrieval of the full-text article, any publication that
investigated an association between childhood SEP and
adult obesity was kept for full-text article review and inde-
pendent secondary assessment by 2 investigators (L. C. S.
and E. B. L.). If there was any doubt about whether or not
the publication assessed the association between childhood
SEP and adulthood obesity, it was saved for the independent
secondary assessment by the 2 investigators (L. C. S. and
E. B. L.). The few discrepancies between researchers were
resolved by consensus. None of the papers examined in this
review overlapped with those in the review published by
Parsons et al. (6) in 1999.
Tabulation and analytical approach
Descriptive information on each study sample and publi-
cation included in this review was summarized (Table 1).
Gender-specific and gender-pooled trends were analyzed by
using a quantitative tally approach in which study samples,
rather than individual papers, were the units of analysis.
This was done in an attempt to eliminate the bias imposed
by studies that were the subject of numerous publications.
Meta-analysis was not appropriate for this review because of
the heterogeneity of exposures and outcomes (14). Trends
were assessed on the basis of 2 measures of effect: effect
22 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

size and statistical significance. Representative effect sizes
for each publication were reported in tabular form (Table 2).
Findings were stated as statistically significant if the statis-
tical test provided a Pvalue of <0.05 or if the 95% confi-
dence intervals for the difference between high and low SEP
did not encompass the point estimate for the reference cat-
egory. Two studies (13, 15) required additional statistical
analyses to determine statistical significance for the associ-
ation of childhood SEP with adult obesity, using Student
ttests to incorporate point estimates and variances reported
in studies (15). In summarizing each study to determine
whether the overall associations between each measure of
SEP and every measure of obesity were statistically signif-
icant, we calculated the proportion of statistically significant
(P<0.05) associations reported in each study (e.g., for
multiple outcomes such as body mass index, waist circum-
ference, and waist/hip ratio, and/or for multiple exposures
such as mother’s education and father’s education). If the
proportion of statistically significant reported observations
in the studies was >0.6, it was considered to show overall
statistically significant associations. If the proportion was
0.4 or 0.6, the study was concluded to show inconsistent
findings. If the proportion was <0.4, the overall study find-
ings were considered not to be statistically significant. The
proportions of significant associations in 4 categories (in-
verse association, direct association, inconsistent associa-
tion, and no association) were summed for 3 possible
analytical approaches: adjusted for age (Table 3), adjusted
for age and adult SEP (Table 4), and adjusted for age, adult
SEP, and other potential obesity risk factors (such as smok-
ing, physical activity, parity, cognitive function, birth
weight, and so on) (Table 5). All health behavior (e.g.,
smoking and physical activity) measurements were obtained
concurrently with obesity measurements; other measures,
such as parity, birth weight, and cognitive function, were
measured either earlier in the life course or concurrently
with obesity.
For longitudinal study publications that reported sequen-
tial results for multiple ages, we presented results only at the
oldest age. Furthermore, as results suggested that there were
gender differences in the associations between childhood
SEP and adulthood obesity, we reported only the gender-
specific findings for summary tables (Tables 3–5) in studies
that presented both gender-specific and gender-pooled find-
ings (described in Table 2).
RESULTS
Overall, 48 publications meeting the inclusion criteria
were identified. These separate analyses drew upon data
from 30 distinct studies carried out in Europe, North Amer-
ica, Australia, New Zealand, and China (Table 1). All of the
studies used observational designs. Although most study
participants were middle aged when obesity was assessed,
birth years ranged from 1892 to 1985 (Table 1). All analyses
were either age adjusted or stratified by discrete age inter-
vals no larger than 10 years. Parental occupation was used as
a measure of childhood SEP in all but 4 studies (15–18).
Father’s occupation was examined in 18 of these 26 studies
(12, 19–49) (Table 1). Other childhood SEP measures
(including parental education level and familial economic
distress) are shown in Table 1. Obesity measures included
body mass index, percentage body fat, weight change, waist/
hip ratio, and waist circumference (Table 1).
In age-adjusted analyses, 70% (14 of 20) of studies of
females demonstrated significant inverse associations be-
tween childhood SEP and adulthood obesity (Table 3). In
contrast, only 27% (4 of 15) of age-adjusted studies in males
showed significant inverse associations (Table 3). Effect
sizes are shown in Table 2. As an example of effect size
in studies of females that showed an inverse association
between childhood SEP and adulthood obesity, typical ef-
fect sizes in age-adjusted analyses for change in body mass
index between highest and lowest SEP were in the range of
1–2 kg/m
2
(exact point estimates and variances shown in
Table 2) (12, 13, 23, 30, 33, 37). In other words, for a female
who is 163 cm (approximately 5 feet, 4 inches) in height,
this would be a difference of 2.7–5.3 kg between the highest
and lowest levels of SEP. For males, the effect sizes were
typically smaller: 0.2–0.5 kg/m
2
for the majority of studies
that used body mass index as an outcome (23, 29, 50, 51),
although 1 study showed a difference of 1.8 kg/m
2
between
the highest and lowest SEP (30) (exact point estimates and
variances shown in Table 2). Other measures of effect, such
as odds ratios, for obesity showed similar differences in
effect sizes between males and females (Table 2). Studies
typically showed somewhat of a decrease in effect sizes
after further adjustment for adulthood SEP (Tables 2 and
4). For example, in the Danish Longitudinal Study on Work,
Employment, and Health, the age-adjusted odds ratio of
obesity for manual compared with nonmanual father’s oc-
cupation was 1.5 (95% confidence interval (CI): 1.1, 2.0)
among females. After additional adjustment for adult SEP,
the odds ratio was reduced to 1.3 (95% CI: 0.9, 1.8) (Table 2)
(44). In females, 47% (8 of 17) of studies showed statisti-
cally significant inverse associations when age and adult
SEP were adjusted for, while in males only 14% (2 of 14)
of studies showed significant inverse associations. Statistical
adjustment for potential obesity risk factors other than age
and adulthood SEP was made in 22 studies, considering
studies of males, females, and gender-pooled studies sepa-
rately (Tables 1, 2, and 5). Effect sizes were typically
slightly further reduced after additional adjustments
(examples of effect sizes described in Table 2). For example,
in the 1958 British birth cohort, the odds for obesity in fe-
males for higher versus lower parental occupation were 1.28
(95% CI: 1.14, 1.43) after adjustment for age and adulthood
SEP and 1.23 (95% CI: 1.10, 1.38) after further adjustment
for parental body mass index (45). Thirty-three percent (4 of
12) of studies (23, 39, 45, 52) of females showed significant
inverse associations between childhood SEP and obesity
after adjustment for age, adulthood SEP, and other potential
risk factors for obesity, while 2 of 8 studies (26, 45) of males
demonstrated significant inverse associations (Table 5).
The large range in sample size, from 603 to 100,330, did
not likely bias results, as the variables that may induce effect
modification on the association between childhood SEP and
adulthood obesity were fairly evenly distributed with
respect to sample size. For example, with a sample size
cutpoint of 2,000, 13 studies among females were based
Child Socioeconomic Position and Adult Obesity 23
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 1. Summary of Study Characteristics
a
Study, Authors,
Year of Publication
(Reference)
No. Gender SEP Measure Adiposity
Measure Statistical Adjustments
Australian Longitudinal
Study on Women’s
Health, 1972–1978
a
,
Australia
Ball and Mishra,
2006 (13)
8,756 Female Highest educational level and
occupation of mother, father,
or main caregiver
BMI Adult SEP (education and occupation),
smoking, alcohol, physical activity,
area-of-residence parity, marital status
b
Whitehall II Longitudinal
Study, 1929–1953,
United Kingdom
Brunner et al.,
1999 (19)
4,774 Male Father’s occupation BMI, WC, WHR Age, adult SEP (occupation)
2,206 Female
Heraclides et al.,
2008 (23)
3,364 Male Father’s occupation BMI, WHR Age, adult SEP (occupation and
education), diet, alcohol consumption,
physical activity, smoking status
1,234 Female
1950s Aberdeen Children’s
Study, 1950–1956,
United Kingdom
Dundas et al.,
2006 (21)
3,387 Male Father’s occupation BMI Age, sex, adult SEP (occupation and
education), IQ test scores in
childhood, smoking
3,708 Female
Lawlor et al.,
2005 (88)
7,183 Male and
female
Father’s occupation BMI Age, gender, adult SEP (occupation,
income, educational attainment),
family size, birth order, birth weight,
childhood height, childhood BMI,
cognitive function at ages 7 and 11,
parental smoking
Pierce and Leon,
2005 (37)
2,968 Female Father’s occupation BMI None
c
The British Women’s Heart
and Health Study, 1919–
1941, United Kingdom
Ebrahim et al.,
2004 (22)
2,936 Female Father’s occupation BMI Age, adult SEP (occupation of head
of household)
Lawlor et al.,
2002 (32)
3,444 Female Father’s occupation BMI, WHR Age, adult SEP (occupation of head of
household)
Lawlor et al.,
2004 (35)
3,444 Female Father’s occupation BMI, WHR Age, adult SEP (occupation of head
of household)
Lawlor et al.,
2005 (33)
4,286 Female Father’s occupation, bathroom in
childhood home, hot water supply
in childhood home, family access
to car and bedroom sharing in
childhood
BMI, WHR Age
Lawlor et al.,
2007 (34)
3,869 Female Father’s occupation and childhood
household amenities (bathroom,
hot water, car, and bedroom sharing)
BMI, WHR Age, diabetes
24 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

The Medical Research
Council National
Survey of Health and
Development, 1946,
United Kingdom
Hardy et al.,
2000 (20)
2,659 Male and
female
Father’s occupation BMI Sex, adult SEP (education),
childhood relative body weights
d
Kuh et al.,
2002 (26)
1,314 Male Father’s occupation WC, WHR Adult SEP (occupation of head
of household), birth weight,
childhood relative body weight,
adult BMI
d
1,318 Female
Langenberg et al.,
2003 (30)
1,299 Male Father’s occupation BMI, WC, WHR Adult SEP (occupation of the head
of the household)
d
1,337 Female
Langenberg et al.,
2006 (31)
1,311 Male Father’s occupation WC Adult SEP (occupation and highest
level of education)
d
1,318 Female
Mishra et al.,
2003 (36)
2,980 Male and
female
Father’s occupation WC Sex, adult SEP (occupation), vitamin E
intake, childhood micronutrient intake,
adult micronutrient intake, region of birth,
smoking status, physical activity, adult
BMI, polyunsaturated fatty acid index
d
Power et al.,
2005 (44)
3,266 Male and
female
Father’s occupation BMI Adult SEP (occupation)
d
1970–1973 Scottish
cohort of workers,
1905–1938, Scotland
Heslop et al.,
2001 (24)
958 Female Father’s occupation BMI Age
16-Year Finnish cohort,
~1967, Finland
Huurre et al.,
2003 (12)
661 Male Father’s occupation (or mother’s
occupation if father’s is missing)
BMI Adult SEP (education and occupation)
d
796 Female
Pitt County Study,
1937–1963,
United States
James et al.,
2006 (57)
679 Female Primary earner’s occupation BMI Age, adult SEP (education, occupation,
employment status, home owning),
marital status, alcohol, smoking, childhood
food insecurity, fruit/vegetable consumption,
strenuous exercise
Bennett et al.,
2007 (54)
751 Female Primary earner’s occupation and
childhood household conditions
(public assistance, no plumbing,
no electricity, food scarcity)
BMI, BMI change Age, marital status, cigarette smoking,
strenuous physical activity, adult SEP
(occupation, employment status, homeowner)
Johns Hopkins Precursors
Study—medical students
from the classes of
