IJE Advance Access originally published online on March 21, 2006
International Journal of Epidemiology 2006 35(5):1272-1277; doi:10.1093/ije/dyl045
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Genetics |
Genetic and environmental influences on the relation between parental social class and mortality
1 Institute of Public Health, University of Copenhagen, Blegdamsvej 3, 2200 Copenhagen N, Denmark
2 Institute of Public Health, University of Southern Denmark, JB Winsløwsvej 9B, 5000 Odense C, Denmark
3 Danish Epidemiology Science Centre, Institute of Preventive Medicine, Copenhagen University Hospital, Centre for Health and Society, Copenhagen, Denmark
4 Department of Psychology, University of Copenhagen, Denmark
* Corresponding author. E-mail: MOsler{at}health.sdu.dk
 |
Abstract |
|---|
 Top Abstract Material and methodsResultsDiscussionReferences |
|---|
Background Genetic and maternal prenatal environmental factors as well as the post-natal rearing environment may contribute to the association between childhood socioeconomic circumstances and later mortality. In order to disentangle these influences, we studied all-cause and cause-specific mortality in a cohort of adoptees, in whom we estimated the effects of their biological and adoptive fathers' social classes as indicators of the genetic and/or prenatal environmental factors and the post-natal environment, respectively.
Methods In all 12 608 children born 1924–47 in Denmark who were placed early in life with adoptive parents were followed up for causes of death until 2000. Hazard ratios for paternal social class retrieved from adoption records were estimated using Cox regression models.
Results Adoptees with biological fathers from higher social classes had a lower rate of mortality after their fifth decade of life, mainly due to a lower risk of cardiovascular, infectious, and respiratory diseases. Adoptive father's social class showed no clear relation with adoptee's mortality risk. The risk estimates for paternal social class were slightly attenuated after adjustment for adoptee's adult social class, which as expected was inversely related to mortality from both natural and external causes.
Conclusion Genetic and/or prenatal environmental factors contribute to the development of the relation of paternal social class to mortality from natural causes later in adult life independently of the effect of own social class, whereas there is no evidence for such long-term effect of the rearing environment.
Keywords Adoption cohort, parental social class, mortality
Accepted 23 February 2006
A number of cohort studies have suggested that childhood socioeconomic circumstances contribute to different causes of deaths.1 In particular, childhood conditions appear related to mortality from natural causes, especially stomach cancer, respiratory and cardiovascular diseases.1–7 The association between childhood socioeconomic circumstances and mortality could come about through processes that may be divided into three groups related to genetic factors, prenatal environmental exposures, and the family environment during childhood. The genetic and prenatal theories postulate that genes or prenatal exposures, such as impaired intrauterine growth, link the socioeconomic position of parents to the health of their adult offspring. The family environment theory addresses the impact of material conditions, behaviour, and social support of the home in which the child is raised. The effect of these processes might act both independently of each other or be mediated through the children's own social position later in life.
Studies of adoptees, who have been separated from their biological parents shortly after birth and transferred without delay to adoptive parents genetically unrelated to them, offer a unique opportunity to isolate the effect of the environment and explore the above theories. In an adoption study a relationship between an adoptive parent's socioeconomic position and mortality of the adopted offspring would argue in favour of family environmental mechanisms. An association between the biological parent's socioeconomic position and mortality of the adopted adult offspring would be consistent with genetic and prenatal influences.
