Last month, the BBC publicised the results of a study which claimed that sugar consumption in early life increases the risk of dementia. People who don’t eat much sugar in the first years of life are supposedly 23 per cent less likely to develop the condition when they are older. “Our findings suggest that limiting sugar during the earliest stages of life may have lasting benefits for brain health,” said the study’s lead author Jiazhen Zheng.
How does he know this? He looked at data from the UK Biobank and found that people born between 1940 and 1953, when sugar was rationed, were less likely to develop dementia than those born afterwards.
Of 15,025 who experienced sugar rationing before birth and during their first year of life, 388 developed dementia. Of 15,256 who experienced sugar rationing before birth and during their first two years of life, 456 developed dementia. Of 23,774 born after sugar rationing, 504 developed dementia.
If you get your calculator out, you will find that these figures show that the rate of dementia actually fell over time, with 2.1 per cent of the youngest cohort suffering from the disease, compared to 2.6 per cent of the oldest cohort and 2.5 per cent of the second group. The claim that the people who endured sugar rationing were 23 per cent less likely to get dementia is based on numbers that have undergone substantial adjustment.
But that is not the real problem with this study. The real problem is something that has doubtless already occurred to you and that would be obvious even to a small child. If people born between 1940 and 1953 are healthier than people born after 1953, there could be any number of explanations for it. There is no more reason to attribute it to sugar rationing than to anything else, including meat rationing and egg rationing. There is no reason to assume that it has anything to do with diet, nor indeed to anything people do in their first two years of life. These individuals grew up and lived in different times when the prevalence of some diseases declined and others increased. Focusing on one event and giving it the credit for everything that happened afterwards is so stupid that there is a name for it: the post hoc ergo proctor hoc fallacy.
This struck me as a new low for nutritional science, but I soon discovered that there was nothing new about it. It turns out to be the latest in a long line of studies making claims about the risks of very young children eating sugar, all of which are based on nothing more than the “natural experiment” of wartime rationing in the United Kingdom. Six such studies were published last month alone. To date, it has been claimed that the before-and-after experience of sugar rationing shows that sugar consumption early in life (and in utero!) causes heart disease (October 2025), cancer (November 2025), worse respiratory health (December 2025), heart failure (January 2026), frailty (July 2026), depression and anxiety (July 2026), dementia (again) (July 2026), early on-set cancer (July 2026), biological ageing and mortality (July 2026) and metabolic dysfunction-associated steatotic liver disease and major adverse liver outcomes (August 2026).
All of these studies use data from approximately 64,000 people who are registered with the UK Biobank, none of whom have been asked how much sugar they actually consumed when they were children, and they are all written by Chinese researchers. Five of them have the same lead author, the aforementioned Jiazhen Zheng. These academics have taken their lead from the work of Tadeja Gracner and Paul Gertler, two economists based in California, who published a working paper titled “The Sweet Life: The Long-Term Effects of a Sugar-Rich Early Childhood” in 2023. The paper claimed that the end of sugar rationing in the UK was associated with higher rates of diabetes, arthritis and chronic inflammation. In 2024, they used the UK Biobank data and concluded, in a study published in Science, that sugar rationing in early life reduced type 2 diabetes and hypertension risk by 35 per cent.
There is now a cottage industry of junk science in which the UK Biobank is dredged for statistics that support this narrative. People who are registered with the UK Biobank are volunteers and it is well known that they are not representative of the general population, but it is an easily accessible source of data and it saves epidemiologists the effort of conducting their own research. It is a valuable resource if used properly, but in the wrong hands it is the handmaid of quackery.
All observational epidemiology rests on the conceit that correlation equals causation. Conscientious academics are careful to include a sentence or two in their studies cautioning the reader against drawing causal inferences, but this is mere throat-clearing. They wouldn’t be publishing their findings if they didn’t think that X caused Y. By the time the public hears about the study via the media, all caution has been thrown to the wind.
These sugar rationing studies are epidemiology at its dumbest. They rest solely on a supposed change in disease prevalence that the authors attribute to an arbitrarily chosen event that supposedly coincided with it. You might as well say that having your town bombed early in life protects you against dementia.
And that brings us to the punchline. Last year, another group of Chinese researchers used the same UK Biobank data in a study and concluded, in all seriousness, that exposure to nighttime blackouts early in life reduces the risk of having a mental disorder by 21 per cent and having diabetes by 38 per cent. They said that their findings demonstrated “the critical role of early-life LAN [light at night] restriction in shaping long-term health trajectories.” The benefits of blackouts were so powerful that even people who were only foetuses at the time had better health outcomes!
More research is needed before we ban headlights and street lamps, but this is clearly an exciting new field of science. The possibilities are almost endless. It would be interesting to see whether listening to Vera Lynn on the radio early in life leads to lower rates of heart disease and cancer. I have a feeling that it might.
This article first appeared in The Critic