1948–1964, ~1921–1938,
United States
Kittleson et al.,
2006 (25)
1,131 Male Father’s occupation BMI Adult SEP (all were physicians,
so shared similar education and
occupation SEP categories)
c
Table continues
Child Socioeconomic Position and Adult Obesity 25
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 1. Continued
Study, Authors,
Year of Publication
(Reference)
No. Gender SEP Measure Adiposity
Measure Statistical Adjustments
Cardiovascular Risk in Young
Finns Study, 1961–1977,
Finland
Kivimaki et al.,
2006 (60)
767 Male Parental occupation BMI, WC, WHR Age, adult SEP (occupation)
992 Female
Polish conscription samples
in 1965, 1986, 1995, and
2001, 1946–1982, Poland
Bielicki et al.,
2000 (59)
21,130 (1965),
29,275 (1986),
30,940 (1995)
Male Father’s education BMI Cohort year and degree of urbanization
of childhood residence
d
Koziel et al.,
2004 (51)
26,440 (1986),
26,245 (1995),
26,528 (2001)
Male Parental education and
occupation
BMI Cohort year, degree of urbanization of
childhood residence, and number
of siblings
d
Koziel et al.,
2006 (50)
4,689 (1986),
3,661 (1995),
2,869 (2001)
Male Parental education and
occupation
BMI None
d
Kaiser Permanente Women
Twins Study, 1892–
1961, United States
Krieger et al.,
2001 (53)
630 Female Occupational class of
head of household
BMI Age, adult SEP (occupation),
marital status, race/ethnicity,
number of siblings
Survey of employed
middle-aged men and
women in Helsinki,
1940–1961, Finland
Laaksonen et al.,
2004 (18)
1,252 Male Parental education level,
self-reported economic
difficulties in childhood
BMI Age, adult SEP (education
and occupation)
4,975 Female
Malmo
¨Diet and Cancer
Prospective Cohort
Study, 1923–1950,
Sweden
Lahmann et al.,
2000 (52)
5,464 Female Parental occupation for
primary wage earner
% Body fat, WC, WHR Age, adult SEP (education, occupation,
employment status), menopausal
status, hormonal therapy, parity,
age at menarche, self-rated health,
smoking status, physical activity,
alcohol intake, past change in diet,
ethnicity, living arrangement
1966 Northern Finland
birth cohort, 1966,
Finland
Laitinen et al.,
2001 (29)
2,876 Male Father’s occupation
(or mother’s occupation,
if missing father’s occupation)
BMI, WC, WHR Maternal age and maternal BMI
d
3,404 Female
26 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Laitinen et al.,
2002 (28)
2,536 Male Father’s occupation (or mother’s
occupation, if missing
father’s occupation)
BMI Unemployment history, BMI at
14 years, school performance at
16 years, area of residence,
whether adult BMI was self-reported
or directly measured, number of
own children
d
2,766 Female
Laitinen et al.,
2004 (27)
2,840 Male Father’s occupation WHR None
d
2,930 Female
Glasgow Alumni Cohort,
~1925–1945,
United Kingdom
Okasha et al.,
2003 (42)
683 Male Father’s occupation BMI Age
Dunedin Multidisciplinary
Health and Development
Study, 1972–1973,
New Zealand
Poulton et al.,
2002 (56)
931 Male and
female
Parental occupation BMI, WHR Sex, adult SEP (occupation),
infant health
d
1958 British birth cohort,
1958, United Kingdom
Power et al.,
2003 (45)
3,267 Male Father’s occupation BMI Adult SEP (education and
occupation) and parental BMI
d
3,808 Female
Power et al.,
2005 (44)
9,350 Male and
female
Father’s occupation BMI Adult SEP (occupation)
d
Power et al.,
2007 (43)
4,665 Male Father’s occupation BMI Sex, adult SEP (occupation)
d
4,712 Female
Thomas et al.,
2007 (48)
7,468 Male and
female
Father’s occupation BMI None
d
Spanish sampling from
census, ~1920–1940,
Spain
Regidor et al.,
2004 (47)
1,477 Male Father’s occupation BMI, WC Age, adult SEP (occupation)
1,755 Female
Regidor et al.,
2004 (46)
1,514 Male Father’s occupation BMI, WC Age, adult SEP (occupation)
1,769 Female
GLOBE Study, 1916–1966,
Netherlands
van de Mheen
et al., 1998 (49)
13,854 Male and
female
Father’s occupation BMI Age, sex, adult SEP (occupation),
marital status, religious affiliation,
and degree of urbanization
Power et al.,
2005 (44)
5,057 Male and
female
Father’s occupation BMI Age, adult SEP (occupation)
Giskes et al.,
2008 (39)
615 Male Father’s occupation BMI, BMI change Age, adult SEP (occupation),
smoking status, presence of one
or more chronic health conditions,
baseline BMI (for BMI change only)
695 Female
Table continues
Child Socioeconomic Position and Adult Obesity 27
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 1. Continued
Study, Authors,
Year of Publication
(Reference)
No. Gender SEP Measure Adiposity
Measure Statistical Adjustments
Harvard Study of Moods
and Cycles, 1949–1961,
United States
Wise et al.,
2002 (15)
603 Female Economic distress in childhood BMI Adult SEP (economic distress)
c
Kuopio Ischemic Heart
Disease Risk Factor
Study, 1930–1947,
Finland
Power et al.,
2005 (44)
1,918 Male Father’s occupation BMI Age, adult SEP (occupation)
Swedish Survey of Living
Conditions, 1930–1959,
Sweden
Power et al.,
2005 (44)
9,218 Male and
female
Father’s occupation BMI Age, adult SEP (occupation)
Danish Longitudinal Study
on Work, Unemployment,
and Health, 1950–1960,
Denmark
Power et al.,
2005 (44)
6,143 Male and
female
Father’s occupation BMI Age, adult SEP (occupation)
Alameda County Study,
1910–1948,
United States
Power et al.,
2005 (44)
2,727 Male and
female
Father’s occupation BMI Age, adult SEP (occupation)
Baltrus et al.,
2007 (38)
608 Male Father’s occupation Weight gain Age, race, height, weight at
baseline, adult SEP (education,
income, and occupation)
712 Female
Survey of female Polish
students, 1981–1985,
Poland
Wronka and
Pawlinska-
Chmara,
2007 (58)
783 Female Parents’ education, occupation,
and economic status in childhood
BMI Degree of urbanization of childhood
residence, mother’s and father’s
educations and occupations,
childhood economic status, number
of siblings, type of childcare received
b
Nurses’ Health Study,
1920–1946,
United States
Lidfeldt et al.,
2007 (41)
100,330 Female Father’s occupation BMI Age, adult SEP (all were nurses,
so shared similar educational
and occupational SEP categories)
28 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

on smaller samples (n<2,000), while 12 used larger sam-
ples (n2,000). Among males, there were 11 smaller
(n<2,000) and 9 larger (n2,000) sample sizes.
Measures of obesity
Associations between childhood SEP and adulthood obe-
sity did not differ markedly on the basis of whether obesity
was measured directly (e.g., directly measured height,
weight, waist circumference, hip circumference, skinfold
thickness) or self-reported (e.g., self-reported height and
weight). Specifically, in age-adjusted analyses for females,
significant inverse associations were found in 8 of 12 studies
that measured obesity directly (19, 23, 26, 27, 29–31, 40,
44–47, 52, 53) and 6 of 8 studies that used self-reported
obesity (12, 13, 18, 37, 44). Among males, 2 of 7 studies
that used self-reported data reported a significant inverse
association (42, 44), while 2 of 8 studies that measured
obesity directly reported such a finding (26, 30, 31, 44,
45). Similar findings were observed for studies that addi-
tionally adjusted for adulthood SEP and other obesity risk
factors.
Importance of prospectively assessed SEP versus
retrospectively recalled SEP
Among both males and females in the age-adjusted stud-
ies, the likelihood of finding a significant inverse association
between childhood SEP and adult obesity was greatest when
SEP was assessed during childhood: 40% (2 of 5) of studies
in males (26, 30, 31, 44, 45) and 100% (6 of 6) of studies in
females (12, 26, 27, 29–31, 37, 40, 44, 45). The proportion
of significant inverse findings when retrospectively recalled
data were used was 22% (2 of 9) of studies of males (42, 44)
and 57% (8 of 14) of studies of females (13, 18, 19, 23, 44,
46, 47, 52, 53) in age-adjusted analyses. As an indication of
whether study sample size (and resulting statistical power)
had an effect on these observations, among studies using
retrospectively recalled childhood SEP, 12 had sample sizes
of <2,000, while 10 had samples of 2,000. For studies
where SEP was assessed during childhood, 5 studies had
smaller sample sizes (n<2,000), and 3 had larger sample
sizes (n2,000). This would suggest that sample size and
the resulting statistical power were not major explanatory
factors in explaining why associations between childhood
SEP and obesity were stronger in studies that measured SEP
during childhood, rather than retrospectively recalled during
adulthood.
Geography
Despite the very broad search strategy used, the majority
of studies that met all of the inclusion criteria were limited
to only a few geographic regions. Most studies were con-
ducted in developed countries, largely Europe and the
United States (Table 1). For example, in age-adjusted anal-
yses, the studies that reported no associations among fe-
males were conducted in the United Kingdom, the United
States, and the Netherlands (24, 38, 44, 54); significant
inverse associations were found in Australia, the United
Health and Retirement Study,
1951, United States
Best et al.,
2005 (17)
7,913 Male Parents’ education and
family financial status
BMI Age, race/ethnicity, mother’s education,
father’s education, childhood health,
family’s financial status, own education
10,319 Female
Midspan Family Study,
1936–1966,
United Kingdom
Hart et al.,
2008 (40)
1,040 Male Father’s occupation BMI, WC Adult SEP (occupation)
b
1,298 Female
Guangzhou Biobank Cohort
Study, 1910–1956, China
Schooling et al.,
2008 (16)
2,735 Male Parental possession of a
watch, a sewing
machine, or a bicycle
WC Height, smoking status, alcohol
drinking status, physical activity,
use of appropriate medication,
adult SEP (occupation and education)
b
7,011 Female
Abbreviations: BMI, body mass index; GLOBE, Gemifloxacin Long-term Outcomes in Bronchitis Exacerbations; IQ, intelligence quotient; SEP, socioeconomic position; WC, waist
circumference; WHR, waist/hip ratio.
a
Range of years after name of study represents birth range.
b
All participants were born within a 5-year age block or were examined in 5-year age blocks; consequently, age should not be an important confounder.
c
All participants were born within a 5- to 10-year age block; consequently, residual confounding by age is a moderate possibility.
d
All participants were born within 1 year and assessed at the same ages; consequently, age adjustment was not needed.
Child Socioeconomic Position and Adult Obesity 29
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Summary of Associations Between Childhood SEP and Adulthood Obesity After Adjustment for Age, Adult SEP, and Other Potential Obesity Risk Factors
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Australian Longitudinal
Study on Women’s
Health
Ball and Mishra,
2006 (13)
Inverse association for
father’s education
(P<0.0001), father’s
occupation (P<0.0001),
mother’s education
(P<0.0001), and
mother’s occupation
(P¼0.0002) with BMI.
Inverse association for
father’s education
(P¼0.01 and P<0.0001)
and father’s occupation
(P¼0.002 and P<0.0001)
with BMI. No association for
mother’s education (P¼
0.08 and P¼0.06) or
mother’s occupation
(P¼0.36 and P¼0.08)
with BMI.
Inverse association for
father’s occupation
(P¼0.0003) with BMI; no
association for father’s
(P¼0.07) or mother’s (P
‘‘nonsignificant’’) education
with BMI.
Inverse association
(female).
Inconsistent
association (female).
No association
(female).
For example, mean BMI
by level of father’s
education: 10
years ¼24.1 (SD, 4.9)
kg/m
2
; university
or higher degree ¼22.9
(SD, 4.2) kg/m
2
.
For example, BMI for father’s
education nonuniversity ¼
24.3 (SD, 5.2) kg/m
2
,
father’s education university/
higher degree ¼23.0 (SD,
4.3) kg/m
2
, in participants
who did not themselves have
a university degree
(P<0.0001).
For example, BMI for father’s
education 10 years ¼25.3
(95% CI: 24.9, 25.7) kg/m
2
;
university or higher
degree ¼24.8 (95%
CI: 24.3, 25.3) kg/m
2
.
Whitehall II Longitudinal
Study, summary of
2 papers listed below
Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Brunner et al.,
1999 (19)
Females: inverse association
for childhood SEP with
BMI (P<0.001), WHR
(P<0.005), and WC
(P<0.001).
Females: inverse association
for childhood SEP with BMI
(P<0.001) and WC
(P¼0.002); no association
with WHR (P¼0.30).
Not available. Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Not available.
Males: inverse association
for childhood SEP with
BMI (P<0.001); no
association with WHR
(P¼0.34) or WC
(P¼0.06).