In order to disentangle the influences of the rearing environment from genetic and/or prenatal environmental factors we followed up a cohort of adult adoptees and analysed the effect of biological and adoptive fathers' social classes on adoptees' mortality from natural and external causes. We also explored whether any such effects were mediated through adoptee's own social class later in life.
|
Material and methods |
|---|
|
TopAbstractMaterial and methodsResultsDiscussionReferences |
|---|
Study population
The study is based on the Danish Adoption Register. This register includes all 14 425 adoptions granted in Denmark from 1924 through 1947 in which neither adoptive parent was a biological relative of the child.8 The Register contains the identity of the biological parents, the adoptive parents, and the adoptee, as well as details of the circumstances of the adoption, including the age at which the adoptee was transferred to the adoptive home. Ninety per cent were transferred to the adoptive parents within the first year of life. Children who were not transferred to adoptive parents immediately after birth usually did not remain with their biological parents but were instead placed in foster homes. The adoptions were arranged anonymously through an authorized intermediary, and there was very little, if any, contact between the adoptees and their biological family later in life. The following is a brief description of the methods, which have been reported in detail elsewhere.9–11 We selected the 12 608 adoptees (5954 males and 6654 females) born during the period 1924 through 1947. In total, 1078 adoptees born before 1924, 78 twins and 661 children lost to follow-up, were not included. Those born before 1924 were excluded because they represented a group with particularly late dates of formal adoption.
Measures of socioeconomic position
The information on biological and adoptive fathers' occupations together with the annual income of the adoptive family at time of the adoption was retrieved from the adoption records.12 This information was available for all adoptive fathers and 80% (n = 10 086) of biological fathers. In 1976 information on current or latest occupation for the adoptees themselves was obtained as recorded from annual tax returns in the Danish Civil Registration System for the period 1969–76. At this time they had a median age of 34.5 (range 28.2–50.9) years. This information was available for 92% (n = 5507) of men and 51% (n = 3408) of women. The occupations were coded on a scale of social class derived from a Danish study of prestige ratings of occupational titles.13 The scale ranges from 0 = low positions (unskilled manual workers) to 7 = high positions (advanced professional). Details of the scale have been reported earlier.12 The initial data analyses showed that a relatively small number of biological parents belonged to the highest social classes (7 and 6) and the risk estimates were nearly the same for the fifth category; therefore, these three categories were merged. The estimates for the four lower social classes were such that they were also combined. This left two categories: high/middle class (advanced professional, academic and sub-academic professions, and self-employed) and working class (subordinate clerk, skilled worker, semiskilled worker, and unskilled worker). The 3050 adopted women who were alive in 1969 and had no information on current occupation were coded as housewives. Adoptive family's household income was used as an additional social indicator of the rearing environment. This variable was analysed in quartiles adjusted by year of adoption in order to correct for inflation.
The age at which the adoptee was transferred to the adoptive home, usually occurring some time before formal adoption, might influence the effects of adoptive father's social class as a measure of the rearing environment14 and consequently, adoptee's age at transfer was included in the adjusted models. This variable was reported in months and recoded into five categories (0, <1, 1–4, >4 years, and missing).
Follow-up
On the basis of the information in each adoption record, the adoptees were traced in population registers and followed up for vital status until November 2000. Of the 12 608 adoptees 42 had disappeared, 352 had emigrated, and 2622 had died during follow-up. The deaths were divided into early (age 0–50 years), which included a major proportion (75%) of the external causes, and later (age 
50 years) in which period natural causes pre-dominated (67%). Biological and adoptive fathers' vital status had also been traced for a random sample of 1641 adoptive fathers and 1491 biological fathers of whom 554 and 605, respectively, had died during follow-up. We also identified causes of death for the period 1969–98 by record linkage to the Cause-of-Death Registry. Causes of death were based on official death certificates coded using the 8th Revision of the International Classification of Diseases for the year 1969–93 and the 10th Revision for 1994–98. Causes of death were classified into the following groups: all natural causes, infectious diseases, respiratory diseases (which included pneumonia), cardiovascular diseases, cancers, all external causes (injury, suicide, and homicide), and suicide.
