ENV-14

Are childhood socio-economic circumstances related to coronary heart disease risk? Findings from a population-based study of older men Sheena E Ramsay,1* Peter H Whincup,2 Richard W Morris,1 Lucy T Lennon1

and SG Wannamethee1

Accepted 6 March 2007

Background The independent influence of childhood social circumstances on health in later

life remains uncertain. We examined the extent to which childhood socio-

economic circumstances are related to the risk of coronary heart disease (CHD)

in older British men, taking account of adult social class and behavioural risk

factors.

Methods A socio-economically representative sample of 5552 British men (52–74 years)

with retrospective assessment of childhood socio-economic circumstances

(father’s occupation and childhood household amenities) who were followed

up for CHD (fatal and non-fatal) for 12 years.

Results Men whose childhood social class was manual had an increased hazard ratio

(HR) 1.34 (95% CI 1.11–1.63)—this effect was diminished when adjusted for

adult social class and adult behavioural risk factors (cigarette smoking, alcohol,

physical activity and body weight) (HR 1.19; 95% CI 0.97–1.46). Men whose

family did not own a car in their childhood were at increased CHD risk

even after adjustments for adult social class and behaviours (HR 1.35, 95% CI

1.04–1.75). Men with combined exposure to both childhood and adult manual

social class had the highest risk of CHD (HR 1.51; 95% CI 1.19–1.91); this was

substantially reduced by adjustment for adult behavioural risk factors (adjusted

HR 1.28; 95% CI 0.99–1.65).

Conclusions Less affluent socio-economic conditions in childhood may have a modest

persisting influence on risk of CHD in later life.

Keywords Childhood social circumstances, adult social class, behavioural risk factors, CHD,

old age

Introduction The growing burden of chronic diseases in addition to the

current increase of an ageing population is a public health

problem faced by many parts of the developed and developing

world. Therefore, researching pathways to ill health in the

elderly will help improve our understanding of ways in which

the burden of chronic diseases in older people may be lessened.

One of the possible pathways is through the environment in

early life which affects the development of chronic diseases

such as coronary heart disease (CHD).1 There is evidence

supporting the association of childhood social class with CHD

independent of adult social class.2–4 But many of these (7 of the

10 studies in a systematic review)2,3 have not additionally taken

into account the role of adult behavioural risk factors, which

could be important influences on the relationship between

childhood social class and CHD. Some studies, which have done

so, have shown mixed results with the association of

childhood social class and CHD either diminishing or having

a weak positive effect when controlling for adult behavioural

* Corresponding author. Department of Primary Care and Population Sciences, Royal Free Hospital and University College Medical School, Rowland Hill Street, London NW3 2PF, UK. E-mail: [email protected]

1 Department of Primary Care and Population Sciences, Royal Free Hospital and University College Medical School, Rowland Hill Street, London NW3 2PF, UK.

2 Department of Community Health Sciences, St George’s University of London, London SW17 0RE, UK.

Published by Oxford University Press on behalf of the International Epidemiological Association

� The Author 2007; all rights reserved. Advance Access publication 17 April 2007 International Journal of Epidemiology 2007;36:560–566

doi:10.1093/ije/dym060

560

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from

risk factors. 5–8

We have previously shown that in middle-age

childhood social class was related to CHD prevalence, indepen-

dent of adult social class and behavioural risk factors.9 Here we

report on the relation of childhood social circumstances to the

incidence of CHD in older men studied from 52–74 years over a

12-year period, using both father’s occupation and childhood

household amenities as markers. We also investigate the

contribution of adult social class and adult behavioural risk

factors to the associations observed, and examine the combined

effect of childhood and adult social class on CHD.

Methods The British Regional Heart Study (BRHS) is a prospective study

of cardiovascular disease comprising a socially and geographi-

cally representative sample of men aged 40–59 years drawn

from one general practice in each of 24 towns representing all

major British regions in 1978–80. Baseline assessment included

standard cardiovascular risk factors measured at study entry in

1978–80. 10

Details of the BRHS are reported elsewhere. 11

In

1992, information on childhood social circumstances in addi-

tion to information on lifestyle factors was collected by postal

questionnaires. For this article, we use follow-up data on

morbidity and mortality from 31 October 1992 to 1 June 2004

and our main outcome of interest was fatal and non-fatal CHD.