Males: inverse association for
childhood SEP with BMI
(P<0.001); no association
with WHR (P>0.5) or WC
(P¼0.10).
For example, the WC
difference between lowest
and highest father’s social
class among females ¼4.54
(SE, 1.03) cm; among
males ¼1.02 (SE, 0.54) cm.
For example, the WC
difference between lowest
and highest father’s social
class among females ¼3.32
(SE, 1.08) cm; among
males ¼0.90 (SE, 0.54) cm.
Heraclides et al.,
2008 (23)
Females: inverse association
for childhood SEP with
BMI (P<0.001) and
WHR (P<0.002).
Females: inverse association
for childhood SEP with BMI
(P¼0.001) and WHR
(P¼0.039).
Females: inverse association
for childhood SEP with BMI
(P¼0.002) and WHR
(P¼0.043).
Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Males: no association for
childhood SEP with BMI
(P¼0.15) or WHR
(P¼0.23).
Males: no association for
childhood SEP with BMI
(P¼0.29) or WHR
(P¼0.94).
Males: no association for
childhood SEP with BMI
(P¼0.38) or WHR
(P¼0.90).
30 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

For example, the mean
difference in BMI between
father’s social class I and
class IV/V ¼2.04 (95%
CI: 0.90, 3.18) kg/m
2
in
females and 0.48 (95% CI:
0.07, 1.04) kg/m
2
in males.
For example, the mean
difference in BMI between
father’s social class I and
class IV/V ¼1.62 (95% CI:
0.41, 2.84) kg/m
2
in females
and 0.43 (95% CI: 0.10,
1.12) kg/m
2
in males.
For example, the mean
difference in BMI between
father’s social class I and
class IV/V ¼1.62 (95% CI:
0.37, 2.80) kg/m
2
in females
and 0.37 (95% CI: 0.19,
0.93) kg/m
2
in males.
1950s Aberdeen Children’s
Study, summary of
3 papers listed below
Inverse association
(female).
b
Inverse association
(male and female).
Inconsistent
association (male
and female).
Dundas et al.,
2006 (21)
Not available. Not available. Inconsistent association for
childhood SEP with BMI.
Not available. Not available. Inconsistent
association (male
and female).
For example, the odds ratio for
obesity for father’s social
class IV vs. class I/II ¼1.56
(95% CI: 1.14, 2.13); for
father’s social class V vs.
I/II ¼1.31 (95% CI: 0.95,
1.79).
Lawlor et al.,
2005 (88)
Inverse association for
childhood SEP with BMI
(P<0.001) and overweight/
obese (BMI, 25 kg/m
2
)
(P<0.001).
Inverse association for
childhood SEP with
overweight (BMI,
25 kg/m
2
).
Inverse association for
childhood SEP with
overweight (BMI,
25 kg/m
2
).
Inverse association
(male and female).
Inverse association
(male and female).
Inverse association
(male and female).
For example, the mean BMI
of adults whose father was
in professional/managerial
class ¼25.4 (SD, 4.1) kg/m
2
;
unskilled manual ¼27.1
(SD, 5.4) kg/m
2
.
For example, the odds ratio
for association of overweight
with manual vs. nonmanual
father’s social class ¼1.27
(95% CI: 1.12, 1.43).
For example, the odds ratio
for association of overweight
with manual vs. nonmanual
father’s social class ¼1.15
(95% CI: 1.00, 1.33).
Pierce and Leon,
2005 (37)
Inverse association for
childhood SEP with BMI
(P<0.001).
Not available. Not available. Inverse association
(female).
Not available. Not available.
For example, the mean BMI
of father’s social class
I¼25.0 (SD, 4.4) kg/m
2
;
father’s social class
V¼27.0 (SD, 6.0) kg/m
2
.
The British Women’s Heart
and Health Study,
summary of 5 papers
listed below
Inconsistent association
(female).
Inconsistent
association (female).
Not available.
Ebrahim et al.,
2004 (22)
Not available. Inconsistent association for
childhood SEP with BMI.
Not available. Not available. Inconsistent
association (female).
Not available.
For example, the prevalence
of obesity among childhood
nonmanual/adult nonmanual
vs. childhood manual/adult
nonmanual ¼15.2% (95%
CI: 12.6, 18.3) vs. 24.0%
(95% CI: 20.6, 27.6). The
prevalence of obesity among
childhood nonmanual/adult
manual vs. childhood
manual/adult manual ¼
23.2% (95% CI: 18.5, 28.7)
vs. 28.6% (95% CI: 25.9,
31.4).
Table continues
Child Socioeconomic Position and Adult Obesity 31
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Lawlor et al.,
2002 (32)
Inverse association for
childhood SEP with BMI;
no association with WHR.
Inverse association for
childhood SEP with BMI; no
association with WHR.
Not available. Inconsistent association
(female).
Inconsistent
association (female).
Not available.
For example, the difference in
BMI per increase in social
class grade (I–V) ¼0.04
(95% CI: 0.02, 0.06) kg/m
2
;
the difference in WHR ¼
0.000 (95% CI: 0.000,
0.001).
For example, the difference in
BMI per increase in social
class grade (I–V) ¼0.03
(95% CI: 0.01, 0.05) kg/m
2
;
the difference in WHR ¼
0.000 (95% CI: 0.000,
0.001).
Lawlor et al.,
2004 (35)
Not available. No association for childhood
SEP with obesity.
Not available. Not available. No association
(female).
Not available.
For example, the odds ratio
of obesity for manual
childhood/nonmanual
adulthood vs. nonmanual
childhood/nonmanual
adulthood ¼1.18 (95%
CI: 0.95, 2.03).
Lawlor et al.,
2005 (33)
Inverse association
for father’s occupation,
bathroom in childhood home,
hot water supply in childhood
home, family access to car,
and bedroom sharing in
childhood with BMI. Inverse
association for bathroom in
childhood home with WHR.
No association for father’s
occupation, hot water supply
in childhood home, family
access to car, and bedroom
sharing in childhood with
WHR.
Not available. Not available. Inconsistent association
(female).
Not available. Not available.
For example, the difference in
mean BMI for females
without vs. with a hot water
supply in the childhood
home ¼1.4 (95%
CI: 0.7, 2.1) kg/m
2
.
Lawlor et al.,
2007 (34)
Inverse association for
childhood SEP with BMI; no
association with WHR.
Not available. Not available. Inconsistent association
(female).
Not available. Not available.
For example, the mean
difference in BMI per greater
number of adverse
childhood SEP
indicators ¼0.28 (95%
CI: 0.17, 0.38) kg/m
2
.
32 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

The Medical Research
Council National
Survey of Health and
Development—1946
Birth Cohort, summary
of 6 papers listed below
Inverse association
(female).
Inverse association
(male).
Inverse association
(female).
b
Inconsistent
association (male).
b
No association
(female).
b
Inverse association
(male).
b
Hardy et al.,
2000 (20)
Not available. Not available. Inverse association for
childhood SEP with BMI.
Not available. Not available. Inverse association
(male and female).
For example, the mean
difference in BMI for
father’s occupation manual
vs. nonmanual ¼0.25 (SD,
0.09) kg/m
2
(P<0.01).
Kuh et al.,
2002 (26)
Females: inverse association
for childhood SEP with
WHR (P<0.001) and
WC (P<0.001).
Not available. Females: no association for
childhood SEP with WHR
(P¼0.06) or WC (P¼0.30).
Inverse association
(female).
Inverse association
(male).
Not available. No association
(female).
Inverse association
(male).
Males: inverse association
for childhood SEP with
WHR (P<0.001) and WC
(P<0.001).
Males: inverse association
for childhood SEP with WHR
(P¼0.05) and WC
(P¼0.02).
For example, the regression
coefficient for father’s
occupation manual vs.
nonmanual in relation to
WHR ¼3.26 (95% CI: 2.41,
4.11) in females and 1.80
(95% CI: 1.16, 2.44)
in males.
For example, the regression
coefficient for father’s
occupation manual vs.
nonmanual in relation to
WHR ¼0.63 (95% CI:
0.01, 1.27) in females and
0.56 (95% CI: 0.02, 1.10)
in males.
Langenberg
et al., 2003 (30)
Females: inverse association
for childhood SEP with
WHR (P
trend
<0.001),
WC (P
trend
<0.001),
and BMI (P
trend
<0.001).
Females: inverse association
for childhood SEP with WHR
(P
trend
¼0.02), WC
(P
trend
¼0.003), and BMI
(P
trend
<0.001).
Not available. Inverse association
(female).
Inverse association
(male).
Inverse association
(female).
Inverse association
(male).
Not available.
Males: inverse association
for childhood SEP with
WHR (P
trend
¼0.001),
WC (P
trend
¼0.04), and
BMI (P
trend
<0.001).
Males: inverse association
for childhood SEP with
WHR (P
trend
¼0.01) and
BMI (P
trend
¼0.001); no
association with WC
(P
trend
¼0.06).
For example, the difference
in mean BMI for father’s
occupation class V vs. I ¼
1.3 (95% CI: 0.4, 2.9)
kg/m
2
in females and 1.8
(95% CI: 0.5, 3.0) kg/m
2
in males.
For example, the difference
in mean BMI for father’s
occupation class V vs.
I¼0.6 (95% CI: 1.1, 2.3)
kg/m
2
in females and 1.7
(95% CI: 0.4, 3.0) kg/m
2
in males.
Langenberg
et al., 2006 (31)
Females: inverse association
for childhood SEP with WC.
Females: inverse association
for childhood SEP with WC.
Not available. Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with WC.
Males: no association for
childhood SEP with WC.
For example, the odds ratio
of elevated WC in high vs.
low father’s occupation ¼
2.4 (95% CI: 1.6, 3.5) in
females and 1.4
(95% CI: 0.9, 2.1)
in males.
For example, the odds ratio
of elevated WC in high vs.
low father’s occupation ¼1.8
(95% CI: 1.1, 2.8) in females
and 1.1 (95% CI: 0.7, 1.7)
in males.
Table continues
Child Socioeconomic Position and Adult Obesity 33
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Mishra et al.,
2003 (36)
Not available. Inverse association for
childhood SEP with WC.
Inverse association for
childhood SEP with WC.
Not available. Inverse association
(male and female).
Inverse association
(male and female).
For example, increased odds
of high WC among childhood
manual/adult nonmanual
compared with childhood
nonmanual/adult nonmanual
individuals; odds ratio ¼1.55
(95% CI: 1.18, 2.03).
For example, increased odds
of high WC among childhood
manual/adult nonmanual
compared with childhood
nonmanual/adult nonmanual
individuals; odds ratio ¼1.45
(95% CI: 1.09, 1.93).
Power et al.,
2005 (44)
Females: inverse association
for childhood SEP with
obesity.
Females: inverse association
for childhood SEP with
obesity.
Not available. Inverse association (female).
No association (male).
Inverse association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with obesity.
Males: no association for
childhood SEP with obesity.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼2.9 (95% CI:
1.8, 4.8) in females and 1.4
(95% CI: 0.9, 2.2) in males.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼2.5 (95% CI:
1.5, 4.0) in females and 1.3
(95% CI: 0.8, 2.3) in males.
1970–1973 Scottish
cohort of workers
Heslop et al.,
2001 (24)
No association for childhood
SEP with BMI.
Not available. Not available. No association (female). Not available. Not available.
For example, the mean BMI
for manual father’s social
class ¼24.6 kg/m
2
and for
nonmanual father’s social
class ¼24.7 kg/m
2
.
16-Year Finnish cohort
Huurre et al.,
2003 (12)
Females: inverse association
for childhood SEP with BMI
(P¼0.003) and proportion
overweight (BMI, 25 kg/m
2
)
(P¼0.003).
Females: inverse association
for childhood SEP with BMI
(P¼0.01) and proportion
overweight (BMI, 25 kg/m
2
)
(P¼0.02).
Not available. Inverse association (female).
No association (male).
Inverse association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with BMI
(P>0.05) or proportion
overweight (BMI, 25 kg/m
2
)
(P>0.05).
Males: no association for
childhood SEP and BMI
(P>0.05) or proportion
overweight (BMI, 25 kg/m
2
)
(P>0.05).