Statistical analysis
Associations between measures of social class and mortality for cohort members were analysed using Cox's proportional hazard regression models. Models were stratified by gender and birth cohort in seven groups (1923–26, 1927–29, 1930–33, 1934–37, 1938–40, 1941–43, and 1944–47), so that each stratification group has a separate underlying hazard but was assumed to have the same effect of social class. Age was used as the underlying time scale with adoptee's age at the time of adoption as entry time in models exploring the effects of paternal social class. All the adjusted models included the social class of both the biological and adoptive fathers together in order to adjust for any bias originating in selective placement of the adoptees. The information on the adoptee's occupational social class was retrieved from the Civil Registration System, which was established in April 1968, and the social class information was thus available only for adoptees, who were alive in 1968. We, therefore, used adoptee's age in 1969 as entry time, and the analysis of models where adoptee's own social class was included was based on the 2105 deaths that occurred after this year. Follow-up for total mortality ended at the age of death, emigration, or November 30, 2000 whichever came first. The proportional hazard assumption was tested in two ways: the standard graphical check based on the log of cumulative hazard and through formal tests of proportionality based on Schoenfeldt residuals. Statistical interaction between parental social class and adoptee's social class was evaluated using likelihood ratio tests. The best evidence on separating social effects will come from the children adopted at very early ages and all analyses were repeated for the 90% transferred under 1 year of age. All statistical analyses were performed using STATA version 8.15
|
Results |
|---|
|
TopAbstractMaterial and methodsResultsDiscussionReferences |
|---|
Working class adoptees more often had biological and adoptive fathers belonging to the working class compared with adoptees from the high/middle social class (Table 1).
|
Adoptive father's social class at the time of the adoption was not associated with adoptee's total mortality before or after age 50 years (Table 2). Adoptive father's social class showed a marginal inverse relationship with adoptee's death from natural causes and cardiovascular diseases. Adoptive father's social class was positively related to adoptee's risk of dying from external causes, apparently due to a higher risk of suicide. In the five-category scale of social class, adoptees with adoptive fathers from the highest social class (academic professions) had a 2.31 (1.35–3.94) higher risk of suicide compared with adoptees with adoptive fathers in the lowest social class (unskilled workers). We also found that adoptees with fathers in the upper income quartile had 1.52 (1.20–1.94) higher risk of dying before age 50 years than those with adoptive fathers in the lowest quartile (Table 3). A similar but weaker tendency was seen for deaths from external causes and suicide (data not shown).
|
|
The adoptee's own social class was inversely related to mortality before and after age 50 years as well as with all the specific causes (Table 4). The estimates for paternal social class changed only slightly when adoptee's own social class was included in the model.
|
Cancer mortality was not related to paternal or adoptee's own social class, neither before nor after adjustments. There were no statistical interactions.
Compared with those transferred immediately after birth, adoptees transferred during the first year of life had higher mortality before age 50 years [Hazard ratio 1.24 (1.04–1.37)], while the risk of dying later in life was not related to age at transfer. There were no notable social class differences in age at transfer, and it did not attenuate the risk estimates for social class and mortality. The estimates from the analyses restricted to adoptees transferred the first year was essentially the same as estimates presented based on all adoptees (data not shown).
In a sub-cohort we explored the ability of the biological and adoptive fathers' social classes to predict their own mortality. This showed a tendency for an inverse relation between adoptive father's social class and mortality, while the social class of the biological father was not associated with mortality risk.
|
Discussion |
|---|
|
TopAbstractMaterial and methodsResultsDiscussionReferences |
|---|
This study found that adoptees with biological fathers from higher social classes at the time of adoption had a lower rate of mortality after their fifth decade of life mainly owing to a lower risk of cardiovascular, infectious, and respiratory diseases. Adoptive father's social class showed no clear relation with adoptee's mortality risk with the single exception that adoptees having adoptive fathers from academic positions had an elevated suicide risk. Adoptive father's income level was positively associated with risk of dying before age 50 years. The effects for paternal social class were only slightly attenuated after adjustment for adoptee's adult social class, which in itself, as expected, was inversely related with mortality in both early and later adult life.