Information on morbidity and mortality has been routinely

collected during the follow-up through general practice records

and the National Health Service Central Register, respectively.

Non-fatal myocardial infarction was defined by the presence of

at least two of—severe prolonged chest pain, ECG evidence of

myocardial infarction and cardiac-enzymes changes consistent

with myocardial infarction. This was ascertained by reviews of

general practitioner records. Information from death certificates

using the International Classification of Diseases, 9th revision

(ICD-9) was used to identify fatal myocardial infarction cases

as deaths with code 410–414 (equivalent to ICD 10th revision

codes I20–I25).

Childhood social class

Subjects were asked in the questionnaire about the kind of job

their father had done for the longest period of his (father’s)

life. This information was used to classify subjects into manual

[3752 (71%)] and non-manual [1436 (27%)] childhood social

class groups using the Office of Population Censuses and

Surveys Classification of Occupations (1980) social class coding

index manual.9,12 Two hundred and sixty-eight (5%) men who

did not report their father’s social class and 115 (2%) men

whose fathers’ longest-held occupation was the Armed Forces

were excluded from the analyses.

Childhood social circumstances

Besides father’s occupation, information was also collected on

childhood household amenities as a proxy for childhood social

circumstances and to enable a better assessment of early-life

socio-economic position. Subjects were asked if till they were

10 years old their home had a bathroom, hot water supply and

family car ownership.

Adult socio-economic status

Subjects’ own adult social class was based on the longest-held

occupation of each man recorded at the study entry and

classified using the Registrar Generals’ Social Class

Classification. In this analysis, we excluded men with longest-

held occupation in the Armed Forces, and categorized social

classes I, II, III non-manual into ‘non-manual social class’ and

III manual, IV and V as ‘manual social class’.

Adult behavioural risk factors

Detailed questions were asked about smoking (number of

cigarettes smoked and changes in smoking habits), alcohol

consumption (frequency and number of alcoholic drinks),

physical activity (frequency and type of activity) and body

weight.11,13,14 Body mass index (BMI) was calculated as body

weight/(height) 2 using measures of body weight (in kilograms)

and height (in metres) measured at the baseline examination.

The men were classified into groups based on their alcohol

intake—none, occasional, light, moderate and heavy. Heavy

drinking was defined as drinking more than six units (1 UK

unit ¼ 10 g) of alcohol daily or on most days in the week. In the questionnaire, subjects were also asked to report their pattern

of physical activity such as walking, cycling and other sporting

activities. Physical activity scores were assigned on the basis of

frequency and type of activity and the men were divided into

six groups: none, occasional, light, moderate, moderately

vigorous and vigorous. Subjects who reported none or

occasional activity were classified as ‘inactive’.

Statistical analyses

Cox proportional hazards model was used to calculate age-

adjusted hazard ratios (HRs) with 95% confidence intervals (CI)

for CHD for those of manual childhood social class compared

with those of non-manual childhood social class. We assessed

the proportionality assumption for the Cox models by carrying

out a test on the Schoenfeld residuals.15 The assumption was

found to be valid for the main explanatory variables—

childhood social class and childhood household amenities.

The model was then separately adjusted first for adult social

class (all six social groups), second for adult behavioural risk

factors and finally both for adult social class and behavioural

risk factors. Similar hazards ratios were calculated according to

childhood household amenities. HRs were also calculated

according to adult social class adjusted for childhood social

class and then behavioural risk factors. To assess the combined

effect of social class and to explore any interaction between

childhood and adult social class we categorized subjects into

four sub-groups according to both childhood and adult social

class—both childhood and adult non-manual social class;

childhood non-manual and adult manual social class; child-

hood manual and adult non-manual; and childhood and adult

manual social class. We calculated HRs for CHD for these

groups with those of childhood and adult non-manual social

class as the reference category. We also carried out a formal test

of interaction between childhood and adult social class.