For example, in females, the
mean BMI for nonmanual
childhood social
class ¼23.3 (SD, 4.0) kg/m
2
and for manual childhood
social class ¼24.2 (SD, 4.6)
kg/m
2
(P¼0.01); in males,
these were 25.2 (SD, 3.6)
and 25.3 (SD, 3.7) kg/m
2
,
respectively (P>0.05).
34 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Pitt County Study,
summary of
2 papers listed
below
No association (female). Not available. Inconsistent association
(female).
James et al.,
2006 (57)
Inverse association for
childhood SEP and obesity.
Not available. Inverse association for
childhood SEP with obesity.
Inverse association (female). Not available. Inverse association
(female).
For example, the odds ratio
of obesity for low vs. high
parental occupation ¼
2.09 (95% CI: 1.27, 3.46).
For example, the odds ratio
of obesity for low vs. high
parental occupation ¼2.21
(95% CI: 1.32, 3.68).
Bennett et al.,
2007 (54)
No association for parental
occupation (P¼0.82),
receipt of public assistance
(P¼0.96), no plumbing
(P¼0.77), no electricity
(P¼0.09), or food scarcity
(P¼0.55) with BMI. Direct
association for parental
occupation (P¼0.01) and no
plumbing (P¼0.01) with
change in BMI; no association
for receipt of public assistance
(P¼0.17), no electricity
(P¼0.28), or food scarcity
(P¼0.85) with change in
BMI.
Not available. Direct association of change
in BMI with parental
occupation (P¼0.04) and
no plumbing (P¼0.02); no
association with receipt of
public assistance
(P¼0.20), no electricity
(P¼0.27), or food scarcity
(P¼0.97).
No association
(female).
Not available. Inconsistent association
(female).
For example, the mean
increase in BMI for parental
low/unskilled ¼4.8 (SE, 0.3)
kg/m
2
and for parental
skilled ¼6.3 (SE, 0.5) kg/m
2
(P¼0.01).
For example, the mean
increase in BMI for parental
low/unskilled ¼4.8 (SE, 0.3)
kg/m
2
and for parental
skilled ¼6.1 (SE, 0.6) kg/m
2
(P¼0.04).
Johns Hopkins Precursors
Study—medical
students from the
classes of 1948–1964
Kittleson et al.,
2006 (25)
Not available. Inverse association for
childhood SEP with BMI.
Not available. Not available. Inverse association
(male).
Not available.
For example, for the 60–69-
year age group, the BMI for
low father’s occupation ¼
25.7 (SD, 3.5) kg/m
2
and for
high father’s occupation ¼
25.0 (SD, 3.4) kg/m
2
(P<0.05).
Cardiovascular Risk in
Young Finns Study
Kivimaki et al.,
2006 (60)
Not available. Females: inverse association
for childhood SEP with WC
(P¼0.02) and WHR
(P¼0.0002); no association
with BMI (P¼0.07).
Not available. Not available. Inverse association
(female).
No association
(male).
Not available.
Males: inverse association
for childhood SEP with WHR
(P¼0.006); no association
with BMI (P¼0.13) or WC
(P¼0.06).
Table continues
Child Socioeconomic Position and Adult Obesity 35
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
For example, the mean WC
for manual vs. upper
nonmanual parental
occupation ¼80.7 vs. 77.2
cm in females, respectively,
and 90.2 vs. 88.1 cm in
males, respectively.
Polish conscription
samples in 1965,
1986, 1995, and
2001, summary of
3 papers listed below
Inconsistent association
(male).
Not available. Inconsistent association
(male).
Bielicki et al.,
2000 (59)
Not available. Not available. For the combined years
1965, 1986, and 1995: no
association for childhood
SEP with BMI.
Not available. Not available. No association (male).
For example, the Fratio for
paternal education in a
3-way analysis of variance
examining variation in BMI
with 2 df ¼0.40 (P>0.05).
Koziel et al.,
2004 (51)
For year 1986: no
association for childhood
SEP with BMI (P¼0.54).
Not available. For the combined years
1986, 1995, and 2001:
association for childhood
SEP and BMI (P¼0.04) but
no clear gradient.
Inconsistent association
(male).
Not available. Inconsistent association
(male).
For year 1995: no
association for childhood
SEP with BMI (P¼0.39).
For year 2001: inverse
association for childhood
SEP with BMI (P¼0.01).
For example, the mean BMI for
parental wealthy vs. poor ¼
21.97 (SD, 2.50) kg/m
2
vs.
21.98 (SD, 2.30) kg/m
2
in
1986 and ¼22.34 (SD,
3.11) kg/m
2
vs. 22.12
(SD, 3.02) kg/m
2
in 2001.
For example, the Fratio for
parental occupation/
education in a 4-way
analysis of variance
examining variation in BMI
with 2 df ¼3.29 (P¼0.04).
Koziel et al.,
2006 (50)
For year 1986: no
association for childhood
SEP with BMI.
Not available. Not available. Inconsistent association
(male).
Not available. Not available.
For year 1995: association for
childhood SEP and BMI but
no clear gradient.
36 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

For example, the mean BMI
differs significantly between
males from intelligentsia
(highest social class) (22.9
(SD, 2.9) kg/m
2
) and those
from entrepreneurial
background (22.6 (SD, 3.0)
kg/m
2
)(P¼0.025) and
between intelligentsia and
farming background (22.4
(SD, 2.7) kg/m
2
)(P¼0.045).
For year 2001: association
between childhood SEP
and BMI but no clear
gradient.
For example, the mean BMI
differs significantly between
intelligentsia (22.3 (SD, 3.0)
kg/m
2
) and entrepreneurial
background (22.7 (SD, 3.3)
kg/m
2
)(P¼0.011).
Kaiser Permanente
Women Twins Study
Krieger et al.,
2001 (53)
Inverse association for
childhood SEP with BMI.
No association for
childhood SEP with BMI.
No association for
childhood SEP with BMI.
Inverse association
(female).
No association
(female).
No association
(female).
For example, the odds ratio
for elevated BMI ¼1.43
(95% CI: 1.01, 2.05) for
parental working class vs.
nonworking class.
For example, among
participants who were non-
working class as adults, the
odds ratio for parental
working class vs.
nonworking class ¼1.33
(95% CI: 0.90, 1.97).
For example, among
participants who were non-
working class as adults, the
odds ratio for parental
working class vs. non-
working class ¼1.31 (95%
CI: 0.88, 1.95).
Survey of employed
middle-aged men
and women in
Helsinki
Laaksonen et al.,
2004 (18)
Females: inverse
association for
childhood economic
difficulties and parental
education with obesity.
Females: inverse
association for
childhood economic
difficulties with obesity;
no association for
parental education
with obesity.
Not available. Inverse association
(female).
Inconsistent association
(male).
Inconsistent
association (female).
No association
(male).
Not available.
Males: inverse association
for parental education with
obesity; no association for
childhood economic difficulty
with obesity.
Males: no association
for parental education or
childhood economic
difficulties with obesity.
For example, the odds ratio
of obesity for economic
difficulties vs. no economic
difficulties ¼1.49 (95% CI:
1.22, 1.82) in females and
1.22 (95% CI: 0.80, 1.85)
in males.
For example, the odds ratio
of obesity for economic
difficulties vs. no economic
difficulties ¼1.39 (95% CI:
1.14, 1.71) in females and
1.10 (95% CI: 0.72, 1.70)
in males.
Table continues
Child Socioeconomic Position and Adult Obesity 37
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Malmo
¨Diet and
Cancer Prospective
Cohort Study
Lahmann et al.,
2000 (52)
Inverse association for
childhood SEP with percent
body fat (P<0.0001), WC
(P<0.0001), WHR
(P¼0.001), and weight
change (P<0.0001).
Not available. Inverse association for
childhood SEP with percent
body fat (P<0.0001), WC
(P<0.0001), WHR
(P¼0.006), and weight
change (P<0.0001).
Inverse association
(female).
Not available. Inverse association
(female).
For example, the mean percent
body fat for parental
unskilled manual worker and
higher nonmanual worker ¼
31.4% (SE, 0.2) and 30.1%
(SE, 0.3), respectively.
For example, the regression
coefficient of parental
occupation with percent
body fat ¼0.122 (SE,
0.033) (P<0.0001).
1966 Northern Finland
birth cohort,
summary of 3 papers
listed below
Inverse association
(female).
Inconsistent
association (male).
Not available. No association
(female).
No association
(male).
Laitinen et al.,
2001 (29)
Females: inverse association
for childhood SEP with
BMI (P¼0.061), WC
(P<0.001), and WHR
(P<0.001).
Not available. Not available. Inverse association
(female).
Inverse association
(male).
Not available. Not available.
Males: inverse association
for childhood SEP with BMI
(P¼0.023) and WHR
(P¼0.023); no association
with WC (P>0.05).
For example, the BMI for social
class level I/II and farming ¼
23.4 (SD, 4.4) and 23.8 (SD,
4.6) kg/m
2
, respectively, for
females and 25.0 (SD, 3.6)
and 25.4 (SD, 3.7) kg/m
2
,
respectively, for males.
Laitinen et al.,
2002 (28)
Not available. Not available. Females: no association
for childhood SEP with
obesity.
Not available. Not available. No association
(female).
No association
(male).
Males: no association for
childhood SEP with obesity.
For example, the odds ratio of
obesity for parental
occupation level IV vs.
I/II ¼0.95 (95% CI: 0.67,
1.36) in females and 0.71
(95% CI: 0.49, 1.03) inmales.
38 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Laitinen et al.,
2004 (27)
Females: inverse association
for childhood SEP with
obesity (P¼0.013).
Not available. Not available. Inverse association (female).
No association (male).
Not available. Not available.
Males: no association for
childhood SEP with obesity
(P¼0.091).
For example, the percent
obese for ‘‘professional’’ and
‘‘unskilled worker/no
occupation/unknown’’ ¼9%
and 13%, respectively, in
females and 8% and 9%,
respectively, in males.
Glasgow Alumni Cohort
Okasha et al.,
2003 (42)
Inverse association for
childhood SEP with BMI.
Not available. Not available. Inverse association
(male).
Not available. Not available.
For example, the odds ratio
for overweight in social
class III/V vs. I ¼1.72
(95% CI: 1.10, 2.68).
Dunedin Multidisciplinary
Health and
Development Study—
1972–1973 cohort
Poulton et al.,
2002 (56)
Not available. Not available. Inverse association for
childhood SEP with
BMI (P<0.0001) and
WHR (P¼0.002).
Not available. Not available. Inverse association
(male and female).
For example, the regression
coefficient for the
association of low vs. high
parental occupation with
BMI ¼1.95 (SE, 0.49).
1958 British birth cohort,
summary of 4 papers
listed below
Inverse association
(female).
b
Inverse association
(male).
b
Inverse association
(female).
b
Inverse association
(male).
b
Inverse association
(female).
Inverse association
(male).
Power et al.,
2003 (45)
Females: inverse association
for childhood SEP with
obesity.
Females: inverse association
for childhood SEP with
obesity.
Females: inverse
association for childhood
SEP with obesity.
Inverse association
(female).
Inverse association
(male).
Inverse association
(female).
Inverse association
(male).
Inverse association
(female).
Inverse association
(male).
Males: inverse association
for childhood SEP with
obesity.
Males: inverse association
for childhood SEP with
obesity.
Males: inverse association
for childhood SEP with
obesity.
For example, the odds ratio
of obesity for higher vs. lower
parental occupation ¼1.29
(95% CI: 1.14, 1.46) in
females and 1.27 (95%
CI: 1.11, 1.44) in males.
For example, the odds ratio
of obesity for higher vs. lower
parental occupation ¼1.28
(95% CI: 1.14, 1.43) in
females and 1.22 (95%
CI: 1.07, 1.39) in males.
For example, the odds ratio
of obesity for higher vs. lower
parental occupation ¼1.23
(95% CI: 1.10, 1.38) in
females and 1.18 (95%
CI: 1.04, 1.35) in males.
Power et al.,
2005 (44)
Females: inverse association
for childhood SEP with
obesity.
Females: inverse
association for childhood
SEP with obesity.
Not available. Inverse association
(female).
Inverse association
(male).
Inverse association
(female).
Inverse association
(male).
Not available.
Males: inverse association for
childhood SEP with obesity.
