The influence of familial factors on the intelligence-mortality association – A twin approach

The aim was to investigate whether the widely reported inverse intelligence-mortality association could be replicated in twin intrapair analyses in which genetic and shared environmental factors are controlled for, as per the study design. Furthermore, the issue of generalizability of twin findings to the general population was investigated by comparing associations for twins and singletons. A large male cohort of 7316 twins and 649,809 singletons, for whom intelligence was assessed at a mean age of 19.5years, was followed for information on mortality until late mid-life (mean age: 63.5years). In unpaired analyses, where each twin was treated as an individual, a higher risk of mortality for each standard deviation decrease in intelligence test-score was found for all twins (hazards ratio (HR)per1SDdecrease=1.28; 95% CI:1.21,1.35), dizygotic (HRper1SDdecrease=1.24; 95% CI:1.16,1.33) and monozygotic (HRper1SDdecrease=1.33; 95% CI:1.18,1.50) twins. The observed associations were of a similar magnitude to those found for singletons (HRper1SDdecrease=1.28; 95% CI:1.27,1.29) and for the group of twins with unknown zygosity (HRper1SDdecrease=1.30; 95% CI:1.13,1.51). The inverse intelligence-mortality associations were replicated in intrapair analyses of all twins and DZ twins where genetic and shared familial factors were partly controlled for, as per the study design. For twin pairs with a difference in intelligence test-score of half a standard deviation or more, the proportion of twin pairs in which the lower-scoring twin died first was 55% (95% CI: 0.52,0.58). These findings suggest a modest inverse association between young adult intelligence and mortality until late mid-life, with genetics and shared environment explaining only a small part of the association. Furthermore, the study suggests that the twin findings were generalizable to singletons.