To further explore the relationship of behavioural risk factors

with childhood and adult social class we calculated the

percentage of men who were current smokers, heavy drinkers,

inactive and obese (BMI 5 30 kg/m2) according to the

CHILDHOOD SOCIO-ECONOMIC CIRCUMSTANCES AND CORONARY HEART DISEASE 561

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from

sub-groups of childhood and adult social class. For the

adjustments, age and BMI were fitted as continuous variables.

Childhood social class (two levels), adult social class (six

levels), smoking (six levels), physical activity (five levels) and

alcohol intake (five levels) were fitted as ordinal variables.

Analyses were carried out using SAS version 8 and STATA

version 7.

Results Among 5552 men aged 64–83 by the end of the follow-up

period (55, 380 person-years), there were 645 (1.2% per

annum) cases of CHD (372 non-fatal cases, 244 fatal cases

and 29 had both a non-fatal and fatal event). Table 1 describes

the demographic characteristics of the subjects according to

childhood and adult social class. Eighty-five per cent of men of

adult manual social class had a manual childhood social class.

The proportion of men lacking childhood amenities and with

adverse behavioural risk factors was higher in those of manual

childhood social class and among those of manual adult

social class.

Table 2 shows the number of CHD events and the relative risk

of CHD according to childhood and adult social class groups.

Those of manual childhood social class had a greater risk of

CHD compared with those of non-manual childhood social

class. This effect, though modest in size, was statistically

significant when adjusted for adult social class. In an age-

stratified analysis, the association appeared to be present both

Table 2 Hazard ratios (HR) with 95% confidence intervals (CI) for coronary heart disease (CHD) according to childhood and adult social class and childhood household amenities in a prospective study of British men aged 52–73 years followed up from 1992 to 2004

Number of CHD events (%)

HR (95% CI) Adjusted for age

HR (95% CI) Adjusted for

age and adult social class

HR (95% CI) Adjusted for age and

adult behavioural risk factorsb

HR (95% CI) Adjusted for age, adult behavioural risk factorsb and adult social class

Childhood social class

Non-manual 136 (9) 1.00 1.00 1.00 1.00

Manual 460 (12) 1.34 (1.11–1.63) 1.26 (1.03–1.55) 1.19 (0.98–1.45) 1.19 (0.97–1.46)

Adult social class

Non-manual 244 (10) 1.00 1.00 1.00 1.00

Manual 377 (13) 1.28 (1.09–1.51) 1.16 (0.98–1.38)a 1.09 (0.92–1.30) 1.09 (0.91–1.33)a

Bathroom in childhood home

Yes 280 (10) 1.00 1.00 1.00 1.00

No 357 (13) 1.11 (0.95–1.29) 1.05 (0.89–1.23) 1.02 (0.87–1.20) 1.01 (0.85–1.20)

Hot water supply in childhood home

Yes 294 (11) 1.00 1.00 1.00 1.00

No 342 (13) 1.07 (0.91–1.25) 1.01 (0.86–1.18) 0.99 (0.85–1.17) 0.98 (0.82–1.16)

Family access to car in childhood

Yes 69 (11) 1.00 1.00 1.00 1.00

No 568 (12) 1.45 (1.15–1.85) 1.36 (1.07–1.74) 1.41 (1.10–1.79) 1.35 (1.04–1.75)

a Adjusted for childhood social class. b Adult behavioural risk factors included smoking, alcohol, physical activity and BMI.

Table 1 Demographic characteristics of subjects in the BRHS aged 52–73 in 1992 according to childhood and adult social class

Childhood social class Adult social class

Non-manual n ¼ 1436 (28%)

Manual n ¼ 3752 (72%) P-value

Non-manual 2385 (44%)

Manual 3017 (56%) P-value

Age—mean in years (SD) 61 (6) 62 (6) 0.62 61 (6) 62 (6) 0.15

Adult manual social class—n (%) 406 (15) 2351 (85) <0.0001 – – –

No bathroom in childhood home—n (%) 376 (14) 2288 (86) <0.0001 941 (34) 1812 (66) <0.0001

No hot water supply in childhood home—n (%) 363 (14) 2212 (86) <0.0001 889 (33) 1772 (67) <0.0001