Males: inverse association for
childhood SEP with obesity.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.8 (95% CI:
1.5, 2.3) in females and 1.5
(95% CI: 1.3, 1.9) in males.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.7 (95% CI:
1.4, 2.2) in females and 1.4
(95% CI: 1.2, 1.8) in males.
Table continues
Child Socioeconomic Position and Adult Obesity 39
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Power et al.,
2007 (43)
Inverse association for
childhood SEP with BMI.
Inverse association for
childhood SEP with BMI.
Not available. Inverse association
(male and female).
Inverse association
(male and female).
Not available.
For example, the mean
difference in BMI per
increase in father’s social
class grade (I–V) ¼0.45
(95% CI: 0.36, 0.53) kg/m
2
.
For example, the mean
difference in BMI per
increase in father’s social
class grade (I–V) ¼0.41
(95% CI: 0.32, 0.49) kg/m
2
.
Thomas et al.,
2007 (48)
Inverse association for
childhood SEP with BMI
(P<0.001).
Not available. Not available. Inverse association
(male and female).
Not available. Not available.
For example, the mean BMI
for father’s social class I/II vs.
IV/V ¼26.22 (95% CI: 25.99,
26.44) kg/m
2
vs. 27.97 (95%
CI: 27.72, 28.23) kg/m
2
.
Spanish sampling from
census, summary of
2 papers listed below
Inverse association
(female).
No association
(male).
No association
(female).
No association
(male).
Not available.
Regidor et al.,
2004 (47)
Females: inverse association
for childhood SEP with
BMI and WC.
Females: no association
for childhood SEP with
BMI or WC.
Not available. Inverse association
(female).
No association
(male).
No association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with BMI
or WC.
Males: no association for
childhood SEP with BMI or
WC.
For example, the difference
in mean BMI per increase in
social class grade ¼0.28
(95% CI: 0.06, 0.49) kg/m
2
in
females and 0.10 (95% CI:
0.10, 0.30) kg/m
2
in males.
For example, the difference
in mean BMI per increase in
social class grade ¼0.06
(95% CI: 0.17, 0.28) kg/m
2
in females and 0.10 (95% CI:
0.12, 0.33) kg/m
2
in males.
Regidor et al.,
2004 (46)
Females: inverse association
for childhood SEP with
elevated BMI and
abdominal obesity.
Females: inverse association
for childhood SEP with
elevated BMI; no association
with abdominal obesity.
Not available. Inverse association
(female).
No association
(male).
Inconsistent association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with
elevated BMI or abdominal
obesity.
Males: no association for
childhood SEP with elevated
BMI or abdominal obesity.
For example, the prevalence
ratio of abdominal obesity for
working class vs.
nonmanual/self-employed
farmers ¼1.12 (95% CI:
1.00, 1.25) in females and
0.99 (95% CI: 0.89, 1.09)
in males.
For example, the prevalence
ratio of abdominal obesity for
working class vs.
nonmanual/self-employed
farmers ¼1.02 (95% CI:
0.97, 1.07) in females and
0.97 (95% CI: 0.87, 1.08)
in males.
40 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

GLOBE Study, summary
of 3 papers listed
below
No association
(female).
No association
(male).
No association
(female).
No association
(male).
Inverse association
(female).
b
No association
(male).
b
van de
Mheen et al.,
1998 (49)
Not available. Not available. Inverse association for
childhood SEP with BMI.
Not available. Not available. Inverse association
(male and female).
For example, the odds ratio
for unskilled manual vs. high
professional ¼1.76
(P
overall effect
<0.01).
Power et al.,
2005 (44)
Females: no association
for childhood SEP with
obesity.
Females: No association
for childhood SEP with
obesity.
Not available. No association
(female).
No association
(male).
No association
(female).
No association
(male).
Not available.
Males: no association
for childhood SEP with
obesity.
Males: no association for
childhood SEP with obesity.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.2 (95% CI:
0.8, 1.7) in females and 1.1
(95% CI: 0.7, 1.5) in males.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼0.9 (95% CI:
0.7, 1.4) in females and 0.8
(95% CI: 0.5, 1.2) in males.
Giskes et al.,
2008 (39)
Not available. Not available. Females: inverse association
for childhood SEP with BMI
(P<0.01), likelihood of
overweight/obesity (BMI,
25 kg/m
2
)(P¼0.01), and
change in prevalence of
overweight/obesity (P<
0.01); no association with
change in BMI (P¼0.07).
Not available. Not available Inverse association
(female).
No association
(male).
Males: no association for
childhood SEP with BMI
(P¼0.86), likelihood of
overweight/obesity (P¼
0.89), change in BMI (P¼
0.86), or change in
prevalence of overweight/
obesity (P¼0.34).
For example, the mean BMI for
father’s professional class
vs. blue collar class ¼22.82
(95% CI: 21.84, 23.80) kg/m
2
vs. 24.72 (95% CI: 24.33,
25.09) kg/m
2
in females,
respectively, and 25.33
(95% CI: 24.40, 26.27) kg/m
2
vs. 25.38 (95% CI: 24.88,
25.88) kg/m
2
in males,
respectively.
Harvard Study of
Moods and Cycles
Wise et al.,
2002 (15)
Not available. Inverse association for
childhood SEP with
elevated BMI.
Not available. Not available. Inverse association
(female).
Not available.
For example, the prevalence of
BMI >30 kg/m
2
differs
between ‘‘no lifetime
economic distress’’ and
‘‘economic distress as child
only’’: 8.5% vs. 21.9%,
respectively (P¼0.008).
Table continues
Child Socioeconomic Position and Adult Obesity 41
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Kuopio Ischemic Heart
Disease Risk Factor
Study
Power et al.,
2005 (44)
No association for childhood
SEP with obesity.
No association for childhood
SEP with obesity.
Not available. No association
(male).
No association
(male).
Not available.
For example, the odds ratio
of obesity for manual vs.
nonmanual ¼1.1 (95%
CI: 0.8, 1.4).
For example, the odds ratio
of obesity for manual vs.
nonmanual ¼1.1 (95%
CI: 0.8, 1.4).
Swedish Survey of
Living Conditions
Power et al.,
2005 (44)
Females: inverse
association for childhood
SEP with obesity.
Females: inverse association
for childhood SEP with obesity.
Not available. Inverse association
(female).
No association
(male).
Inverse association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with obesity.
Males: no association for
childhood SEP with obesity.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.7 (95% CI:
1.2, 2.3) in females and 1.2
(95% CI: 0.9, 1.6) in males.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.4 (95% CI:
1.1, 1.9) in females and 1.2
(95% CI: 0.9, 1.5) in males.
Danish Longitudinal
Study on Work,
Unemployment,
and Health
Power et al.,
2005 (44)
Females: inverse
association for childhood
SEP with obesity.
Females: no association
for childhood SEP with
obesity.
Not available. Inverse association
(female).
Inverse association
(male).
No association
(female).
No association
(male).
Not available.
Males: inverse association
for childhood SEP with
obesity.
Males: no association for
childhood SEP with obesity.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.5 (95% CI:
1.1, 2.0) in females and 1.3
(95% CI: 1.0, 1.8) in males.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.3 (95% CI:
1.0, 1.8) in females and 1.2
(95% CI: 0.9, 1.7) in males.
Alameda County Study,
summary of 2 papers
listed below
No association
(female).
Inconsistent
association (male).
No association
(female).
No association
(male).
Not available.
Power et al.,
2005 (44)
Females: no association
between father’s
occupation and obesity.
Females: no association
between father’s
occupation and obesity.
Not available. No association
(female).
Inverse association
(male).
No association
(female).
No association
(male).
Not available.
Males: inverse association
between father’s occupation
and obesity.
Males: no association
between father’s occupation
and obesity.
42 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.4 (95% CI:
0.8, 2.3) in females and 1.5
(95% CI: 1.0, 2.3) in males.
For example, the odds ratio
of obesity for manual vs.
nonmanual father’s
occupation ¼1.2 (95% CI:
0.7, 2.0) in females and 1.4
(95% CI: 0.9, 2.3) in males.
Baltrus et al.,
2007 (38)
Females: no association
for childhood SEP with
weight gain.
Females: no association
for childhood SEP with
weight gain.
Not available. No association
(female).
No association
(male).
No association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with weight
gain.
Males: no association
for childhood SEP with
weight gain.
For example, the mean
weight gain per year for low
vs. high childhood
SEP ¼0.01 kg (P¼0.698)
in males and 0.05 kg
(P¼0.099) in females.
For example, the mean
weight gain per year for low
vs. high childhood
SEP ¼0.01 kg
(P¼0.801) in males and
0.03 kg (P¼0.454) in
females.
Survey of female
Polish students
Wronka and
Pawlinska-
Chmara,
2007 (58)
Inverse association for
mother’s education and
childhood economic status
with BMI. No association for
father’s education, mother’s
occupation, or father’s
occupation with BMI.
Not available. Inverse association for
mother’s education
(P¼0.0031) with BMI. No
association for father’s
education (P¼0.7778) or
general economic status
(P¼0.2792) with BMI.
Inconsistent
association (female).
Not available. No association
(female).
For example, the mean BMI
for mother’s having a primary
education vs. a university
education ¼21.63 (95% CI:
21.16, 22.10) kg/m
2
vs.
20.65 (95% CI: 20.28, 21.02)
kg/m
2
and, for father’s
having a primary education
vs. a university education, it
is 20.97 (95% CI: 20.61,
21.33) kg/m
2
vs. 21.10
(95% CI: 20.70, 21.50) kg/m
2
.
For example, the regression
coefficient for BMI in relation
to mother’s education ¼
0.4115 (P¼0.0031) and
father’s education ¼0.0356
(P¼0.7778).
Nurses’ Health Study
Lidfeldt et al.,
2007 (41)
Not available. Inverse association for
childhood SEP with BMI
(P<0.001).
Not available. Not available. Inverse association
(female).
Not available.
For example, the mean BMI for
laborer father’s social
class ¼24.7 kg/m
2
and for
professional father’s social
class ¼23.7 kg/m
2
(P
trend
<
0.001).
Table continues
Child Socioeconomic Position and Adult Obesity 43
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Table 2. Continued
Study, Authors,
Year of Publication
(Reference)
Statistical Adjustments Conclusions on Overall Association
Age Age and
Adult SEP
Age, Adult SEP,
and Other Obesity
Risk Factors
a
Adjusted
for Age
Adjusted for
Age and
Adult SEP
Adjusted for Age,
Adult SEP, and
Other Risk Factors
a
Health and Retirement
Study
Best et al.,
2005 (17)
Not available. Not available. Females: direct association
for father’s education with
prevalence of overweight
(P<0.05) and obesity
(P<0.001). No association
for mother’s education or
family financial status with
prevalence of overweight or
obesity, respectively.
Not available. Not available. No association
(female).
Inconsistent
association (male).
Males: inverse association
for mother’s education with
prevalence of obesity
(P<0.05); no association
with prevalence of
overweight. No association
for father’s education with
prevalence of overweight or
obesity. Inverse association
for family financial status
with prevalence of
overweight (P<0.01) and
obesity (P<0.001).
For example, the odds of being
obese by father’seducation ¼
0.26 (SE, 0.07) in females.
The odds of being obese b y
mother’s education ¼0.18
(SE, 0.09) in males.
Midspan Family
Study
Hart et al.,
2008 (40)
Females: inverse association
for childhood SEP with BMI
(P¼0.002), WC
(P¼0.016), and percent
obese (P¼0.001).
Females: no association
for childhood SEP with BMI
(P¼0.16), WC (P¼0.20),
or percent obese (P¼0.08).
Not available. Inverse association
(female).
No association
(male).
No association
(female).
No association
(male).
Not available.
Males: no association for
childhood SEP with BMI
(P¼0.87), WC (P¼0.54),
or percent obese (P¼0.73).
Males: no association
for childhood SEP with BMI
(P¼0.57), WC (P¼0.56),
or percent obese (P¼0.68).
For example, the mean WC
for father’s nonmanual vs.
father’s manual social
class ¼78.9 vs. 80.7 cm
(P¼0.016) for females,
respectively, and 93.9 vs.
93.4 cm (P¼0.54) for
males, respectively.
