Last week saw a flurry of media reports, of whose headlines one of the worst preceded one of the best reports. ‘Eating too many ultra-processed foods has been linked to a higher risk of early death,’ barked the Telegraph – but went on to explain carefully and fairly a ground-breaking report. Other broadsheets opted for the easy option: big report, ultra-processed food, death.
Food-type blaming can be a comforting evasion of a simple truth: overeating makes you fat
The report caught my eye because I’ve been consistently sceptical about sensationalist books and statements demonising in wholesale terms the consumption of foods categorised, in pseudo-scientific language, as ‘ultra-processed’. I question the usefulness of the category. Food-type blaming can be a comforting evasion of a simple truth: that overeating makes you fat and unhealthy. Instead, we’re told food manufacturers are to blame. Telling people that obesity and ill-health are somebody else’s fault sells books and newspaper articles. But what if this area of knowledge is a house built on sand?
So I read the whole Harvard report, usefully set out in the British Medical Journal. Its findings are almost grotesquely inconsistent with those headlines. The report pours a bucket of cold (if academically phrased) water on generalised and simplistic linkages between ultra-processed food (UPF) and ill-health. (Here I should state that I have no vested interest in defending any of this often unwholesome stuff, nor the remotest association with the food industry.)
In summary, the Harvard report found an astonishingly weak link (4 per cent) between heavy intake of UPF and reduced longevity; and when processed meats, dairy and high sugar content were subtracted from the equation, there was no link at all. Nor was there any link with cancers or heart disease. In fact a good deal of UPF, for instance nut-based, wholegrain cereal-based and even wholegrain bread, turns out to be positively life-prolonging. As the authors say: ‘We aim to rectify the potential misperception that all ultra-processed food products should be universally restricted, and to avoid oversimplification when formulating dietary recommendations.’
To put it in numbers: the research found the mortality rate among the quarter of humans who ingested the most UPF was 1,536 per 100,000 person years; among the quarter who ingested the least UPF it was 1,472. The authors concluded: ‘This study found a higher intake of ultra-processed foods was associated with slightly higher all-cause mortality, driven by causes other than cancer and cardiovascular diseases. The associations varied across subgroups of ultra-processed foods, with meat/poultry/seafood based ready-to-eat products showing particularly strong associations with mortality.’
These results are the more compelling because the research has followed the food intake of some 74,000 women and 39,000 men in America for more than 30 years. The 4 per cent linkage was obtained from comparing the quarter with the diets highest in UPF and the quarter with the lowest. Even as regards that small apparent linkage, your mind may have gone, as did mine, to a problem with all such studies. What if the kind of people who eat a lot of junk food tend to have unhealthier than average lifestyles: smoking, obesity, sedentary habits, for example, spring to mind? Might these factors rather than the UPF they eat be the primary link with mortality?
The researchers have not overlooked this: ‘We adjusted for race/ethnicity, marital status, physical activity, body mass index, smoking status and pack [cigarette consumption] years, alcohol consumption…’ – the list goes on. But ‘controlling for’ extraneous factors that might be the real culprits is always a hazardous business for data analysis, importing (as it sometimes must) a measure of guesswork about the likelihood and strength of other connections. In Britain there’s probably a notable but counter-intuitive correlation between scratch-card use and morbidity, before you control for social class.
The study is surprising, too, for some of the factors that it turned out didn’t need to be controlled for: ‘We found no interaction by body mass index or physical activity [and] no associations for mortality due to cancer or cardiovascular diseases.’ The findings ‘remarkably’ suggested no association overall between smoking and morbidity, but I wonder whether that’s because the young smoke less but eat more junk so the two cancel each other out?
And here’s a surprise: ‘Compelling evidence shows that nuts and (dark) chocolate, common constituents of “sweet snacks and desserts”, are inversely associated with cardiovascular diseases. We observed that dark chocolate in the [UPF] subgroup “packaged sweet snacks and desserts” was associated with decreased mortality.’
UPF studies rest upon a classification system – ‘Nova’ – that originated with a Brazilian academic. The more I read this Harvard study, which has all the hallmarks of honest and rigorous enquiry, and the more tangled, self-confounding and counter-intuitive the results seem, the more convinced I became that medical science has a root-cause problem here. My next thought was triggered by this hint: ‘The Nova classification is based on broad categories that do not capture the full complexity of food processing, leading to potential misclassification.’
‘Aha,’ I thought, ‘are you stopping short of a possible destination for your reasoning? What if the problem is not how we classify the processing of what we eat, but the whole concept: “ultra-processed” food? What if it’s not a useful catch-all term at all? What if, whatever the health risks undoubtedly are with some foods, they don’t arise from the simple fact of processing? Perhaps they have no single cause at all?’ That would help explain another suggestion in this report: that mortality rates may be more associated with the absence of a healthy diet than with the presence of UPF within that diet. Indeed, the report points out that some of the components of a healthy diet are in fact processed foods. They conclude: ‘Dietary quality was observed to have a more predominant influence on mortality outcomes than ultra-processed food consumption.’
In our hunt for what foods, or what components in food, reduce lifespan, could ultra-processing be a red herring? I’m a rank amateur here but I do understand marketing. UPF is a great marketing tool for treatises of popular science, but it is no basis for new government taxes.
Listen to Matthew Parris and Dr Chris van Tulleken debate ultra-processed foods on The Edition podcast:
Comments