No family access to car in childhood—n (%) 895 (21) 3436 (79) <0.0001 1792 (40) 2699 (60) <0.0001

Current smokers—n (%) 195 (20) 787 (80) <0.0001 308 (30) 729 (70) <0.0001

Heavy drinkers—n (%) 60 (29) 145 (71) 0.60 96 (45) 116 (55) 0.73

Physically inactive—n (%) 424 (26) 1215 (74) 0.05 689 (40) 1021 (60) 0.0001

Overweight (BMI 5 25 kg/m2)—n (%) 805 (26) 2289 (74) 0.001 1355 (42) 1854 (58) 0.0006

562 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from

in the older (>63 years) and younger subjects (<63 years) in

the cohort (data not shown). Including additional measures of

adult socio-economic position (car and house ownership) made

little difference to these estimates (data not shown). The effect

of childhood social class was diminished when adjusted for

adult behavioural risk factors (Table 2). Adjustment for both

adult social class and behavioural risk factors did not

substantially alter the effect estimates after adjustment for

behavioural risk factors alone (Table 2). Further adjustment for

town of birth or region of residence in adult life (at the time of

recruitment in the study) made very little difference to the

reported effect of childhood social class on CHD risk.

Adjustment for other cardiovascular risk factors measured at

baseline screening including blood cholesterol, blood pressure

and blood glucose also did not materially affect these results.

Men of manual adult social class had a greater risk of CHD

compared with non-manual, but the strength of this association

was weaker than that seen for childhood social class. The effect

of adult social class was diminished when adjusted for

childhood social class.

Table 2 also shows the HRs for CHD in relation to different

childhood household amenities. Those whose family did not

own a car had a higher CHD risk and this remained statistically

significant even after adjusting for adult social class and

behavioural risk factors. When we combined these childhood

social amenities and father’s social class in a score, this showed

no consistent evidence of a higher CHD risk with increasing

number of these adverse childhood social circumstances

(P ¼ 0.19). We examined the combined effect of social class in childhood

and adulthood on risk of CHD. Table 3 shows CHD rates per

1000 person-years and age-adjusted relative risk of CHD

according to childhood and adult social class with non-

manual childhood and adult social class group as the reference

group (group 1). CHD risks were lowest in this reference group

and highest in those with both childhood and adult manual

social class (group 4). Exposure to manual social class either in

childhood or as adults (groups 2 and 3) was also associated

with increased CHD risk. However, higher levels of current

smoking, physical activity and obesity were found in these

groups (Table 4) with the highest levels in those exposed to

both childhood and adult manual social class (group 4). Adult

manual workers had higher levels of smoking and physical

activity than all non-manual workers irrespective of childhood

social class (groups 2 and 4 vs groups 1 and 3); childhood social

class had an influence on obesity irrespective of adult manual

social class.

Table 3 also shows the relative risk for CHD adjusted for these

behavioural risk factors according to childhood and adult social

class. Adjustment for adult behavioural risk factors substan-

tially reduced the increased relative risk seen in those of both

childhood and adult manual social class (group 4); the

increased risk was of borderline significance (HR 1.28; 95% CI

0.99–1.65). The greater relative risks of CHD in those of manual

social class either in childhood or as adults (groups 2 and 3)

were also attenuated when adjusted for adult behavioural risk

factors. A test for interaction between childhood and adult

social class showed no evidence that the effect of childhood

social class was different in those of adult non-manual and

manual groups (P ¼ 0.48). There was no evidence that the relation of adult behavioural risk factors with CHD incidence

differed according to childhood social class (P-value for tests of

interaction for smoking ¼ 0.27, alcohol ¼ 0.17, physical activity ¼ 0.21, BMI ¼ 0.25). There was no association between

Table 4 Adult behavioural risk factors according to childhood and adult social classes in a prospective study of British men aged 52–73 years followed-up from 1992 to 2004

Groups according to childhood and adult social class Current smokers (n ¼ 931)

Heavy drinking (n ¼ 199)

Inactive (n ¼ 1589)

Obese (BMI530 kg/m2) (n ¼ 533)

Group Childhood Adult n (%) n (%) n (%) n (%)