For example, the mean change
in WC by father’s social
class ¼0.38 (95% CI:
0.20, 0.96) cm in females
and 0.18 (95% CI: 0.78,
0.42) cm in males.
44 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

Kingdom, Finland, the United States, Sweden, Spain, and
Denmark (12, 13, 18, 19, 23, 26, 27, 29–31, 37, 40, 44–47,
52, 53). Among males, no association was found in studies
conducted in the United Kingdom, Finland, the Netherlands,
Sweden, and Spain (12, 19, 23, 40, 44, 46, 47); significant
inverse associations were found in the United Kingdom and
Denmark (26, 30, 31, 42, 44, 45). One notable exception
was the study from China, which examined the association
between childhood SEP (measured as retrospectively re-
called parental possession of a watch, a sewing machine,
and a bicycle) and waist circumference, controlling for
a number of potential risk factors including smoking status,
alcohol consumption, and adult SEP (16). This study re-
ported no association among females and an unprecedented
direct association among males. Apart from this unique
finding, no striking trends among the developed countries
in terms of the association between childhood SEP and
adulthood obesity were apparent (Tables 1 and 2).
Race/ethnicity
In line with the somewhat limited geographic range of the
studies in this review, samples appeared to be quite ethni-
cally homogenous, and race/ethnicity was infrequently ad-
justed for. Only papers based on 6 studies made specific
references to their largely white samples (22, 23, 25, 27–
29, 32–35, 53, 55, 56). However, many stated that they were
working with random samples from a given national or
municipal population. Because of the geographic regions
covered by the studies, it is assumed that results apply pri-
marily to those of white race/ethnicity. An exception is the
Pitt County Study, which focused on black females living in
the United States (54, 57). It found that females with parents
who had unskilled or farm labor occupations had an odds
ratio of 2.21 (95% CI: 1.32, 3.88) for obesity after adjust-
ment for age, adulthood SEP, and other obesity risk factors,
compared with females with at least 1 parent who had an
occupation classified as skilled (57).
Age and birth year effects
Only 23% (7 of 30) of studies examined the relation
between childhood SEP and obesity in exclusively younger
subjects aged 19–39 years (12, 13, 27–29, 50, 51, 56, 58–
60). Of the studies in younger individuals, 3 of 4 studies
showed significant inverse associations after adjustment for
age in females (12, 13, 27–29), and none of the 3 studies
among males showed significant inverse associations (12,
27, 29, 50, 51, 59).
It was clear in 3 (20, 25, 42) of the 8 longitudinal analyses
(12, 25, 38, 39, 42, 52, 54, 61) that the association between
childhood SEP and obesity was increasingly apparent as the
age at which obesity was measured increased. Hardy et al.
(20) found that the annual rate of increase in body mass
index was 0.03 (95% CI: 0.02, 0.04) kg/m
2
higher among
males and females whose father had a manual occupation
than among those whose fathers had nonmanual occupa-
tions, indicating that the disparity between the 2 groups
was increasing with time. Similarly, Kittleson et al. (25)
found significant associations between father’s occupation
Guangzhou Biobank
Cohort Study
Schooling et al.,
2008 (16)
Not available. Not available. Females: no association
for parental possession of
watch, sewing machine, and
bicycle with WC (P¼0.12).
Not available. Not available. No association (female).
Direct association (male).
Males: direct association
for parental possession of
watch, sewing machine, and
bicycle with WC (P¼0.02).
For example, the difference in
WC between parental
possession of 3 items vs.
0 items ¼0.29 (95% CI:
0.81, 0.24) cm in females
and 1.34 (95% CI: 0.38,
2.33) cm in males.
Abbreviations: BMI, body mass index; CI, confidence interval; GLOBE, Gemifloxacin Long-term Outcomes in Bronchitis Exacerbations; SD, standard deviation; SE, standard error; SEP, socioeconomic position; WC,
waist circumference; WHR, waist/hip ratio.
a
Please refer to Table 1 for adjustment covariates.
b
Gender-specific analyses were prioritized for inclusion in Tables 3–5; consequently, in cases where both gender-specific and pooled analyses were reported, only the results for gender-specific analyses were
included in Tables 3–5.
Child Socioeconomic Position and Adult Obesity 45
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

and obesity when obesity was measured later in adulthood
(40–69 years) but not earlier in adulthood.
Approximately 57% (17 of 30) of studies were focused on
samples that consisted primarily of participants born prior
to 1950, while 13 studies examined participants who were
largely born in the second half of the 20th century (Table 1).
Among age-adjusted analyses of females born before 1950,
64% (7 of 11) of studies showed an inverse association
(19, 23, 26, 30, 31, 40, 44, 46, 47, 52, 53), while for those
born after 1950, this value was 78% (7 of 9 studies) (12, 13,
18, 27, 29, 37, 44, 45). Among males born in the first half of
the 20th century, 22% (2 of 9) of age-adjusted studies reported
an inverse association (26, 30, 42), while 33% (2 of 6) of
studies in which the majority of participants were born after
1950 reported such an association (44, 45).
DISCUSSION
This systematic review of 30 studies published from
January 1998 through September 2008 demonstrated con-
sistent inverse associations between childhood SEP and
adulthood obesity in females after adjustment for age. Fur-
ther statistical adjustment for other obesity covariates
(including adulthood SEP, health behaviors, cognitive abil-
ities, maternal obesity, among others) typically reduced the
magnitude of effect, suggesting that these may be some of
the mechanisms responsible for the observed associations
between childhood SEP and adulthood obesity in females.
In males, findings were weaker and less consistent. There
was some indication that associations between childhood
SEP and adulthood obesity may be stronger when childhood
SEP is measured in childhood rather than recalled during
adulthood.
Prior literature
A review of 12 studies published up to the year 1998 by
Parsons et al. (6) found consistent inverse associations be-
tween childhood SEP and adulthood obesity in males (8 of 9
studies) and females (4 of 5 studies). Our analyses of 30
studies published from 1998 through September 2008
showed inverse associations between childhood SEP and
obesity in females after adjustment for age, which is con-
sistent with the findings of Parsons et al. However, studies
published since 1998 have shown a general lack of consis-
tent associations between childhood SEP and adulthood
obesity in males, unlike studies in the review by Parsons
et al. Furthermore, the effects in females and males were
substantially reduced following adjustment for adulthood
SEP and other potential obesity risk factors. Recent evi-
dence has suggested that socioeconomic gradients in obesity
may be declining, indicative of the widespread nature of the
obesity epidemic (9). In her 2007 systematic review that
updated a previous review published by Sobal and Stunkard
(62) in 1989, McLaren (10) noted the recent finding that
63% of studies of females showed an inverse associa-
tion between adulthood SEP and obesity in developed coun-
tries. This was less than that reviewed earlier by Sobal and
Stunkard (62), which showed that 93% and 75% of studies
Table 3. Summary of Statistically Significant Associations
a
Between Childhood Socioeconomic Position and Adulthood Obesity
After Adjustment for Age
Direction of
Association
a
No. of Studies
Females Males
Males and
Females
Combined
Total
Inverse association 14 4 0 18
Inconsistent association 2 4 0 6
No association 4 7 0 11
Direct association 0 0 0 0
Total 20 15 0 35
a
Associations are considered statistically significant when P<
0.05 or 95% confidence intervals do not encompass the reference
point estimate. Please refer to Table 2 for the statistical significance
of the associations for each study, which determined the categoriza-
tion of studies summarized in this table.
Table 4. Summary of Statistically Significant Associations
a
Between Childhood Socioeconomic Position and Adulthood Obesity
After Adjustment for Age and Adulthood Socioeconomic Position
Direction of
Association
a
No. of Studies
Females Males
Males and
Females
Combined
Total
Inverse association 8 2 1 11
Inconsistent association 3 1 0 4
No association 6 11 0 17
Direct association 0 0 0 0
Total 17 14 1 32
a
Associations are considered statistically significant when P<
0.05 or 95% confidence intervals do not encompass the reference
point estimate. Please refer to Table 2 for the statistical significance
of the associations for each study, which determined the categoriza-
tion of studies summarized in this table.
Table 5. Summary of Statistically Significant Associations
a
Between Childhood Socioeconomic Position and Adulthood Obesity
After Adjustment for Age, Adulthood Socioeconomic Position, and
Other Potential Obesity Risk Factors
Direction of
Association
a
No. of Studies
Females Males
Males and
Females
Combined
Total
Inverse association 4 2 1 7
Inconsistent association 1 2 1 4
No association 7 3 0 10
Direct association 0 1 0 1
Total 12 8 2 22
a
Associations are considered statistically significant when P<
0.05 or 95% confidence intervals do not encompass the reference
point estimate. Please refer to Table 2 for the statistical significance
of the associations for each study, which determined the categoriza-
tion of studies summarized in this table.
46 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

in females had inverse associations in the United States and
other developed countries, respectively. It may currently be
that, although many of the causes of obesity have been
identified (e.g., high caloric diets and sedentary activities),
societies and people of all social classes continue to struggle
to alter obesity’s fundamental causes.
Mechanisms
Gender differences. There are potential mechanisms that
may explain the gender differences in their relation between
childhood SEP and adulthood obesity, including parity and
socioeconomic-patterned pressure to be slim in females.
With regard to parity, there is consistent evidence that chil-
dren are likely to have a socioeconomic position (e.g., edu-
cation and income) similar to that of their parents (63).
Furthermore, women’s educational attainment is inversely
related to birth rates (64). Childbirth is associated with in-
creased long-term central obesity compared with before
pregnancy (65). There is evidence to suggest that there
may be stronger social pressure against obesity in females
compared with males, and that pressure is greater among
females with high SEP than low SEP (62, 66). Even in
youth, this stigma may be socioeconomically distributed
particularly in females. For example, in a study of 1,248
female adolescents living in England, participants with high
SEP (measured through self-report of family affluence, in-
cluding car ownership, ownership of a computer, housing
tenure, and whether the student had the option of a free
school meal) had a greater awareness of social ideals of
slimness, defined a lower body mass index as ‘‘fat,’’ and
were more likely to have used healthy weight control meth-
ods compared with participants of lower SEP (67). In a study
of Australians 18 years of age, for females, the frequency of
consumption of cereals, fruits, and vegetables was greatest
in participants living in neighborhoods with high SEP, while
the consumption of high-fat foods was highest in partici-
pants living in neighborhoods with low SEP; for males,
there were no significant differences related to SEP (68).
Other mechanisms. Maternal socioeconomic disadvan-
tage predicts low birth weight, likely due to a clustering of
risk factors that are more likely to occur in pregnant women
of low SEP, including malnutrition, smoking, alcohol con-
sumption, drug abuse, and stress (69–75). Low birth weight
can act as a predictor of a ‘‘catch-up’’ phase during which
body weight increases more so than height, leading to in-
creased prevalence of obesity among those of low birth
weight compared with others (76).
Unhealthy behaviors including lack of leisure-time
physical activity (77, 78), unhealthy diets (79–83), and
smoking (84) tend to be higher in adults with low SEP
compared with high SEP. These behaviors can be modeled
as normative behaviors to offspring (85). Early childhood
is a critical period for the development of food and flavor
preferences, as well as the ability to self-regulate food
consumption (86). Childhood socioeconomic disadvantage
is inversely related to smoking in adulthood (44, 87, 88).
This inverse gradient would tend to counteract the gradient
in obesity, as smoking is protective against elevated body
mass index; however, smoking is predictive of central
obesity (89, 90).
There is consistent evidence that offspring’s SEP is
influenced by their parents’ SEP, where parental SEP and
offspringSEParepositivelyassociated(63).SEPinadult-
hood was shown in a recent systematic review of 333 pa-
pers to be inversely associated with obesity in women, and
it had a flat or curvilinear association with obesity in males
(10). Adulthood SEP provides several mechanisms that can
potentially influence obesity, including inverse gradients
of SEP with risk for smoking (84), leisure-time seden-
tary activity (77, 78, 91), and obesogenic diet (79–83), as
well as parity (64). Furthermore, obesity is reported to be
stigmatizedmorehighlyinwomenthanmen(although
there are racial and ethnic differences in the stigmatiza-
tion), and consequently obesity may limit upward social
mobility more so in women than men for many groups of
women (66).