The inverse relation of biological father's social class to adult offspring mortality of cardiovascular, infectious, and respiratory diseases, which act independently of the subject's own social class later in life, suggests an underlying common genetic or a prenatal environmental mechanism. The genetic effects could operate through pleiotropic genes affecting both characteristics. Thus, there are genetic effects for social class12 as well as for mortality16 especially from the causes that were associated with biological father's social class in the present study. It is also possible that the mechanism is genetic factors leading to poor health, which itself is a barrier for attaining a high social status. The fact that the relation between biological father's social class and mortality of the adopted offspring was only seen for deaths occurring later in life could also support the finding of genetic effects developing over time previously noted for educational level17 and body mass index18 in the same cohort. Inherited personality traits influencing both social class and health-related behaviours, and risk factors could be mediators of the observed effects. Intelligence (IQ) could also be such a trait and has been suggested as the mechanism behind the association between biological father's social class and adoptee's own social class.19 A number of studies have shown an inverse relationship between childhood IQ and mortality in adult life.7,20 In a recent Danish study of men born in 19537 father's social class was associated with mortality both before and after age 35 years, while an inverse relation with IQ only remained clear for the later deaths. These findings suggest that IQ could be the mediator of genetic effects. Another mechanism behind the association is a genetic component in smoking, a known risk factor that varies with socioeconomic circumstances.21
An alternative hypothesis is that this reflects a prenatal programming effect. Mother's lifestyle and health during pregnancy may determine her child's fetal development and later adult disease. One explanation as to why biological father's social class may reflect this is because of assortative mating; men of higher social classes are likely to marry women of higher social classes and, hence, have healthier babies. In this scenario, father's social class is a proxy marker of maternal social class. A relatively high correlation between biological father's social class and biological mother's social class in the present study12 suggest this could be a pathway. On the other hand the biological mothers were often not sharing the family environment with the biological father as a reflection of the reason for giving the child up for adoption, and the effect of biological father's social class was not influenced by adjustment for social class of the biological mother. However, the social class information on biological mothers of adoptees was frequently poorly defined or unavailable.
The absence of a clear inverse association between adoptive father's social class and adoptee mortality from natural causes suggests that family environment factors such as material living conditions, parental behaviour, and support are not important mechanisms in themselves within the range observed here. This is contrary to studies of intact families showing that different measures of family deprivation predict adult mortality and account for some of the relationship between father's social class and mortality.1,22
The elevated risk found for adoptees with adoptive fathers from the highest social class in the present study could be a chance finding or reflect special conditions for adoptees. A few other studies have reported on associations between childhood socioeconomic circumstances and risk of suicide in adulthood. Barraclough et al.23 found that suicides without preceding mental illness were more frequent in males of a better social class than suicides with preceding mental illness. However, later cohort studies from the UK24 and Finland3 showed that suicide was more likely to occur among those with fathers of lower social class. If the result of the present study is true, it suggests that the environment in adoptive families in the highest social class is in some way unfavourable or challenging for the adoptees.
As expected, the observed effect for the biological father's social class was rather small. On the other hand, the results should be seen against the background of the relatively small risk estimates of 
0.70 found for adults with fathers of higher social classes in intact families.1,4,7 In accordance with previous studies,1 we found no evidence for an association between paternal social class and overall cancer mortality. However, the small number of deaths from stomach cancer, which has been strongly related to early life social circumstances,1,4 did not allow analyses with sufficient power. The present study meets the important requirements for valid assessment of the separate effects of the genetic and prenatal environmental factors, and the post-natal family environment on adoptee's health, whereas the opportunities for assessment of the isolated effects of the prenatal social environment are weak. The suitability of the adoption cohort for addressing these questions has been demonstrated in several previous studies.10–12,17–19 There was evidence of only some small degree of selective placement of the adoptees in the matching of the social class of their biological and adoptive parents.12 This might potentially mask an independent effect of the biological parent's social class, but in the present cohort a simple correlation between biological father's social class and adoptive father's social class (in eight categories) was weak (r = 0.18),12,17,19 and, moreover, the adjusted analysis included both biological and adoptive fathers' social classes.
