The biggest-ever study investigating the effect of lifestyle on how we age came to some surprising conclusions
This is Everyday Science with Clare Wilson, a subscriber-only newsletter from The i Paper. If you’d like to get this direct to your inbox, every single week, you can sign up here.
Hello, and welcome back to Everyday Science.
Whether or not we stick to it, most of us are aware of the official healthy eating advice: try to avoid too much red meat, salt and fat, especially saturated fat, and eat plenty of fruit and vegetables and wholemeal grains.
But how much of an impact does such a diet have? I was intrigued to see a study released this week which claimed to be the largest ever to have investigated the impact of lifestyle factors on healthy ageing and our risk of dying. For those who believe that diet is all-important, the results may come as a surprise.
The research was designed to find out the relative contributions of lifestyles and genes on people’s risk of death.
It used a large ongoing project called UK Biobank, which two decades ago recruited half a million middle-aged Britons to start regularly filling out questionnaires and undergo a barrage of medical tests. The data has been used in countless studies by scientists around the world.
In the latest analysis, researchers at the University of Oxford tried to quantify people’s exposure to everything in the environment that could conceivably affect their health – their “exposome”.
They analysed 164 factors, from obvious things like smoking and what they ate and drank, through to what their home was like, noise levels in the street and whether they attended religious services.
They also analysed the participants’ DNA sequences to calculate their genetic propensity to 22 of the most common medical conditions that can kill us, such as cancer and heart attacks.
In terms of outcomes, the researchers recorded whether the person had died since the study began and the medical cause. As most were still living, they also measured levels of 25 proteins in their blood that have previously been linked with healthy ageing.
I should point out that I personally think these kind of ageing “biomarkers” are not yet a proven way to measure health status. Biomarkers are a common approach in ageing research, but different scientists use different groups of proteins, so how do we know which is right?
Ageing biomarkers certainly seem to measure something to do with health, but you might as well just call them health biomarkers. Calling them ageing biomarkers gives an undeserved aura of scientific credibility to “biological age” tests touted online by snake-oil salesmen. But let’s just roll with calling them ageing biomarkers for now, and consider them a proxy for overall health status.
Nature or nurture
So what were the results? The analysis has now been published in the journal, Nature Medicine. The finding highlighted by the researchers is that lifestyle factors are implicated much more than genes with people’s risk of death and ageing biomarkers.
If we assume lifestyle factors were directly responsible, they accounted for 17 per cent of the variation in people’s risk of death, while genes could explain only 2 per cent.
In other words, if you are some kind of genetic determinist, who thinks that DNA controls every aspect of our lives, these figures should give you pause. If your parents died young, that doesn’t necessarily mean you will too.
But I think there are more interesting conclusions we can draw from this study. I was struck by how little of our susceptibility to disease is affected by our lifestyle.
Considering how much we get told these days that it is crucial to eat healthily, sleep well and get in our 10,000 steps a day, 17 per cent seems a small number.
By far the biggest impact on people’s risk of dying was their age and sex – factors outside their control. And when you put these contributions together with the lifestyle and genetic factors, it adds up to only 68 per cent of the variance in risk.
Que será, será
The other 32 per cent – more than double the impact from lifestyle – seems to be down to just chance influences, like accidents, or other things you can’t control. Which would seem to support a “Que será, será” approach to life.
The next interesting finding comes from looking at exactly which lifestyle factors are implicated.
Most of the results, you won’t find surprising. The largest lifestyle contributor to health is a decision that many people make when they are still just a teenager – whether to start smoking. The next item in the list of exposures that would be considered a lifestyle choice is physical activity levels. I’m sure you can guess in which direction both of those factors influenced death rates.
But the big surprise – to me – was that in general, diet was not found to correlate with healthy ageing.
With the exception of one food group, which I’ll get to shortly, this study did not point the finger at any of the usual suspects – not red meat, not processed meat, not fibre, and not vegetables.
That might be an artefact of the way the researchers analysed the figures: they concluded that a food group was linked with health only if it appeared to affect both death risk and ageing biomarkers in the same way, either positively or negatively.
So for instance, vegetable intake was judged not have a significant effect because it correlated with a lower risk of death, but worse ageing biomarkers – which isn’t really plausible. Alternatively, you might think that supports my contention that ageing biomarkers can’t yet be trusted.
Weak evidence
That aside, there’s an even bigger caveat with this study to consider, which also affects nearly the whole of nutritional science.
It is that most research in this field is a rather weak kind of medical evidence known as “observational” studies. By contrast, the most authoritative kind of nutrition study is a randomised trial, where you change something about people’s diet to see if that has a health effect. If so, you can conclude that the new diet caused the effect.
In observational studies, you just observe people’s diet and observe their health outcomes, and you may observe a correlation. But that doesn’t prove that any diet caused the health outcome – something else could be the real cause.
The classic example is that certain patterns of diet are popular among those with higher wealth, and wealth also correlates with good health for a host of reasons unrelated to diet.
That means observational studies should not guide us on how to live our lives. “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,” said Professor Kevin McConway, a statistician at the Open University who was not involved with the study
These problems with dietary observational studies is partly why they are notorious for giving contradictory results. One week a study will suggest that drinking full-fat milk is bad for us, for instance, while the next week a different study finds it is good for us.
“That is a problem in dietary research, that the findings are not very reproducible,” said Professor Cornelia van Duijn at the University of Oxford. “We don’t see consistency of findings even in our study, let alone over [several] studies.”
Her colleague Dr Austin Argentieri said their study doesn’t prove diet has no effect. It could be that the Biobank study “just did not have reliable enough assessments of diet”. But UK Biobank is widely acknowledged to be one of the largest and most reputable health studies in the world – if this can’t identify how diet affects health, I’m not sure where we go from here.
If you are a cynic (like me), you might conclude that the entire body of nutritional observational research is untrustworthy – and that you can pick and choose studies to support whatever dietary ideology you favour.
If you think red meat is the root of all evil, you can find observational studies supporting that. Or you could pick studies that show the real dietary sins are carbohydrates, or fat or, more recently, ultra-processed food.
As for the latest study, maybe I’ll just place my faith in the one dietary result that emerged from the analysis as having an impact on health. That food is cheese.
According to NHS health advice, we should avoid eating too much cheese because it is high in saturated fat and salt.
What did the study find? People who ate more cheese had a lower risk of death and had a better result for their ageing biomarkers.
I think I’ll start believing in ageing biomarkers after all.
What else I’ve written recently
With a few weeks left of winter, it looks like the predicted “quad-demic” didn’t materialise. What will health officials do next year to ratchet up the threat level… predict a “quin-demic”?
I’m reading
A Gentleman in Moscow by Amor Towles. If you think it might be hard to empathise with a Russian aristocrat living in a luxury Russian hotel, think again. I’m staying up far too late every night reading this book to find out what happens next.