1 Non-manual Non-manual 122 (12) 46 (5) 281 (29) 72 (7)

2 Non-manual Manual 65 (16) 13 (3) 131 (32) 38 (9)

3 Manual Non-manual 172 (13) 44 (3) 380 (29) 112 (9)

4 Manual Manual 577 (24) 98 (4) 801 (34) 311 (13)

Table 3 Coronary heart disease (CHD) rate per 1000 person-years and HR with 95% CI for CHD according to childhood and adult social classes in a prospective study of British men aged 52–73 years followed-up from 1992 to 2004

Groups according to childhood and adult social class

CHD rate per 1000 person-years

Hazard ratio (HR) (95%CI)

Hazard ratio (HR) (95%CI)

Group n Childhood Adult Age-adjusted Adjusted for age and adult

behavioural risk factorsa

1 985 Non-manual Non-manual 9 1.00 1.00

2 404 Non-manual Manual 11 1.25 (0.87–1.79) 1.18 (0.82–1.69)

3 1298 Manual Non-manual 11 1.33 (1.02–1.73) 1.29 (0.99–1.68)

4 2341 Manual Manual 13 1.51 (1.19–1.91) 1.28 (0.99–1.65)

aAdult behavioural risk factors included smoking, alcohol, physical activity and BMI.

CHILDHOOD SOCIO-ECONOMIC CIRCUMSTANCES AND CORONARY HEART DISEASE 563

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from

any of the childhood socio-economic measures and stroke (data

not shown).

Discussion We have previously shown that in middle-aged men

(52–74 years) childhood social class was related to prevalent

CHD independent of adult social class and behavioural risk

factors.9 In this report we extend our observations to incident

CHD in this population of older men, examine other measures

of socio-economic position in childhood as well as investigate

the combined effect of social class in early life and adulthood

with incident CHD. We observed that the relation of childhood

social class with subsequent CHD risk was diminished when

adult behavioural risk factors were taken into account. The

combination of manual childhood and adult social class

appeared to increase CHD risk further but this was attenuated

by adult behavioural risk factors. Lack of family car ownership

in childhood had an independent relationship with increased

CHD risk.

Our findings are consistent with previous studies of the

relationship between childhood social circumstances and

CHD. 16

Previous studies have indicated that less favourable

childhood socio-economic conditions were associated with

increased CHD risk; the size of the associations were also

mostly weak as seen in our study. 2,16

A small number of earlier

studies have taken into account the additional role of adult

behavioural risk factors (which can be important influences on

the relation between childhood social class and CHD), showing

that adjustment reduces the influence of childhood social

circumstances on CHD.5,6,8 Previous studies have also shown

that social circumstances in childhood were associated with

increased accumulation of adult risk factors17–19 and some have

reflected a similar difference in the influence of childhood and

adult social class on adult risk factors. 20–22

The absence of any

relation between most childhood household amenities (access

to bathroom, hot water and overcrowding) with CHD risk is

consistent with recent findings from the British Women’s Heart

and Health Study, which suggested that childhood infections as

a result of poor household conditions are unlikely to be an

important pathway to increased CHD risk in adulthood. 7

Our results are based on a population-based socioeconomi-

cally representative sample of men from across Britain, with

high rates of follow-up for morbidity and mortality. The

childhood social class measure was based on the longest-held

occupation of the father, which is likely to be a stable measure

of childhood social status, with social mobility in the father’s

generation probably less marked than among men in the

generation of our study, who were more influenced by widened

educational opportunities. A limitation of our study is that the

measures of childhood social class are based on retrospective

collection of information, raising the possibility both of random

error and recall bias. Recall bias, particularly a tendency to

overestimate social status, has been previously demonstrated

for father’s occupation, when compared with information

collected in early life. 23

However, the validity of the father’s

social class measurement is suggested by its strong relationship

with educational attainment of the subjects; a markedly lower

proportion of subjects with fathers in manual occupations were

educated after 18 years of age and a higher proportion left

education at 14 years. Recall bias is less likely to affect recall of

family amenities in childhood, particularly car ownership. The

accuracy of adult social class (based on occupation and used in

adjusted analyses) is also important. Our measure was based on

longest-held occupation recorded at study entry in 1978–80

when the subjects were aged 40–59 years. We have already

established that this measure was stable over a 20-year period,

changing in only a small proportion (<10%) of subjects.24

Moreover, the addition of other measures of adult social status

(including car ownership and housing tenure) had little effect

on the results. Our measures of social status after retirement

are however limited, allowing the possibility of some residual

social confounding.

