IJE Advance Access originally published online on January 30, 2006
International Journal of Epidemiology 2006 35(3):665-670; doi:10.1093/ije/dyi321
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Article |
The association between cognitive ability measured at ages 18–20 and mortality during 30 years of follow-up—a prospective observational study among Swedish males born 1949–51
1 Division of Occupational Health, Department of Public Health Sciences, Karolinska Institute, Stockholm, Sweden
2 National Institute for Working Life, Stockholm, Sweden
3 Department of Psychology, Stockholm University, Sweden
4 Division of Social Medicine, Department of Public Health Sciences, Karolinska Institute, Stockholm, Sweden
* Corresponding author: National Institute for Working Life, SE-113 91 Stockholm, Sweden. E-mail: tomas.hemmingsson{at}niwl.se
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Objectives An association between childhood cognitive ability measured with IQ tests and mortality has been reported recently. It is not clear from those studies if the relative risk is increased only among those in the lower end of the IQ score scale or if there is graded increase in mortality from the lowest to the highest. This study aims to investigate the association between cognitive ability measured at age 18–20 and mortality during a 30 year period of follow-up.
Methods Data on cognitive ability was collected from 49 323 men, born in 1949–51, who were conscripted for compulsory military training in 1969/70. Data on mortality were obtained from the Causes of Death register 1971–2000.
Results Cognitive ability was a strong predictor of all-cause mortality, cardiovascular disease (CVD)-mortality, mortality from violent causes, and alcohol-related mortality. A striking finding was a pronounced gradient in mortality risk across all IQ score groups. Adjustment for adult socioeconomic position attenuated the increased risk somewhat [for all-cause mortality: crude hazard ratio (HR) 1.16 (1.13–1.19), adjusted HR 1.12 (1.09–1.15)].
Conclusion IQ test score measured in late adolescence (only males) was a significant predictor of all-cause, as well as cause-specific (CVD and injuries), mortality during 30 years of follow-up. The risk increased from high to low IQ test score results for all outcomes.
Keywords IQ, cognitive ability, mortality, socioeconomic group, cohort
Accepted 3 January 2006
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The association between childhood cognitive ability and mortality in old age1 and in mid-life2–7 has recently been documented in studies based on populations born 1921–55. In most studies cognitive ability refers to outcomes of psychometrically validated instruments aimed at measuring IQ. In two studies the association between cognitive ability and mortality was markedly stronger for those in the lowest IQ quartile than in the other quartiles,5,6 while in one study a graded increase of mortality was observed from high to low IQ test score.7 Those studies were rather small and the inconsistency between studies may be due to a lack of power. Previous studies on the subject have mainly focused on all-cause mortality, probably due to a limited number deaths in those studies.
Based on findings from the British 1946 birth cohort, Kuh et al.6 suggest that the observed relation between low cognitive ability and mortality might be explained by a greater cumulative exposure to poor life-time socioeconomic conditions among those in the lowest quarter of IQ test scores. In other studies adjustment for differences concerning indicators of a poor childhood environment had only a small effect on the risk estimates.3,7
Some causes of death, e.g. cardiovascular disease (CVD) mortality, are associated with early and later life circumstances,8,9 while others, e.g. mortality from injuries, are more strongly associated with adult circumstances.10
In the present study, we aim to investigate the association between IQ test score, measured at ages 18–20, and mortality, all-cause and cause-specific, during 30 years of follow-up in a cohort of 49 323 middle-aged Swedish males, born in 1949–51. With a study population of this size we have the possibility to investigate the following two questions:
- Is there an association between cognitive ability and mortality in this population and, if so, is the relative risk of mortality increased only among those in the lower end of the IQ score scale or is there a graded increase in mortality from the lowest to the highest scores?
- Is the strength of the association between IQ test scores and mortality similar for three major causes of death, i.e. mortality from cancer, CVD mortality, violent deaths, and alcohol related mortality?
We used data on fathers' socioeconomic position (SEP) from the 1960 census as a proxy for influences from social circumstances in childhood and data on adult SEP from the 1985 census, both measures associated with all-cause and CVD mortality in this material,8 as a proxy for influences from social circumstances in adult life. Cognitive ability is associated with achieved SEP in this material.11 Other potential risk factors have different associations with the included causes of mortality and will not be investigated in this study.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Study population
The study was based on data from a nation-wide survey of 49 323 Swedish males, born 1949–51, who were conscripted for compulsory military service in 1969/70. The background of the Swedish conscription surveys and the variables included have been presented in detail elsewhere.12,13At that time, only 2–3% of all Swedish men were exempted from conscription, in most cases owing to severe handicaps or congenital disorders. Those included in this study accounted for 97.7% of all conscripts in 1969/70, and the remaining 2.3% were born before 1949.