In the articles summarized in this systematic review,
adjustment for the aforementioned potential mechanisms
typically reduced the effect size between childhood SEP
and obesity, suggesting that some or all may play a contrib-
uting role. Many studies individually adjusted for adult-
hood SEP, and doing so typically reduced effect sizes.
This suggests that adulthood SEP is important in explain-
ing the observed associations between childhood SEP and
obesity. Few studies adjusted for other potential obesity
risk factors individually and, although these further adjust-
ments usually slightly reduced effect sizes, it is difficult to
know which specific obesity risk factors may be particu-
larly important.
Strengths and limitations
The strengths of this review include the methodological
approach (92, 93). Specifically, we used the services of
a professional librarian to devise and execute the literature
search strategy, searched 5 different health databases, hand-
searched key relevant journals, contacted experts in the
field, and allowed papers from any language (which can
increase the number of null findings found) to be included.
Two researchers independently performed the secondary
selection of manuscripts to include in the review. Further-
more, effect sizes and the statistical significance of findings
in each publication are reported in this review, which ena-
bles transparency of methods to allow the readers to better
assess for themselves if they agree with the methodological
approach and summaries provided by the authors.
With regard to limitations, ideally 2 researchers would
independently perform the initial paper selection from all
identified publications; only 1 researcher did this. However,
we used an approach where any manuscript that had even
a slight indication of being relevant was included in the list
for secondary selection by 2 investigators. The complete-
ness of this approach was in part confirmed through the
observation that no contacted experts had further publica-
tions to suggest for inclusion. A limitation of the publica-
tions summarized in this review is that the majority of
analyses were cross-sectional (i.e., obesity was measured
at only 1 time point), and all studies were based on
Child Socioeconomic Position and Adult Obesity 47
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

observational data. Consequently, causality cannot be at-
tributed to the observed associations between childhood
SEP and adulthood obesity solely on the basis of these
observational studies. Furthermore, the findings from this
review are expected to suffer from positive publication
bias, where significant findings are typically more likely
to be written up and published than null findings. We would
expect this phenomenon to induce bias toward a significant
association between childhood SEP and obesity. In addi-
tion, the majority of studies implemented parental occupa-
tion (and usually father’s occupation in particular) as the
measure of childhood SEP, and consequently the findings
are reflective primarily of the association between parental
occupation and adulthood obesity, rather than other meas-
ures of childhood SEP, such as parental education, eco-
nomic distress, housing conditions, or measures of
father’s versus mother’s SEP (2, 3).
Future directions, population health, and research
implications
Overall, observational studies show inverse associations
between childhood SEP and adulthood obesity after adjust-
ment for age in developed countries, primarily in Caucasian
females, and not in Caucasian males. Little is known about
associations in racial/ethnic minorities or developing coun-
tries. The best practices for research in this field appear to
measure childhood SEP during childhood rather than retro-
spectively during adulthood whenever possible. Most stud-
ies used parental occupation as a measure of childhood SEP,
and consequently less is known about the effects of other
measures of childhood SEP, such as parental education, pa-
rental income, and economic distress. The mechanisms re-
sponsible for gender differences in the association between
childhood SEP and obesity are still not well defined. Further
work in this regard will better elucidate the reasons for the
socioeconomic and gender disparities. Additional work in-
vestigating the association between childhood SEP and lon-
gitudinal trajectories of weight gain will provide further
mechanistic knowledge of the association between child-
hood SEP and obesity.
Causal inference for the fairly consistently found associ-
ations of childhood SEP with cardiovascular disease in ob-
servational studies (4, 5) is advanced when plausible
biologic mechanisms are identified. Obesity, particularly
in females, may be 1 potential mechanism by which child-
hood SEP may influence cardiovascular disease. Adulthood
SEP appears to be an important explanatory mechanism for
the association between childhood SEP and adulthood obe-
sity. Second, because of the current epidemic of obesity in
many nations throughout the world, it is important to un-
derstand the etiology by which this is occurring. Childhood
socioeconomic disadvantage in developed countries may be
important in the development of adulthood obesity, partic-
ularly in females. Finally, evidence of socioeconomic health
disparities is often hoped to impact policy decisions that
influence the socioeconomic distribution of resources in soci-
ety. Research on specific socioeconomic interventions is re-
quired to determine whether changes in specific policies or
practices truly influence health outcomes such as car-
diovascular disease. Measures of obesity are candidate
biomarkers that havethe potential toserve as early approximate
indicators of the potential effectiveness of these interventions,
particularly in females, in conjunction with other health mea-
surement tools.
ACKNOWLEDGMENTS
Author affiliations: Department of Epidemiology, Biosta-
tistics, and Occupational Health, McGill University, Montre
´al,
Que
´be
´c, Canada (Laura C. Senese, Nisha D. Almeida,
Brendan T. Smith); Department of Family Medicine,
McMaster University, Hamilton, Ontario, Canada (Anne
Kittler Fath); and Department of Community Health, Center
for Population Health and Clinical Epidemiology, Brown
University, Providence, Rhode Island (Eric B. Loucks).
This work was supported by a Canadian Institutes of
Health New Investigator Award (E. B. L.) and Canadian
Institutes of Health operating grants MOP-81239 and
MOP-89950.
The authors are grateful to Angella Lambrou, a profes-
sional librarian at the McGill Health Sciences Library, for
assisting with devising and performing the systematic liter-
ature search.
Conflict of interest: none declared.
REFERENCES
1. Commission on Social Determinants of Health. Closing the
Gap in a Generation: Health Equity Through Action on the
Social Determinants of Health. Geneva, Switzerland: World
Health Organization; 2008.
2. Galobardes B, Shaw M, Lawlor DA, et al. Indicators of so-
cioeconomic position (part 2). J Epidemiol Community Health.
2006;60(2):95–101.
3. Galobardes B, Shaw M, Lawlor DA, et al. Indicators of so-
cioeconomic position (part 1). J Epidemiol Community Health.
2006;60(1):7–12.
4. Galobardes B, Lynch JW, Davey Smith G. Childhood socio-
economic circumstances and cause-specific mortality in
adulthood: systematic review and interpretation. Epidemiol
Rev. 2004;26(1):7–21.
5. Galobardes B, Lynch JW, Smith GD. Is the association
between childhood socioeconomic circumstances and cause-
specific mortality established? Update of a systematic review.
J Epidemiol Community Health. 2008;62(5):387–390.
6. Parsons TJ, Power C, Logan S, et al. Childhood predictors of
adult obesity: a systematic review. Int J Obes Relat Metab
Disord. 1999;23(suppl 8):S1–S107.
7. Eckel RH, York DA, Ro
¨ssner S, et al. Prevention Conference
VII: obesity, a worldwide epidemic related to heart disease and
stroke: executive summary. Circulation. 2004;110(18):
2968–2975.
8. World Health Organization. Obesity: Preventing and Manag-
ing the Global Epidemic. Geneva, Switzerland: World Health
Organization; 2000.
9. Zhang Q, Wang Y. Trends in the association between obesity
and socioeconomic status in U.S. adults: 1971 to 2000. Obes
Res. 2004;12(10):1622–1632.
48 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

10. McLaren L. Socioeconomic status and obesity. Epidemiol Rev.
2007;29(1):29–48.
11. Davey Smith G, Lynch JW. Life course approaches to socio-
economic differentials in health. In: Kuh D, Ben-Shlomo Y,
eds. A Life Course Approach to Chronic Disease Epidemiol-
ogy. 2nd ed. Oxford, United Kingdom: Oxford University
Press; 2004:77–115.
12. Huurre T, Aro H, Rahkonen O. Well-being and health behav-
iour by parental socioeconomic status: a follow-up study of
adolescents aged 16 until age 32 years. Soc Psychiatry
Psychiatr Epidemiol. 2003;38(5):249–255.
13. Ball K, Mishra GD. Whose socioeconomic status influences
a woman’s obesity risk: her mother’s, her father’s, or her own?
Int J Epidemiol. 2006;35(1):131–138.
14. Egger M, Schneider M, Davey Smith G. Spurious precision?
Meta-analysis of observational studies. BMJ. 1998;316(7125):
140–144.
15. Wise LA, Krieger N, Zierler S, et al. Lifetime socioeconomic
position in relation to onset of perimenopause. J Epidemiol
Community Health. 2002;56(11):851–860.
16. Schooling CM, Jiang CQ, Lam TH, et al. Life-course origins
of social inequalities in metabolic risk in the population of
a developing country. Am J Epidemiol. 2008;167(4):
419–428.
17. Best LE, Hayward MD, Hidajat MM. Life course pathways to
adult-onset diabetes. Soc Biol. 2005;52(3-4):94–111.
18. Laaksonen M, Sarlio-La
¨hteenkorva S, Lahelma E. Multiple
dimensions of socioeconomic position and obesity among
employees: the Helsinki Health Study. Obes Res. 2004;12(11):
1851–1858.
19. Brunner E, Shipley MJ, Blane D, et al. When does cardio-
vascular risk start? Past and present socioeconomic circum-
stances and risk factors in adulthood. J Epidemiol Community
Health. 1999;53(12):757–764.
20. Hardy R, Wadsworth M, Kuh D. The influence of childhood
weight and socioeconomic status on change in adult body mass
index in a British national birth cohort. Int J Obes Relat Metab
Disord. 2000;24(6):725–734.
21. Dundas R, Leyland AH, Macintyre S, et al. Does the primary
school attended influence self-reported health or its risk factors
in later life? Aberdeen Children of the 1950s Study. Int J
Epidemiol. 2006;35(2):458–465.
22. Ebrahim S, Montaner D, Lawlor DA. Clustering of risk factors
and social class in childhood and adulthood in British women’s
heart and health study: cross sectional analysis [electronic
article]. BMJ. 2004;328(7444):861.
23. Heraclides A, Witte D, Brunner EJ. The association between
father’s social class and adult obesity is not explained by
educational attainment and an unhealthy lifestyle in adulthood.
Eur J Epidemiol. 2008;23(8):573–579.
24. Heslop P, Smith GD, Macleod J, et al. The socioeconomic
position of employed women, risk factors and mortality. Soc
Sci Med. 2001;53(4):477–485.
25. Kittleson MM, Meoni LA, Wang NY, et al. Association of
childhood socioeconomic status with subsequent coronary
heart disease in physicians. Arch Intern Med. 2006;166(21):
2356–2361.
26. Kuh D, Hardy R, Chaturvedi N, et al. Birth weight, childhood
growth and abdominal obesity in adult life. Int J Obes Relat
Metab Disord. 2002;26(1):40–47.
27. Laitinen J, Pietilainen K, Wadsworth M, et al. Predictors of ab-
dominal obesity among 31-y-old men and women born in
northern Finland in 1966. Eur J Clin Nutr. 2004;58(1):180–190.
28. Laitinen J, Power C, Ek E, et al. Unemployment and
obesity among young adults in a northern Finland 1966 birth
cohort. Int J Obes Relat Metab Disord. 2002;26(10):
1329–1338.
29. Laitinen J, Power C, Ja
¨rvelin MR. Family social class, ma-
ternal body mass index, childhood body mass index, and age at
menarche as predictors of adult obesity. Am J Clin Nutr. 2001;
74(3):287–294.
30. Langenberg C, Hardy R, Kuh D, et al. Central and total obesity
in middle aged men and women in relation to lifetime socio-
economic status: evidence from a national birth cohort.
J Epidemiol Community Health. 2003;57(10):816–822.
31. Langenberg C, Kuh D, Wadsworth ME, et al. Social circum-
stances and education: life course origins of social inequalities
in metabolic risk in a prospective national birth cohort. Am J
Public Health. 2006;96(12):2216–2221.
32. Lawlor DA, Ebrahim S, Davey Smith G. Socioeconomic
position in childhood and adulthood and insulin resistance:
cross sectional survey using data from British women’s heart
and health study [electronic article]. BMJ. 2002;325(7368):
805.
33. Lawlor DA, Ebrahim S, Davey Smith G. Adverse socioeco-
nomic position across the lifecourse increases coronary heart
disease risk cumulatively: findings from the British women’s
heart and health study. J Epidemiol Community Health. 2005;
59(9):785–793.
34. Lawlor DA, Patel R, Fraser A, et al. The association of life
course socio-economic position with diagnosis, treatment,
control and survival of women with diabetes: findings from the
British Women’s Heart and Health Study. Diabet Med. 2007;
24(8):892–900.