The strength and statistical significance of the association of

childhood social class with CHD risk in our study was strongly

dependent on whether adjustment was made for adult social

class and, particularly adult behavioural risk factors. The

interpretation of these adjusted analyses depends on whether

adult social class and adult behavioural risk factors are

regarded as confounders of the childhood social class–CHD

association, or mediators of it. Childhood social status is

strongly related to adult social position, 17

and the adult

behavioural risk factors are themselves affected both by early

life and adult social conditions.17,21 In the present study,

obesity in particular was strongly related to childhood social

class, while the other behavioural risk factors were more

strongly influenced by adult social class. While we have to an

extent explored the different effects of childhood and adult

social class on risk factors, it was not possible in our study to

fully disentangle the issue of whether these risk factors are

mediators or confounders, though the former remains a strong

possibility, suggesting that unadjusted analyses may provide a

truer indicator of the association. In the case of family car

ownership in childhood, the association with CHD risk was

however substantially independent of adjustment, though (like

that for childhood social class) limited in strength. The

persistence of this effect of childhood circumstances in our

older population is noteworthy. There are possible reasons why

lack of family car ownership retained an independent relation-

ship with CHD. First, the question on family car ownership may

have been less prone to recall bias or misclassification

compared with other questions on father’s occupation or

other childhood amenities. Second, family car ownership

during the childhood of our subjects (approximately in the

1930s and 1940s) may be a better or stronger marker of

material wealth or social affluence. Thus, owning a car probably

discriminated the very affluent from the rest, something that

was not fully captured using the father’s occupation-based

social class distinction. It has been previously shown in our

study that in adults material wealth such as car ownership

discriminates mortality even within occupational social class

groups.25 A higher CHD risk in those lacking family car access,

not necessarily indicating poverty as such, probably reflects the

relative difference in wealth when compared with those who

had a family car. It is possible that the effect of adverse

childhood social circumstances would be more apparent using

more precise markers of resource income. This also highlights

that the strength of the association of childhood social

564 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from

circumstances with CHD in later life in observational studies

can differ according to the measures used to assess social class

in early life as was pointed out in a recent report.26

Different mechanisms have been postulated to understand

how childhood or early life factors affect health in later life. 27,28

Exposures acting during a specific period which influence the

development of chronic diseases forms the basis of the ‘critical

period model’ or ‘critical period with later effect modifiers’ if

modified by exposures in later life. The other pathway to

chronic diseases is the ‘accumulation of risk model’. According

to this model, adverse exposures accumulate over the lifecourse

gradually increasing the risk to worse adult health out-

comes.16,28 This accumulation of risk can either occur in a

dose-response fashion or through clustering of exposures such

as low birthweight, poor diet, lower educational attainment

which are all associated with poorer childhood social condi-

tions. 28

Adverse exposures can also accumulate by forming

chains of risk where one exposure increases the risk of another.

Although the present study does not allow discrimination

between these models, the results (particularly the combined

influence of childhood and adult social class) would be

consistent with a cumulative model of risk with social

circumstances at different stages of the lifecourse contributing

to overall risk. However, it remains possible that social

exposures, particularly early in life, are critical in their timing.

Cohorts with more measures at different stages of the

lifecourse are needed to fully substantiate the accumulation

of risk model.

Implications of our findings and conclusions Our results show that the effect of less affluent childhood social

circumstances on CHD risk persists in an older population.

Combined exposure to both adverse childhood and adult social

circumstances is associated with the most unfavourable lifestyle

behaviour. The findings add to the current literature since we

have shown that the influence of social conditions in child-

hood, though modest, persists in old age. Moreover, by this age

behavioural risk factors, which can have their origins in

childhood and adulthood, play an important role in developing

risk of CHD. Thus, as regards to public health policy, a dual

approach to improve childhood socio-economic circumstances

as well as to target social disparities in behavioural risk factors

in adult life will not only help reduce the burden from CHD in

older people but also narrow health inequalities.

