Source: United Kingdom – Executive Government & Departments
A paper published in Nature Medicine looks at the contribution of genetic and lifestyle factors to risk of premature death, ageing, and age-related diseases.
Prof Felicity Gavins, Professor of Pharmacology at Brunel University of London, and Royal Society Wolfson Fellow, said:
“This is an exciting study. The fact that most of the risk factors identified are modifiable highlights an enormous opportunity for prevention. By addressing social inequalities, promoting healthy behaviours and reducing harmful exposures, we can really make a meaningful difference in reducing age-related diseases and premature mortality.
“However, some caution is needed. This is an observational study, so further research is needed to confirm causal relationships, especially before any long-term policy changes are made. Furthermore, targeted interventions will be essential to translating these findings into real-world impact.”
Dr Stephen Burgess, Group Leader at the MRC Biostatistics Unit, University of Cambridge, said:
“This is a large and detailed investigation into the predictors of major causes of mortality in a UK-based population. It provides further demonstration supporting previous research that, in the majority of cases, our genes do not determine our future. There are exceptions, including rare conditions that are caused by a single genetic variation. But for the majority of conditions that Western people die from, disease risk is more strongly attributable to modifiable risk factors and our wider environment, as shaped by our upbringing and choices. Genetics can load the dice, but it is up to us how we play our hand.
“A limitation of the work is that it does not highlight particular risk factors, nor can it make specific causal claims about what would happen if we changed our risk factors and environment.”
Prof Frances Flinter, Emeritus Professor of Clinical Genetics, Guy’s and St Thomas’ NHS Foundation Trust; and Member of the Nuffield Council on Bioethics, said:
“This is a very impressive, thorough and detailed analysis of a vast amount of genetic and non-genetic data from the UK Biobank. The authors compare the relative contributions to ageing and premature mortality of genetic susceptibility markers (polygenic risk scores) and environmental factors, which they refer to as the ‘exposome’ (including alcohol, diet, smoking, housing, type of heating, weight in childhood etc).
“Overall, polygenic risk scores (PRS) for twenty-two major diseases explained less than 2% of additional variation in mortality, whereas the exposome explained 17%. In particular, the exposome explained a greater proportion of the variation than polygenic risk scores for the incidence of disease of the lung, heart and liver, whereas polygenic risk scores explained a greater proportion of the variation than the exposome for dementia and breast, prostate and colorectal cancers.
“The risk of premature mortality was lower in Black, Asian and ethnicities other than white, even after adjustment for socio-demographic deprivation factors, which is currently unexplained.
“With so much focus on genetic determinism these days, it is good to be reminded of the significance of environmental contributions to health, particularly as the risk factors are known and many can be modified.”
Prof Ilaria Bellantuono, Professor in Musculoskeletal Ageing; and Co-Director of The Healthy Lifespan Institute, University of Sheffield, said:
“This important study comprehensively confirms what smaller studies have suggested: multiple socioeconomic and environmental factors significantly influence the risk of developing age-related diseases. More critically, it highlights that health is shaped by multiple interacting factors. This has important policy implications, meaning that policies targeting only one or two of these factors will have limited impact on extending healthspan. The findings support the need for an integrated, multi-faceted approach to prevention and to identify the most influential domains for intervention (smoking, socioeconomic status and deprivation, physical activity, sleep and mental and physical wellness including tiredness, as well as early life exposures including height and body size at 10 years and maternal smoking around birth).
“The study is rigorously conducted and transparently acknowledges its limitations, which are inevitable in research of this nature.”
Dr Julian Mutz, King’s Prize Research Fellow at the Social, Genetic & Developmental Psychiatry Centre, King’s College London, said:
“The study by Argentieri, van Duijn, and colleagues sought to tease apart the relative contributions of environmental exposures (termed the “exposome”) and genetic risk on biological ageing and premature mortality.
“The authors analysed data from the UK Biobank, a unique resource with a wealth of information on sociodemographic characteristics, health records, genetics and biomarker data from half a million UK residents.
“They employed a complex analytical design to identify environmental exposures that were independently associated with biological ageing (defined using a proteomic ageing clock that they developed in a previous high-profile study) and mortality, while minimising the risk of reverse causation, confounding and correlation between exposures. The approach is elegant, though certain assumptions warrant caution. For example, the finding that many exposures independently associated with mortality (e.g., diet or mental health) were not associated with the proteomic ageing clock (or had an association in the opposite direction) does not necessarily mean that these exposures do not impact ageing biology.
“Key findings from the study were that a higher income, Asian or Black ethnic background, higher levels of physical activity and living with a partner were associated with lower mortality risk and a protein-predicted age younger than chronological age. Smoking, living in council housing (reflecting socio-economic status) and the frequency of feeling tired were associated with higher mortality risk and a protein-predicted age older than chronological age.
“Each of the 25 independent exposures that the authors identified was associated with incident diseases and ageing biomarkers.