Information collected at the 1969–70 conscription examination
The IQ tests performed included tests on logic/general intelligence; verbal test of synonym detection; tests of visuospatial/geometric perception; and technical/mechanical skills with mathematical/physics problems that have been described in detail elsewhere.14,15 The outcome of each test was ranked 1–9.14 These standard-nine values were transformed into a new standard-nine scale as a measure of general ability and corresponding to approximate IQ bands of: <74, 74–81, 82–89, 90–95, 96–104, 105–110, 111–118, 119–126, >126.15 Of the men 49 262 (99.9%) had a score on mental ability. The IQ test used is included in the Swedish Enlistment Battery (SEB) that has been used in continuously updated versions since 1944 to classify 18-year-old men into military position (e.g. privates and officers) for their compulsory service. Throughout the years the batteries have been influenced by the changes in the factor analytically derived models of cognitive abilities. The SEB-67 used in this study was a further development of earlier tests to match the concept of general ability.14
Census data on adulthood SEP
Information on socioeconomic group for each conscript was obtained by record linkage with the National Population and Housing Census of 1985 held by Statistics Sweden. This census had a response rate of over 98%. The record linkage was possible because of the unique personal identification number of every citizen in Sweden.
The classification into the following eight socioeconomic groups in 1985 was conducted at Statistics Sweden and is based on information on occupation and the educational level required: (i) unskilled workers, (ii) skilled workers, (iii) assistant non-manual employees, (iv) non-manual employees at intermediate level, (v) non-manual employees at higher level, (vi) farmers, (vii) self-employed people, and (viii) those for whom no occupation was reported (e.g. unemployed, early retired, or disabled).
Through the personal identification numbers, the subjects of the study could be linked to census information concerning their parents. Information on father's SEP was obtained from the National Population and Housing Census of 1960 (response rate 99%), i.e. when the subjects were 9–11 years old. A second classification, into the following six socioeconomic groups, was based on information on fathers' occupation: (i) unskilled workers, (ii) skilled workers, (iii) assistant non-manual employees, (iv) non-manual employees at intermediate or higher level, (v) farmers, and (vi) those not classified in a socioeconomic group.
Outcomes
With personal identification numbers we linked records with the National Cause of Death Register 1971–2000 held by the National Board of Health and Welfare. The study population was followed with regard to all-cause mortality and to CVD mortality [ICD code, 8th (390–459), ICD code, 9th (390–459), and 10th (I00–I99) revisions], coronary heart disease (CHD) mortality (ICD code, 8th (410–414), ICD code, 9th (410–414), and 10th (I20–I25) revisions), CVD mortality other than CHD, mortality from violent causes [ICD code, 8th (800–999), ICD code, 9th (800–999) and 10th (V–Y) revisions], suicide mortality [ICD code, 8th (950–959, 980–989), ICD code, 9th (950–959, 980–989) and 10th (X60–X84, Y12–Y34) revisions], mortality from violent causes other than suicides, cancer mortality [ICD code, 8th (139–209), ICD code, 9th (139–209) and 10th (C) revisions], and to alcohol-related mortality [ICD code, 8th (291, 303), ICD code, 9th (291, 303) and 10th (F10, K70, K74) revisions].
Data analysis
The association between level of cognitive ability and mortality 1971–2000 was estimated on the basis of Cox's proportional-hazards model (both the univariate and the multivariate model) using the PHREG-procedure in the SAS computer package. The follow-up period was also divided into two periods, 1971–85 and 1986–2000 (i.e. at the approximate ages 20–35 and 36–50). In the multivariate models the relative hazard ratio (HR) associated with level of cognitive ability was estimated adjusted for childhood SEP and during the second period of follow-up, i.e. 1986–2000, we also adjusted for adult SEP. Cognitive ability is modelled as a continouos variable, with HRs given for an increase of 1 in the 9-point scale, as well as with dummy variables for each of nine categories using the highest test score as the reference category. All analyses were also conducted using the LOGIST-procedure and using this procedure we calculated the association between IQ score and mortality as effect size of 1 SD difference in IQ score. Confounding by childhood and achieved social class was assessed by fitting the socioeconomic groups as a series of dummy variables.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In Table 1 the distribution of the IQ scores within the 9-point categories is shown. The distribution is somewhat skewed towards the higher test scores in the sense that more persons than expected are found in the higher than in the lower bands.