35. Lawlor DA, Smith GD, Ebrahim S. Association between
childhood socioeconomic status and coronary heart disease
risk among postmenopausal women: findings from the British
Women’s Heart and Health Study. Am J Public Health. 2004;
94(8):1386–1392.
36. Mishra GD, Malik NS, Paul AA, et al. Childhood and adult
dietary vitamin E intake and cardiovascular risk factors in mid-
life in the 1946 British Birth Cohort. Eur J Clin Nutr. 2003;
57(11):1418–1425.
37. Pierce MB, Leon DA. Age at menarche and adult BMI in the
Aberdeen children of the 1950s cohort study. Am J Clin Nutr.
2005;82(4):733–739.
38. Baltrus PT, Everson-Rose SA, Lynch JW, et al. Socioeconomic
position in childhood and adulthood and weight gain over 34
years: the Alameda County Study. Ann Epidemiol. 2007;17(8):
608–614.
39. Giskes K, van Lenthe FJ, Turrell G, et al. Socioeconomic
position at different stages of the life course and its influence
on body weight and weight gain in adulthood: a longitudinal
study with 13-year follow-up. Obesity (Silver Spring). 2008;
16(6):1377–1381.
40. Hart C, McConnachie A, Upton M, et al. Risk factors in the
Midspan family study by social class in childhood and adult-
hood. Int J Epidemiol. 2008;37(3):604–614.
41. Lidfeldt J, Li TY, Hu FB, et al. A prospective study of
childhood and adult socioeconomic status and incidence of
type 2 diabetes in women. Am J Epidemiol. 2007;165(8):
882–889.
42. Okasha M, McCarron P, McEwen J, et al. Childhood social
class and adulthood obesity: findings from the Glasgow
Alumni Cohort. J Epidemiol Community Health. 2003;57(7):
508–509.
43. Power C, Atherton K, Strachan DP, et al. Life-course influ-
ences on health in British adults: effects of socio-economic
position in childhood and adulthood. Int J Epidemiol. 2007;
36(3):532–539.
Child Socioeconomic Position and Adult Obesity 49
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

44. Power C, Graham H, Due P, et al. The contribution of child-
hood and adult socioeconomic position to adult obesity and
smoking behaviour: an international comparison. Int J
Epidemiol. 2005;34(2):335–344.
45. Power C, Manor O, Matthews S. Child to adult socioeconomic
conditions and obesity in a national cohort. Int J Obes Relat
Metab Disord. 2003;27(9):1081–1086.
46. Regidor E, Banegas JR, Gutie
´rrez-Fisac JL, et al. Socioeco-
nomic position in childhood and cardiovascular risk factors in
older Spanish people. Int J Epidemiol. 2004;33(4):723–730.
47. Regidor E, Gutie
´rrez-Fisac JL, Banegas JR, et al. Obesity and
socioeconomic position measured at three stages of the life
course in the elderly. Eur J Clin Nutr. 2004;58(3):488–494.
48. Thomas C, Hyppo
¨nen E, Power C. Prenatal exposures and
glucose metabolism in adulthood: are effects mediated through
birth weight and adiposity? Diabetes Care. 2007;30(4):
918–924.
49. van de Mheen H, Stronks K, Looman CW, et al. Does child-
hood socioeconomic status influence adult health through
behavioural factors? Int J Epidemiol. 1998;27(3):431–437.
50. KoziełS, Szklarska A, Bielicki T, et al. Changes in the BMI of
Polish conscripts between 1965 and 2001: secular and socio-
occupational variation. Int J Obes (Lond). 2006;30(9):
1382–1388.
51. KoziełS, Welon Z, Bielicki T, et al. The effect of the economic
transition on the body mass index of conscripts in Poland.
Econ Hum Biol. 2004;2(1):97–106.
52. Lahmann PH, Lissner L, Gullberg B, et al. Sociodemographic
factors associated with long-term weight gain, current body
fatness and central adiposity in Swedish women. Int J Obes
Relat Metab Disord. 2000;24(6):685–694.
53. Krieger N, Chen JT, Selby JV. Class inequities in women’s
health: combined impact of childhood and adult social
class—a study of 630 US women. Public Health. 2001;115(3):
175–185.
54. Bennett GG, Wolin KY, James SA. Lifecourse socioeconomic
position and weight change among blacks: the Pitt County
Study. Obesity (Silver Spring). 2007;15(1):172–181.
55. Brunner E. Stress and the biology of inequality. BMJ. 1997;
314(7092):1472–1476.
56. Poulton R, Caspi A, Milne BJ, et al. Association between
children’s experience of socioeconomic disadvantage and
adult health: a life-course study. Lancet. 2002;360(9346):
1640–1645.
57. James SA, Fowler-Brown A, Raghunathan TE, et al. Life-
course socioeconomic position and obesity in African Amer-
ican women: the Pitt County Study. Am J Public Health. 2006;
96(3):554–560.
58. Wronka I, Pawlinska-Chmara R. Childcare, height and BMI
among female Polish university students, 2005. Econ Hum
Biol. 2007;5(3):435–542.
59. Bielicki T, Szklarska A, Welon Z, et al. Variation in the
body mass index among young adult Polish males between
1965 and 1995. Int J Obes Relat Metab Disord. 2000;24(5):
658–662.
60. Kivima
¨ki M, Smith GD, Juonala M, et al. Socioeconomic
position in childhood and adult cardiovascular risk factors,
vascular structure, and function: Cardiovascular Risk in Young
Finns Study. Heart. 2006;92(4):474–480.
61. Baltrus PT, Lynch JW, Everson-Rose S, et al. Race/ethnicity,
life-course socioeconomic position, and body weight trajec-
tories over 34 years: the Alameda County Study. Am J Public
Health. 2005;95(5):1595–1601.
62. Sobal J, Stunkard AJ. Socioeconomic status and obesity: a
review of the literature. Psychol Bull. 1989;105(2):260–275.
63. Beller E, Hout M. Intergenerational social mobility: the United
States in comparative perspective. Future Child. 2006;16(2):
19–36.
64. Mathews TJ, Venura SJ. Birth and fertility rates by educational
attainment: United States, 1994. Mon Vital Stat Rep. 1997;
45(10 suppl) (DHHS publication no. (PHS) 97-1120).
65. Gunderson EP, Murtaugh MA, Lewis CE, et al. Excess gains in
weight and waist circumference associated with childbearing: the
Coronary Artery Risk Development in Young Adults Study
(CARDIA). Int J ObesRelat Metab Disord. 2004;28(4):525–535.
66. Finkelstein EA, Ruhm CJ, Kosa KM. Economic causes and
consequences of obesity. Annu Rev Public Health. 2005;26:
239–257.
67. Wardle J, Robb KA, Johnson F, et al. Socioeconomic variation
in attitudes to eating and weight in female adolescents. Health
Psychol. 2004;23(3):275–282.
68. Milligan RA, Burke V, Beilin LJ, et al. Influence of gender and
socio-economic status on dietary patterns and nutrient intakes
in 18-year-old Australians. Aust N Z J Public Health. 1998;
22(4):485–493.
69. Campbell J, Torres S, Ryan J, et al. Physical and nonphysical
partner abuse and other risk factors for low birth weight among
full term and preterm babies: a multiethnic case-control study.
Am J Epidemiol. 1999;150(7):714–726.
70. Chomitz VR, Cheung LW, Lieberman E. The role of lifestyle
in preventing low birth weight. Future Child. 1995;5(1):
121–138.
71. Longo DR, Kruse RL, LeFevre ML, et al. An investigation of
social and class differences in very-low birth weight outcomes:
a continuing public health concern. J Health Care Finance.
1999;25(3):75–89.
72. Millar WJ, Chen J. Maternaleducation and risk factors for small-
for-gestational-age births. Health Rep. 1998;10(2):43–51.
73. Orr ST, James SA, Miller CA, et al. Psychosocial stressors and
low birthweight in an urban population. Am J Prev Med. 1996;
12(6):459–466.
74. Paarlberg KM, Vingerhoets AJ, Passchier J, et al. Psychosocial
predictors of low birthweight: a prospective study. Br J Obstet
Gynaecol. 1999;106(8):834–841.
75. Politzer RM, Yoon J, Shi L, et al. Inequality in America: the
contribution of health centers in reducing and eliminating
disparities in access to care. Med Care Res Rev. 2001;58(2):
234–248.
76. Baird J, Fisher D, Lucas P, et al. Being big or growing fast:
systematic review of size and growth in infancy and later
obesity [electronic article]. BMJ. 2005;331(7522):929.
77. Pekkanen J, Tuomilehto J, Uutela A, et al. Social class, health
behaviour, and mortality among men and women in eastern
Finland. BMJ. 1995;311(7005):589–593.
78. Kuh DJ, Cooper C. Physical activity at 36 years: patterns and
childhood predictors in a longitudinal study. J Epidemiol
Community Health. 1992;46(2):114–119.
79. Kawachi I, Berkman LF. Neighborhoods and Health.New
York, NY: Oxford University Press; 2003.
80. Drewnowski A, Specter SE. Poverty and obesity: the role of
energy density and energy costs. Am J Clin Nutr. 2004;79(1):
6–16.
81. Braddon FE, Wadsworth ME, Davies JM, et al. Social and
regional differences in food and alcohol consumption and their
measurement in a national birth cohort. J Epidemiol Commu-
nity Health. 1988;42(4):341–349.
82. Travers KD. The social organization of nutritional inequities.
Soc Sci Med. 1996;43(4):543–553.
83. Turrell G, Hewitt B, Patterson C, et al. Socioeconomic dif-
ferences in food purchasing behaviour and suggested
50 Senese et al.
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from

implications for diet-related health promotion. J Hum Nutr
Diet. 2002;15(5):355–364.
84. Update on the 1987 Task Force Report on High Blood Pressure
in Children and Adolescents: a working group report from the
National High Blood Pressure Education Program. National
High Blood Pressure Education Program Working Group on
Hypertension Control in Children and Adolescents. Pediatrics.
1996;98(4 pt 1):649–658.
85. Hanson MD, Chen E. Socioeconomic status and health be-
haviors in adolescence: a review of the literature. J Behav
Med. 2007;30(3):263–285.
86. Ventura A, Savage J, May A, et al. Early behavioural, familial
and psychosocial predictors of overweight and obesity. In:
Tremblay R, Barr R, Peters R, eds. Encyclopedia on Early
Childhood Development. Montreal, Canada: Center of Excel-
lence for Early Childhood Development; 2005. 1–10. (http://
www.child-encyclopedia.com/documents/Ventura-Savage-
May-BirchANGxp.pdf). (Accessed September 16, 2008).
87. Jefferis BJ, Power C, Graham H, et al. Effects of childhood
socioeconomic circumstances on persistent smoking. Am J
Public Health. 2004;94(2):279–285.
88. Lawlor DA, Batty GD, Morton SM, et al. Childhood
socioeconomic position, educational attainment, and adult
cardiovascular risk factors: the Aberdeen children of the
1950s cohort study. Am J Public Health. 2005;95(7):1245–
1251.
89. Shimokata H, Muller DC, Andres R. Studies in the distribution
of body fat. III. Effects of cigarette smoking. JAMA. 1989;
261(8):1169–1173.
90. Willi C, Bodenmann P, Ghali WA, et al. Active smoking and
the risk of type 2 diabetes: a systematic review and meta-
analysis. JAMA. 2007;298(22):2654–2664.
91. Li F, Fisher KJ, Bauman A, et al. Neighborhood influences on
physical activity in middle-aged and older adults: a multilevel
perspective. J Aging Phys Act. 2005;13(1):87–114.
92. Cochrane handbook for systematic reviews of interventions
4.2.6. Higgins JPT, Green S, eds. The Cochrane Library.
Vol 4. Chichester, United Kingdom: John Wiley & Sons, Ltd;
2006.
93. Pai M, McCulloch M, Gorman JD, et al. Systematic reviews
and meta-analyses: an illustrated, step-by-step guide. Natl Med
J India. 2004;17(2):86–95.
Child Socioeconomic Position and Adult Obesity 51
Epidemiol Rev 2009;31:21–51
by guest on May 18, 2011epirev.oxfordjournals.orgDownloaded from