Some possible important differences between adoptive and intact biological families must also be considered. Parents who adopted might have been chosen because they possess additional family resources and better living conditions, while parents who give up their child might have fewer resources than usual for the social class to which they belong. Furthermore, the social classes of biological and adoptive fathers were somewhat below and above, respectively, those of the general population for their generation, and for the biological fathers there was some evidence of reduced variance.12 Thus, the study might not enable us to estimate effects of the full range of social class. However, in a small subsample of our cohort, there was no significant association between social class and mortality for the adoptive or biological father. However, both the adoptive and biological fathers' social classes have been predictive for other outcomes such as adoptee's educational level,17 IQ,18 and obesity in previous studies based on larger subsamples of the same cohort.16–18
The information on social class was missing for 20% of biological fathers, but the risk estimates found for this did not differ compared with the risk estimates for working class fathers. This might reflect that these fathers are not without occupation, but are those who had little or no contact to the biological mothers, wherefore they could not report on his social class during the adoption process.
In conclusion, genetic and possibly prenatal environmental factors contribute to the relationship of paternal social class to mortality from natural causes later in adult life in particular from respiratory and cardiovascular diseases. These influences were independent of the effect of own social class, whereas it remains uncertain whether the parental social class has any separate prenatal environmental influence. There is no evidence for a long-term effect of the environment, as indicated by the social class of the adoptive father, on the risk of dying from any of the major causes of death, except perhaps for the risk of suicide.
Conflicts of Interests: None.
KEY MESSAGES
|
|
References |
|---|
|
TopAbstractMaterial and methodsResultsDiscussionReferences |
|---|
1 Garlobardes B, Lynch JW, Davey Smith G. Childhood socioeconomic circumstances and causespecific mortality in adulthood: Systematic review and interpretation. Epidemiol Rev 2004;26:7–21.
2 Vagerø D, Leon D. Effect of social class in childhood and adulthood on mortality. Lancet 1994;343:1224–25.[ISI][Medline]
3 Pensola TH, Martikainen P. The effect of parental social class, own education and social class on mortality among young men. Eur J Public Health 2002;12:29–36.
4 Davey Smith G, Hart C, Blane D, Hole D. Adverse socioconomic conditions and cause specific adult mortality: prospective observational study. BMJ 1998;316:1631–35.
5 Frankel S, Davey Smith G, Gunnell D. Childhood socioeconomic position and adult cardiovascular mortality: the Boyd Orr cohort. Am J Epidemiol 1999;150:1081–84.
6 Claussen B, Davey Smith G, Thelle D. The impact of childhood and adulthood socio-economic position on cause-specific mortality: the Oslo mortality study. J Epidemiol Community Health 2003;57:40–45.
7 Osler M, Nybo Andersen A, Due P, Lund R, Damsgaard MT, Holstein BE. Socioeconomic position in early life, birth weight, childhood cognitive funtion, and adult mortality. A longitudinal study of Danish men born in 1953. J Epidemiol Community Health 2003;57:681–86.
8 Kety SS, Rosenthal D, Wender PH, Schulsinger F. The types and prevalence of mental illness in the biological and adoptive families of adopted skizophrenics. J Psychiatr Res 1967;6(Suppl 1):345–62.[Medline]
9 Stunkard AJ, Sørensen TIA, Hanis C et al. An adoption study of human obesity. New Engl J Med 1986;314:193–98.[Abstract]
10 Sørensen TIA, Price RA, Stunkard AJ, Schulsinger F. Genetics of obesity in adult adoptees and their biological siblings. BMJ 1989;298:87–90.[ISI][Medline]
11 Sørensen TIA, Holst C, Stunkard AJ. Adoption study of environmental modifications of the genetic influences on obesity. Int J Obes 1998;22:73–81.[CrossRef][ISI][Medline]
12 Teasdale TW. Social class correlations among adoptees and their biological and adoptive parents. Behav Genet 1979;9:103–10.[CrossRef][ISI][Medline]