Acknowledgements The British Regional Heart Study is a British Heart Foundation

Research Group and also received funding from the Department

of Health, England. SR is funded by a UK Medical Research

Council Special Training Fellowship in Health Services Research

and Health of the Public. The views expressed in this

publication are those of the authors and not necessarily those

of the funding bodies.

Conflict of Interest: None declared.

References 1 Lawlor DA, Ben-Shlomo Y, Leon DA. Pre-adult influences on

cardiovascular disease. In: Kuh D, Ben-Shlomo Y (eds). A Life

Course Approach to Chronic Disease Epidemiology. 2nd edn., Oxford:

Oxford University Press, 2004, pp. 41–76. 2 Galobardes B, Lynch JW, Davey Smith G. Childhood

socioeconomic circumstances and cause-specific mortality in adult-

hood: systematic review and interpretation. Epidemiol Rev

2004;26:7–21. 3 Naess O, Claussen B, Davey Smith G. Relative impact of childhood

and adulthood socioeconomic conditions on cause specific mortality

in men. J Epidemiol Community Health 2004;58:597–98. 4 Galobardes B, Davey Smith G, Jeffreys M, McCarron P. Childhood

socioeconomic circumstances predict specific causes of death in

adulthood: the Glasgow student cohort study. J Epidemiol Community

Health 2006;60:527–29. 5 Gliksman MD, Kawachi I, Hunter D et al. Childhood

socioeconomic status and risk of cardiovascular disease in middle

aged US women: a prospective study. J Epidemiol Community Health

1995;49:10–15. 6 Hart CL, Davey Smith G. Relation between number of siblings

and adult mortality and stroke risk: 25 year follow up of men

in the Collaborative study. J Epidemiol Community Health

2003;57:385–91. 7 Lawlor DA, Ebrahim S, Davey Smith G. Adverse socioeconomic

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;59:785–93. 8 Smith GD, Hart C, Blane D, Hole D. Adverse socioeconomic

conditions in childhood and cause specific adult mortality: prospec-

tive observational study. Br Med J 1998;316:1631–35. 9 Wannamethee SG, Whincup PH, Shaper G, Walker M. Influence of

fathers’ social class on cardiovascular disease in middle-aged men.

Lancet 1996;348:1259–63. 10 Shaper AG, Pocock SJ, Walker M, Cohen NM, Wale CJ, Thomson AG.

British Regional Heart Study: cardiovascular risk factors in middle-

aged men in 24 towns. Br Med J 1981;283:179–93.

KEY MESSAGES

� Less affluent childhood socioeconomic conditions may have a persistent effect on CHD risk in old age.

� Combined exposure to adverse childhood and adult social circumstances is associated with the most unfavourable lifestyle behaviour and CHD risk.

� Adult behavioural risk factors play an important role in influencing the relation of childhood social factors with CHD risk in later life.

CHILDHOOD SOCIO-ECONOMIC CIRCUMSTANCES AND CORONARY HEART DISEASE 565

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from

11 Walker M, Whincup PH, Shaper AG. The British Regional Heart

Study 1975-2004. Int J Epidemiol 2004;33:1185–92. 12

Office of Population Censuses and Surveys. Classification of occupations and

coding index. London: HM Stationery Office, 1980. 13

Wannamethee SG, Lowe GDO, Whincup PH, Rumley A, Walker M,

Lennon L. Physical Activity and Hemostatic and Inflammatory

Variables in Elderly Men. Circulation 2002;105:1785–90. 14

Shaper AG, Pocock SJ, Walker M, Cohen NM, Wale CJ, Thomson AG.