“To investigate the relative contribution of the environmental exposures compared to genetics, the authors calculated polygenic scores for 22 diseases. Polygenic scores aggregate the small effects of many common genetic variants to estimate an individual’s predisposition to specific traits or diseases. However, there are several caveats to this approach: first, polygenic scores only capture part of the genetic risk; and second, many environmental exposures also have a genetic component. The broad headline of the press release that “environmental factors affect health and ageing more than our genes” should be viewed in light of these limitations.
“One of the most interesting findings from this study is the comparison of the contributions of chronological age and sex (both non-modifiable risk factors), environmental exposures and polygenic scores across several disease endpoints. For example, for certain diseases (e.g., dementia), genetics appears to be more important.
“A key implication of the study is that there is a broad range of modifiable risk factors that could be targeted to reduce the risk of premature mortality and age-related disease. How successful this will be remains to be seen. We already know much about the health-promoting effects of lifestyle interventions, such as physical activity and smoking cessation, but a significant intention–behaviour gap remains.
“The authors have, for the most part, carefully highlighted that the observed associations may not be causal.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:
“This new study involves a large dataset, using data from almost half a million participants in the UK Biobank, data on 164 different environmental exposures (using ‘exposure’ in the broad epidemiologists’ sense, from smoking and intake of various foods, to how plump they were at age 10, to their ethnicity) and (for some of them) genetic and blood measures too. It’s big data, and the researchers use some big-data methods.
“The aim was to quantify the contributions of environmental exposures and genetics to aging and premature mortality, taking into account many aspects of people’s environment rather than concentrating on a few risk factors determined in advance.
“The results are interesting, and I think they do support the researchers’ view that we can learn more by looking at many environmental exposures together rather than trying to pick them off one (or a few) at a time. However, there are some important limitations (as the researchers make clear).
“It would be easy to dismiss this new research by saying that all they have really found is that, if you want to be healthy in old age, you need to give up smoking, do some exercise and not be poor, and we already knew that. But that’s not (in my view) the important finding at all. The important finding is that you get more by looking at more aspects of the environment, if you have enough good data to do that – but that needs careful statistical analysis, including aspects that this study could not do itself. However I think there are good reasons not to pay too much attention to the exact numerical results in the paper, for reasons I’ll come to.
“This is an observational study – the UK Biobank researchers did not choose how the participants acted, but only observed and recorded what they said and did. Like all observational studies, the findings are about correlations and associations, not about cause and effect. The statistical methods used by the researchers can’t determine whether the associations between exposures and ill health and mortality, that they observed, are there because the exposures cause the ill health and mortality. They might, or they might not.
“The way the researchers filtered out exposures that might have showed up as associated with ill health only because they were correlated with other exposures, or because the exposure was actually caused by ill health (reverse causation, as it’s called), does to some extent make it a bit more likely that the associations they mainly report on are ones of cause and effect – but they certainly can’t confirm that they are cause and effect. The researchers say, in their conclusion, that their results indicate that interventions based on environmental exposures are possibly (my emphasis) the best starting point for improving age-related health, but they add that “future causal modelling [that is, research that specifically looks at cause and effect, which uses different methodology] will be needed to study specific exposures of interest.”
“In view of these issues about cause, it’s unfortunate that the press release uses a lot of language that implies the associations are indeed reflecting cause and effect. They talk about the impact of environmental factors on mortality and aging. If something isn’t causing the ill health, ‘impact’ is the wrong word – if you change a factor that is correlated with ill health but doesn’t cause it, you won’t change the level of ill health.
“And when the release says that environmental factors explained 17% of the variation in risk of death, compared to less than 2% for genetic predisposition, this is presenting a misleading picture of what is reported in the research paper. The paper talks about additional mortality variation (in addition to the variation explained by age and biological sex, which are the most important factors, unsurprisingly, along with smoking). And in this context, statisticians are using ‘variation explained’ to mean something statistically technical that has nothing direct to do with cause and effect, even though it sounds as if it does.
“There are other important limitations. The UK Biobank population isn’t typical of the general UK population. And the exposures were all measured at only one time point, when people first entered the UK Biobank study. Therefore, even though the UK Biobank is a major study that goes on through time, these findings can’t, for instance, look at the impact on ill health if someone gives up smoking, or becomes wealthier, or changes what they eat. The researchers emphasise the importance of studying what leads to ill health across the life course, not just at one or a few time points, but like most studies using UK Biobank data, they could not actually do that in this study, beyond looking at some things that participants said about their childhood when they entered the study.
“There is no implication that the 25 independent environmental factors that were identified in this research are the most important environmental factors, or the only important ones. The filtering process that removed factors that might have been correlated to strongly with other factors, or might have been liable to reverse causation, may have removed some that were in fact important to health. (I’m not saying that they should not have been removed, in the light of the overall aims of this study – just that removing them could have led to something being missed.)
“And obviously the researchers could only take into account environmental exposures that were recorded in the UK Biobank data, and that’s not everything. The early life exposures, mentioned in the press release and the paper as being important, were actually recorded alongside all the others when people entered the study, so based on what they recalled, and not actually followed up over time.