|
In Table 2 the associations between IQ score and health outcomes are presented for the period 1971–2000 (at ages 20–50). In the crude analyses, each point increase in the 9-point aggregate IQ test score was associated with increased risk of all-cause mortality [HR 1.15; 95% confidence interval (95% CI) 1.13–1.18], CVD mortality (HR 1.20; 95% CI 1.1.4–1.27), mortality from violent causes (HR 1.16; 95% CI 1.13–1.20), and alcohol-related mortality (HR 1.28; 95% CI 1.17–1.39). There was no significant association between IQ score and cancer mortality (HR 1.04; 95% CI 0.99–1.10). Adjustment for childhood SEP had almost no effect on the risk estimates. The association between IQ score and CHD mortality was more pronounced than the association between IQ score and CVD mortality other than CHD.
|
Associations between IQ score and cause-specific mortality are also presented separately for the period 1986–2000 (at ages 36–50) (Table 2). The association between IQ score and mortality from some causes was more marked during the later follow-up period [e.g. CVD mortality other than CHD (+25%) and mortality from injuries other than suicide (+36%)]. Adjustment for childhood SEP had almost no effect on the risk estimates. Adjustment for SEP in 1985 attenuated the increased risks for the later period somewhat (25% for all-cause mortality, 18% for CVD mortality, 28% for mortality from violent causes, and 37% for alcohol-related mortality). The association between IQ score and CHD mortality (HR 1.24 95% CI 1.14–1.36) was marginally attenuated by adjustment for childhood SEP (–8%) but not further lowered by adjustment for adult SEP. The association between IQ score and CVD mortality other than CHD was not changed by adjustment for childhood SEP but lowered by 25% by adjustment for adult SEP. A strong graded increase of HR was seen from high to low test scores for all-cause mortality and for mortality from violent causes during the period 1971–85 as well as during the period 1986–2000, i.e. the approximate ages 20–35 and 36–50 (Figure 1). The slightly stronger association in the second period seems to be explained by the fact that suicides are more strongly associated with IQ and that a larger proportion of mortality from injuries consists of suicides during the second period of follow-up. Moreover, the associations between IQ and accidents are stronger during the second period of follow-up.
|
We also conducted the same analyses using logistic regression and the results were the same. The increased risk of all-cause mortality calculated as effect size of 1 SD difference in IQ score in those analyses was 1.34 (95% CI 1.29–1.41), for CVD mortality 1.45 (95% CI 1.29–1.62), and for mortality from violent causes 1.36 (95% CI 1.28–1.45) for the period 1971–2000.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In this cohort of Swedish males, born 1949–51, IQ test scores at ages 18–20 were related to all-cause mortality, CVD mortality, and mortality from violent causes during a 30 years follow-up period, and a gradient in mortality risk was seen across almost all IQ score groups for these outcomes. IQ score was not significantly associated with cancer mortality. The increased risks were not affected by control for childhood SEP. During the last 15 years of the follow-up period the increased risks were more or less attenuated by adjustment for adult SEP.
The study population is large and only a very small proportion of Swedish men are exempted from conscription, i.e. the cohort is highly representative of males born around 1950. Apart from the size of the study another major advantage is the rigour in standardization of the IQ measurement score. However, in this study only men are included. In the study based on the British 1946 birth cohort6 men but not women showed an increased risk of mortality with low IQ. Since the cohort is still comparatively young and the number of deaths is limited, the relation between IQ score and cause specific mortality is so far still difficult to test.
