Purpose: Studies of middle-aged and particularly older-aged adults found that those with higher scores on tests of IQ (cognitive function) had lower rates of later mortality. Interpretation of such findings potentially is hampered by the problem of reverse causality: such somatic diseases as diabetes or hypertension, common in older adults, can decrease cognitive function. Studies that provide extended follow-up of the health experience of individuals who had their (premorbid) IQ assessed in childhood and/or early adulthood minimize this concern. The purpose of the present report is to systematically locate, evaluate, and interpret the findings of all such studies.
Methods: We systematically identified individual-level studies linking premorbid IQ with later mortality by using four approaches: search of electronic databases (MEDLINE, EMBASE, and PSYCHINFO); scrutiny of the reference sections of identified reports; search of our own files; and contact with researchers in the field. Study quality was assessed by using predefined criteria.
Results: Nine cohort studies met the inclusion criteria. Overall, study quality was moderate. All reports showed an inverse IQ-mortality relation; i.e., higher IQ scores were associated with decreased mortality risk. The nature of this relation (i.e., dose-response or threshold) and whether it differs by sex was unclear. The IQ-mortality association did not appear to be explained by reverse causality or selection bias. Confounding by other early-life factors also did not seem to explain the association, although some studies were not well characterized in this regard. Adult socioeconomic position appeared to mediate the IQ-mortality association in some studies, but this was not a universal finding.
Conclusions: In all studies, higher IQ in the first two decades of life was related to lower rates of total mortality in middle to late adulthood. Some plausible mechanistic pathways exist, but further examination is required. The precise nature of the IQ-mortality relation (particularly in ethnic minorities and women) and the link between IQ and disease-specific outcomes also warrants further research.