British Regional Heart Study: cardiovascular risk factors in middle-

aged men in 24 towns. Br Med J 1981;283:179–93. 15 Schoenfeld D. Partial residuals for the proportional hazards regres-

sion model. Biometrika 1982;69:239–41. 16

Smith GD, Lynch J. Life course approaches to socioeconomic

differentials in health. In: Kuh D, Ben-Shlomo Y (eds). A Life

Course Approach to Chronic Disease epidemiology. 2nd edn., Oxford:

Oxford University Press, 2004, pp. 77–115. 17

Power C, Matthews S. Origins of health inequalities in a national

population sample. Lancet 1997;350:1584–89. 18 Kivimaki M, Davey Smith G, 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;92:474–80. 19

van de Mheen H, Stronks K, Looman CWN, Mackenbach JP. Does

childhood socioeconomic status influence adult health through

behavioural factors? Int J Epidemiol 1998;27:431–37. 20

Blane D, Hart CL, Smith GD, Gillis CR, Hole DJ, Hawthorne VM.

Association of cardiovascular disease risk factors with socioeconomic

position during childhood and during adulthood. BMJ

1996;313:1434–38. 21

Brunner E, Shipley MJ, Blane D, Smith GD, Marmot MG. When does

cardiovascular risk start? Past and present socioeconomic circum-

stances and risk factors in adulthood. J Epidemiol Community Health

1999;53:757–64. 22

Power C, Graham H, Due P et al. The contribution of childhood and

adult socioeconomic position to adult obesity and smoking behaviour:

an international comparison. Int J Epidemiol 2005;34:335–44. 23

Batty GD, Lawlor DA, Macintyre S, Clark H, Leon DA. Accuracy of adults’

recall of childhood social class: findings from the Aberdeen children

of the 1950s study. J Epidemiol Community Health 2005;59:898–903. 24 Emberson JR, Whincup PH, Morris RW, Walker M. Social class

differences in coronary heart disease in middle-aged British men:

implications for prevention. Int J Epidemiol 2004;33:289–96. 25 Wannamethee SG, Shaper AG. Socioeconomic status within social

class and mortality: a prospective study in middle-aged British men.

Int J Epidemiol 1997;26:532–41. 26

Kauhanen L, Lakka HM, Lynch JW, Kauhanen J. Social

disadvantages in childhood and risk of all-cause death and

cardiovascular disease in later life: a comparison of historical and

retrospective childhood information. Int J Epidemiol 2006;35:962–68. 27

Kuh D, Ben-Shlomo Y. Introduction. In: Kuh D, Ben-Shlomo Y (eds).

A Lifecourse Approach to Chronic Disease Epidemiology. 2nd edn., Oxford:

Oxford University Press, 2004, pp. 3–14. 28

Lynch J, Smith GD. A life course approach to chronic disease

epidemiology. Annu Rev Public Health 2005;26:1–35.

Published by Oxford University Press on behalf of the International Epidemiological Association

� The Author 2007; all rights reserved. Advance Access publication 17 May 2007 International Journal of Epidemiology 2007;36:566–568

doi:10.1093/ije/dym099

Commentary: Selected samples and nebulous measures: some methodological difficulties in life-course epidemiology M Maria Glymour

Accepted 17 April 2007

Introduction Ramsay et al.’s article1 contributes to a growing body of research

on early life characteristics that predict health in adulthood.2–4

This research demonstrates that adults who lived in deprived

socio-economic circumstances as children are more likely to suffer

from cardiovascular disease than adults who had privileged

childhoods; an association that holds even among people with

comparable indicators of adult socio-economic position (SEP).

These articles provide a welcome impetus to consider how adult

health is shaped by early life experiences, but they face a number

of methodological problems that compromise causal inference

regarding the effects of childhood social conditions on adult

health. In this commentary, I wish to focus on just two of these

problems: loss to follow-up and inadequate measurement.

Throughout, I assume that the primary causal question of

interest is how adult health would differ if we intervened to

change childhood SEP, and a secondary causal question is how

Department of Epidemiology, Mailman School of Public Health at Columbia University, 722 W. 168th St room 1603, New York, New York 10032. E-mail: [email protected] Maria Glymour is a Robert Wood Johnson Foundation Health and Society Scholar at Columbia University.

566 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

by guest on A pril 19, 2014

http://ije.oxfordjournals.org/ D

ow nloaded from