A striking finding in the present study was the gradually increased mortality by decreasing test score on almost all nine levels followed between 20 and 35 as well as between 36 and 50 years of age. In the British 1946 birth cohort,6 and in the Midspan cohorts born 1921,5,16 the excess mortality risk found was concentrated only in the lowest-scoring quarter. In the Danish study including males born in 1953 a graded increase of mortality between 15 and 49 years of age from the lowest childhood IQ score quartile to the highest was seen.7 However, when divided into two periods of follow-up, 15–34 and 35–49 years of age, it was shown that the excess mortality risk found during the second period of follow-up (when most of the cases of death occurred) was concentrated to the lowest-scoring quarter also in the Danish study. The absence of an IQ-mortality gradient in all three previous studies where such associations were presented might be attributable to certain characteristics of the persons lost to follow-up or missing at base-line. Several studies have shown that non-participation in epidemiological studies has been related to SEP, as measured by occupational class or income, in a number of studies from several countries.17–19
Whalley and Deary1 suggested four possible mechanisms for the observed relationship between IQ score and mortality: (i) cognitive ability as possible mediator between negative childhood experience and adult mortality, (ii) a poor cognitive ability as a predictor of a less safe adult environment, (iii) cognitive ability as a predictor of healthy behaviours, and (iv) cognitive ability as an indicator of ‘system integrity’. Our finding of a strong gradient in mortality risk over a wide span of IQ scores would suggest that poverty is not the sole mediating factor of the relationship between IQ scores and death. In this study adjustment for childhood social position as measured by father's occupation at age 9–11 had almost no effect on the risk estimates. Thus one may conclude that fathers' social class either is not a good indicator of unfavourable childhood life circumstances or that such circumstances are not important in the explanation of IQ-related mortality. In other studies indicators of negative circumstances in childhood such as low childhood social position,3,6,7 overcrowding in the school catchment area,20 and low birth weight,7 contributed only slightly to lower the increased risk of mortality associated with decreasing IQ score. Furthermore, in our analyses adjusting for SEP in 1985, (i.e. at ages 34–36) the relative risk for all-cause mortality and mortality from injuries decreased by <30% and CVD mortality decreased by 18%, during the second period of follow-up.
In the British 1946 birth cohort study the association between childhood IQ (found only in the lowest childhood IQ score quartile) and mortality decreased and lost statistical significance when adjusted for adult social position. Hart et al.5 showed that 
30% of the increased relative risk found in the lowest IQ quarter could be explained by adjustment for adult social circumstances. In a similar way Hart et al.16 found that adjustment for adult social factors only slightly (<14%) attenuated the association between childhood IQ score and CVD (mortality and hospital admissions combined).
IQ test score measured in late adolescence was a significant predictor of all-cause, as well as cause-specific, mortality during 30 years of follow-up. A gradient in mortality risk was seen across the nine cognitive ability groups. Cognitive ability seems to be an important early life determinant of mortality.
In the analyses of the association between cognitive ability and mortality in this study we adjusted for the influence from broad indicators of social circumstances in childhood and in adulthood (childhood and adult SEP). In future studies we will explore specific explanations for the association between cognitive ability and mortality and other health outcomes.
|
Acknowledgments |
|---|
The ethics committee at the Karolinska Institute, Stockholm, approved the study. This study was financed by the Swedish Council for Working Life and Social Research (Project No 2003–0382).
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1 Whalley LJ, Deary IJ. Longitudinal cohort study of childhood IQ and survival up to age 76. Bmj 2001;322:819.
2 O'Toole B, Stankov L. Ultimate validity of psychological tests. Pers Individ Diff 1992;13:699–716.[CrossRef]
3 Batty GD, Clark H, Morton S, Campell D, Macintyre S. Intelligence in childhood and mortality, migration, questionnaire response rate, and self-reported morbidity and risk factor levels in adulthood-preliminary findings from the Aberdeen ‘Children of the 1950s’ study. J Epidemiol Community Health 2002;56 (Suppl. 2):A1.[CrossRef]
4 Nishiwaki Y, Clark H, Morton SM, Leon DA. Early life factors, childhood cognition and postal questionnaire response rate in middle age: the Aberdeen Children of the 1950s study. BMC Med Res Methodol 2005;5:16.[CrossRef][Medline]
5 Hart CL, Taylor MD, Davey Smith G, Whalley LJ, Starr JM, Hole DJ et al. Childhood IQ, social class, deprivation, and their relationships with mortality and morbidity risk in later life: prospective observational study linking the Scottish Mental Survey 1932 and the Midspan studies. Psychosom Med 2003;65:877–83.
6 Kuh D, Richards M, Hardy R, Butterworth S, Wadsworth ME. Childhood cognitive ability and deaths up until middle age: a post-war birth cohort study. Int J Epidemiol 2004;33:408–13.
7 Osler M, Andersen AM, Due P, Lund R, Damsgaard MT, Holstein BE. Socioeconomic position in early life, birth weight, childhood cognitive function, and adult mortality. A longitudinal study of Danish men born in 1953. J Epidemiol Community Health 2003;57:681–86.
8 Hemmingsson T, Lundberg I. How far are socioeconomic differences in coronary heart disease hospitalization, all-cause mortality and cardiovascular mortality among adult Swedish males attributable to negative childhood circumstances and behaviour in adolescence? Int J Epidemiol 2005;34:260–67.
9 Galobardes B, Lynch JW, Davey Smith G. Childhood socioeconomic circumstances and cause-specific mortality in adulthood: systematic review and interpretation. Epidemiol Rev 2004;26:7–21.