13 Svalastoga K. Prestige, Class and Mobility. Copenhagen: Gyldendal, 1959.
14 Eldred CA, Rosenthal D, Wender PH et al. Some aspects of adoption in selected sampels of adult adoptees. Am J Orthopsychiatry 1976; 46:279–90.[ISI][Medline]
15 Stata Corporation. Stata Statistical Software: Release 8.0. College station, TX: Stata Press, 2003.
16 Sørensen TIA, Nielsen GG, Andersen PK, Teasdale TW. Genetic and environmental influences on premature death in adult adoptees. N Engl J Med 1988;318:727–32.[Abstract]
17 Teasdale TW, Sørensen TIA. Educational attainment and social class in adoptees: genetic and environmental contributions. J Biosoc Sci 1983;15:509–18.[ISI][Medline]
18 Teasdale TW, Sørensen TIA, Stunkard AJ. Genetic and early environmental components in sociodemographic influences on adult body fatness. BMJ 1990;300:1615–18.[ISI][Medline]
19 Teasdale TW. Familial influences in social class, educational level and intelligence. Disertation, Gleerup, Malmø, Sweden, 1985.
20 Batty GD, Deary IJ. Early life intelligence and adult health. BMJ 2004;329:585–86.
21 Osler M, Holst C, Prescott E, Sørensen TIA. Influences of genes and family environment on adult smoking behaviour assessed in an adoption study. Genet Epidemiol 2001;21:1993–2000.
22 Osler M, Nybo Andersen AM, Batty GD, Holstein B. The relation between early life socio-economic position and all cause mortality in two generations. A longitudinal study of Danish men born in 1953 and their parents. J Epidemiol Community Health 2005;59:38–41.
23 Barraclough B, Bunch J, Nelson B, Sainbury P. A hundred cases of suicide: clinical aspects. Br J Psychiatry 1974;125:355–73.
24 Neeleman J, Wessely S, Warthwood M. Predictors of suicide, accidental death and premature natural deaths in a general population birth cohort. Lancet 1998;357:93–97.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  B Galobardes, J W Lynch, and G D. Smith Is the association between childhood socioeconomic circumstances and cause-specific mortality established? Update of a systematic review J. Epidemiol. Community Health, May 1, 2008; 62(5): 387 - 390. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D Gimeno, J E Ferrie, M Elovainio, L Pulkki-Raback, L Keltikangas-Jarvinen, C Eklund, M Hurme, T Lehtimaki, J Marniemi, J S A Viikari, et al. When do social inequalities in C-reactive protein start? A life course perspective from conception to adulthood in the Cardiovascular Risk in Young Finns Study Int. J. Epidemiol., April 1, 2008; 37(2): 290 - 298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. P Casas Commentary: Social class, C-reactive protein and coronary heart disease Int. J. Epidemiol., April 1, 2008; 37(2): 299 - 300. [Full Text] [PDF]
|
|
|
|
|
|
S. Ebrahim The riches of cohorts Int. J. Epidemiol., April 1, 2008; 37(2): 223 - 224. [Full Text] [PDF]
|
|
|
|
|
|
M Osler, A-M Nybo Andersen, and M Nordentoft Impaired childhood development and suicidal behaviour in a cohort of Danish men born in 1953 J. Epidemiol. Community Health, January 1, 2008; 62(1): 23 - 28. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Melchior, T. E. Moffitt, B. J. Milne, R. Poulton, and A. Caspi Why Do Children from Socioeconomically Disadvantaged Families Suffer from Poor Health When They Reach Adulthood? A Life-Course Study Am. J. Epidemiol., October 15, 2007; 166(8): 966 - 974. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||