August 31, 2017

expert reaction to study investigating the relationship between higher education and lower risk of heart disease

Publishing in The BMJ, researchers looked at educational attainment, using genetic data as a proxy, and the development of coronary heart disease. They reported an association between lower education and the development of coronary heart disease.

Dr James Doidge, Senior Research Associate, Administrative Data Research Centre for England, Great Ormond Street Hospital Institute of Child Health and Farr Institute of Health Informatics Research, UCL, said:

“I think this study has one major flaw. The authors report that the population can be split into groups with different genetics that are associated with how much education they will have. They say that these 162 DNA markers for education are not linked to potential confounders such as physical activity or diet and so they argue this creates a form of natural randomised controlled trial, and that any differences in heart disease are because of the education, not diet or activity or any other factor.

“However, these differences in education are probably working by affecting brain function. But brain function influences mental health, behaviour and lifestyle choices, which in turn affect cardiovascular health, and the cardiovascular system is of course largely controlled by the brain. Brain function and mental health clearly also affect educational attainment and therefore are likely to confound the relationship between educational attainment and cardiovascular health.

“Neither the study design (Mendelian randomisation) nor the tests for genetic pleiotropy (multiple effects of single genes) address this confounding by brain function or factors relating to it. While Mendelian randomisation may address confounding by socioeconomic disadvantage, it does not address confounding by brain function, mental health or intelligence, all of which are likely to be affected by the 162 genetic markers because there are few other plausible ways that the genes could affect education.

“Essentially, because educational attainment is so far removed from the genes that influence it, the authors may have stretched the principle of Mendelian randomisation beyond its limits and broken its most basic assumption of ‘no effect of the instrument on the outcome other than through the exposure.”

Prof Tim Frayling, Professor of Human Genetics, University of Exeter said:

“There are strong associations between health and wealth, and inevitably that means health and education. A key question for public health is whether, and to what extent, improving education could help improve people’s health. However it is very difficult to test this in a real world situation because you cannot perform a randomised controlled trial of people’s education.

“This study is important because it uses the next best thing to a randomised controlled trial – the authors effectively randomised people to different educational outcomes by comparing their genetic propensity to continue for longer in education. They then showed that the people with the ‘longer education’ genes had fewer heart attacks compared to the people with the ‘shorter education’ genes. What they don’t discuss in detail is whether or not this could be down simply to higher educational ability, not length in education itself – if people are better able to understand and act on health advice they might have spent longer in education and have fewer heart attacks”

Dr Louise Brown, Senior Statistician, MRC Clinical Trials Unit, UCL, said:

“This is an interesting and well-conducted piece of work demonstrating a link between increased length of education and reduced incidence of coronary heart disease. Whilst I think it is likely that education lies on the causal pathway to coronary disease, it may well be mediated through increased wealth and financial security which we know stem from higher levels of education. Perhaps if the same levels of financial security could be attained without the need for additional years of education we may also see a similar drop in the incidence of coronary disease. Factors can be on the causal pathway but not be directly causal e.g. an individual may have an extensive education but then end up living a poor and deprived life despite this education. Mediation explores whether both or one of these factors (i.e. education and financial security) need to be in place to have an impact on coronary disease. It is encouraging that the authors allude to this in their discussion on potential mechanisms and I hope this work moves forward as they suggest.

“I am all for encouraging further education but to leap in and demand it of all whilst not tackling the underlying issues of possible poverty and poor financial security I feel sends us down the wrong path. We cannot all go to university, nor do we all wish to, and it might be more helpful to understand the relationship between education, financial well-being and health in general before considering major educational change.”

* ‘Education and coronary heart disease: mendelian randomisation study’ by Tillmann et al. was published in The BMJ  on Thursday 31 August. 

Declared interests

Prof Tim Frayling: None received.

Dr Louise Brown: No conflicts of interest.