10 Pensola TH, Valkonen T. Effect of parental social class, own education and social class on mortality among young men. Eur J Public Health 2002;12:29–36.
11 Hemmingsson T, Lundberg I, Diderichsen F, Allebeck P. Explanations of social class differences in alcoholism among young men. Soc Sci Med 1998;47:1399–405.[CrossRef][ISI][Medline]
12 Andreasson S, Allebeck P, Romelsjo A. Alcohol and mortality among young men: longitudinal study of Swedish conscripts. Br Med J (Clin Res Ed) 1988;296:1021–25.
13 Larsson D, Hemmingsson T, Allebeck P, Lundberg I. Self-rated health and mortality among young men: what is the relation and how may it be explained? Scand J Public Health 2002;30:259–66.[CrossRef][ISI][Medline]
14 Carlstedt B. Cognitive Abilities—Aspects of Structure, Process and Measurement. Göteborg: ACTA Universitatis Gothoburgensis, 2000.
15 Zammit S, Allebeck P, David AS, Dalman C, Hemmingsson T, Lundberg I et al. A longitudinal study of premorbid IQ Score and risk of developing schizophrenia, bipolar disorder, severe depression, and other nonaffective psychoses. Arch Gen Psychiatry 2004;61:354–60.
16 Hart CL, Taylor MD, Smith GD, Whalley LJ, Starr JM, Hole DJ et al. Childhood IQ and cardiovascular disease in adulthood: prospective observational study linking the Scottish Mental Survey 1932 and the Midspan studies. Soc Sci Med 2004;59:2131–38.[CrossRef][Medline]
17 Friedman GD, Cutter GR, Donahue RP, Hughes GH, Hulley SB, Jacobs DR Jr et al. CARDIA: study design, recruitment, and some characteristics of the examined subjects. J Clin Epidemiol 1988;41:1105–16.[CrossRef][ISI][Medline]
18 Ronmark E, Lundqvist A, Lundback B, Nystrom L. Non-responders to a postal questionnaire on respiratory symptoms and diseases. Eur J Epidemiol 1999;15:293–99.[CrossRef][ISI][Medline]
19 Goldberg M, Chastang JF, Leclerc A, Zins M, Bonenfant S, Bugel I et al. Socioeconomic, demographic, occupational, and health factors associated with participation in a long-term epidemiologic survey: a prospective study of the French GAZEL cohort and its target population. Am J Epidemiol 2001;154:373–84.
20 Deary IJ, Whiteman MC, Starr JM, Whalley LJ, Fox HC. The impact of childhood intelligence on later life: following up the Scottish mental surveys of 1932 and 1947. J Pers Soc Psychol 2004;86:130–47.[CrossRef][ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K Modig-Wennerstad, K Silventoinen, D Batty, P Tynelius, L Bergman, and F Rasmussen Association between offspring intelligence and parental mortality: a population-based cohort study of one million Swedish men and their parents J. Epidemiol. Community Health, August 1, 2008; 62(8): 722 - 727. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G D Batty, M J Shipley, L H Mortensen, S H Boyle, J Barefoot, M Gronbaek, C R Gale, and I J Deary IQ in late adolescence/early adulthood, risk factors in middle age and later all-cause mortality in men: the Vietnam Experience Study J. Epidemiol. Community Health, June 1, 2008; 62(6): 522 - 531. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R. Gale, G. D. Batty, and I. J. Deary Locus of Control at Age 10 Years and Health Outcomes and Behaviors at Age 30 Years: The 1970 British Cohort Study Psychosom Med, May 1, 2008; 70(4): 397 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. R Gale, I. J Deary, I. Schoon, and G D. Batty IQ in childhood and vegetarianism in adulthood: 1970 British cohort study BMJ, February 3, 2007; 334(7587): 245 - 245. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Batty, K Modig Wennerstad, G Davey Smith, D Gunnell, I. Deary, P Tynelius, and F Rasmussen IQ in early adulthood and later cancer risk: cohort study of one million Swedish men Ann. Onc., January 1, 2007; 18(1): 21 - 28. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G D. Batty, I. J Deary, and S. Macintyre Childhood IQ and life course socioeconomic position in relation to alcohol induced hangovers in adulthood: the Aberdeen children of the 1950s study. J. Epidemiol. Community Health, October 1, 2006; 60(10): 872 - 874. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. J Deary and G D. Batty Commentary: Pre-morbid IQ and later health--the rapidly evolving field of cognitive epidemiology Int. J. Epidemiol., June 1, 2006; 35(3): 670 - 672. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||