“Ideally in a study like that using a big and complicated data set, researchers would model the data statistically using just part of the data set, and then check with the rest of the data set whether the findings hold and are not just a statistical fluke. These researchers did that, splitting the data on English UK Biobank participants into two and checking the results from one half on the other half, and then checking several aspects of the statistical modelling by validating the results on data from UK Biobank participants in Scotland and Wales. That’s good, but not ideal, because the Scottish and Welsh participants are likely to be too similar to the English participants to give an independent enough validation.
“It’s interesting that the research paper says that they sought to validate the findings using a different study based in Rotterdam, which would have been much better than the Scottish and Welsh UK Biobank data. But they could not do that because the Rotterdam study did not have enough recorded environmental exposures that matched those in the UK Biobank. They point out that this is likely to be a more widespread problem, because there’s no standard way across different studies of this kind to choose which exposures to record, or how to define them.
“I have to say that I personally wouldn’t pay too much attention to any of the exact figures on associations that are given in the paper. That’s partly because of the limitations I’ve mentioned (and the researchers give more limitations in the paper). But it’s mainly because the data set is big and complicated, and the statistical methods used involved many stages and are complicated. The researchers had to make a long series of choices on which data to analyse and how to analyse it. Another team of researchers would not have made the same choices in each case. That doesn’t mean that this team is wrong and another team would be right – just that there often isn’t a clear best choice to be made. And other choices would have led to different findings, in terms of the detailed numbers at least.
“Statisticians sometimes refer to the series of choices of how to analyse a data set, not entirely seriously, as ‘researcher degrees of freedom’. This study has a lot of researcher degrees of freedom. The researchers did check out some of their choices by carrying out sensitivity analyses, but that doesn’t get near to dealing with every choice they had to make. If time and money were no object, it would be very interesting to see what a different research team made of the same data – but in the real world, that’s not going to happen.
“One final point about the press release. It says that 23 of the 25 independent environmental factors, identified in the research as contributing to the association between environmental exposure and ill health, ‘are modifiable’. The research paper says only that they are potentially modifiable. This sounds like a nit-pick, and maybe it is – but look at the factors (in Figure 2d in the paper, which shows the 25 along with age and biological sex). Smoking is modifiable, even if it can be hard for individuals to make that modification. But for some of the others it’s not easy to see what the modification might be. How do you modify things so that you are living with a partner, if you currently aren’t? (Living with a partner is associated with better health.) How do you modify how often you feel fed up, or how often you feel unenthusiastic? These potential modifications could maybe be done, but saying they are ‘modifiable’ is too much of a simplification. And it’s certainly important to understand that modifying some of them would be possible only by changes in society – it’s not just a question of individuals choosing what to do. (It also bears repeating that this study, because of the issues about cause and effect, can’t actually tell us with any certainty whether modifying these facts would actually change health anyway.)”
Dr Divyangana Rakesh, Lecturer and Researcher in the Department of Neuroimaging, Institute of Psychiatry, Psychology & Neuroscience, King’s College London, said:
“This study makes clear just how much our environment shapes aging and mortality, and it is not surprising that environmental risk often outweighs genetic risk. The authors used a rigorous approach to show that while genetics play a role in specific diseases, our environment – from socioeconomic status to lifestyle factors – shapes overall health trajectories in powerful ways. We see this in developmental research as well, where environmental factors, including socioeconomic status and deprivation, play a crucial role in shaping children’s outcomes. Findings like these reinforce the urgent need to address environmental determinants of health if we want to support healthy development and aging for everyone.”
Prof Joyce Harper, Head of the Reproductive Science and Society Group, UCL Institute for Women’s Health, UCL, said:
“This extensive study systematically examined environmental factors linked to aging using data from the UK Biobank. The researchers conducted an exposome-wide analysis of all-cause mortality in a cohort of 492,567 individuals and investigated how these exposures influenced a proteomic age clock. Their findings identified 25 independent environmental factors associated with both mortality risk and proteomic aging.
“It is so great to see this brilliant study from Oxford Population Health. In today’s society, so many are trying to get a quick fix to improve health and longevity, but this study and others are showing the importance of our lifestyle and environment on healthy aging. It is the first study to show how the combined effect of individual exposures affects us through the life course. I hope people are listening.”
‘Integrating the environmental and genetic architectures of aging and mortality’ by M. Austin Argentieri et al. was published in Nature Medicine at 10.00am UK time on Wednesday 19 February 2025.
DOI: 10.1038/s41591-024-03483-9
Declared interests
Prof Felicity Gavins: “No conflicts.”
Prof Frances Flinter: “No CoI.”
Prof Ilaria Bellantuono: “I am funded by the Michael J Fox Foundation, Dunhill Medical Trust. I co-lead UkAgeNet (https://ukagenet.co.uk/ ) and I am co-director of the Healthy Lifespan Institute.”
Dr Julian Mutz: “I report no conflicts.”
Prof Kevin McConway: “Previously a Trustee of the SMC and a member of its Advisory Committee.”
Dr Divyangana Rakesh: “I have no conflicts of interest to declare.”
Prof Joyce Harper: “No conflicts. I am writing a book on health and happiness over 50 but I do not think that conflicts.”
For all other experts, no reply to our request for DOIs was received.