Have the benefits of intermittent fasting been overhyped?
22 minutes ago
BBC/ Javier Hirschfeld/ Getty Images
Proving that intermittent fasting works, like any lifestyle intervention, is notoriously difficult. But new technologies will better help us to understand what works and what doesn’t
Every few months, it seems, bring a new round of debate on the benefits of intermittent fasting
This popular and wide-ranging diet, which involves either fasting on certain days of the week, or restricting food intake to certain hours of the day, has acquired a reputation for making weight loss easier than standard calorie reduction. But the alleged benefits are meant to extend much further, for example reducing the risk of cancer and cognitive decline. Little wonder the regime has become so popular
Recent evidence, however, has raised questions about the strength of these claims. Depending on the study, intermittent fasting may be shown to be more effective than other diets, or less effective. One even suggested that the practice can be harmful, increasing the risk of cardiovascular disease. The scientists don’t always agree on all the facts, says Krista Varady, a professor of nutrition at the University of Illinois at Chicago, never mind the influencers. “It’s not magic.”
So has intermittent fasting been oversold? Or is it a promising diet that could add years to our lives? And why has the evidence been so inconsistent? The biggest obstacle to making the data more robust, many researchers say, is also the hardest to solve: “We have no idea what people are actually eating,” says Varady. And that applies to many other, if not all, dietary interventions
Prehistoric eating habits
Intermittent fasting is meant to reflect how our ancestors evolved to eat. “We didn’t evolve in an environment where we ate three meals a day plus two snacks,” says Mark Mattson, a neuroscientist whose work at Johns Hopkins University, in Baltimore, was foundational to the field
Fasting causes the metabolism to shift from first burning sugar, a process known as glycolysis, to burning fat circulating in the blood. This metabolic switch changes cellular behaviour. Healthy cells start recycling worn-out and damaged proteins – a process called autophagy – which has the effect of rejuvenating cells.
In theory, this translates directly into the claimed health benefits. In an influential review published in theNew England Journal of Medicine, Mattson tied this metabolic switch to increased longevity and a lower incidence of diseases, including diabetes and cancer
There is only one problem. The most robust evidence for this neat narrative came from studying mice, flies and worms. For them, intermittent fasting was a silver bullet. Mice could eat the same number of calories as a different group who fed freely, but as long as they kept their eating hours restricted, they stayed svelte and healthy into old age while their counterparts became obese and sick
Unfortunately, the evidence from human studies has been less conclusive
Of mice and men
Consider the effects on obesity. Depending on the kind of intermittent fasting, “it produces around 5% weight loss” in humans, says Varady. This is “just as effective as any other diet strategy”, she says, but the human evidence has yet to replicate the dramatic outcomes seen in mice (or social media anecdotes)
Even a small amount of weight loss from intermittent fasting can undoubtedly lead to metabolic benefits. “There are some decreases in metabolic risk markers,” says Varady, including cholesteroland blood pressure. It even seems to lessen the symptoms ofpolycystic ovary syndrome. The less healthy a person is at the start of the intervention, she says, the more health benefits they will accrue. But these, too, tend to be less pronounced in humans compared with those observed in mice. What’s more, the more intense forms of intermittent fasting – such as eating every other day – may make a person less metabolically healthy by stripping away lean body mass.
A familiar story emerges for the claims of improved cognition: there is a noticeable gap between mouse and human data. And the cancer research yields similarly mixed messages. Animal trials had linked intermittent fasting to cancer prevention as well as greater resilience to withstand toxic chemotherapy. But a major study into the mechanism that worked in mice failed in humans. (Researchers had devised a special diet meant to mimic the biological response to long-term fasting. In theory, it would make healthy cells more resilient while weakening cancer cells. However, the subjects never entered the intended metabolic switch.)
Getty Images
That said, the 5:2 diet might let women with metastatic breast cancer fare better when receiving chemotherapy, both in terms of their quality of life and a slowed progression of disease.
Among the confusing and contradictory findings, one clear pattern emerges: the benefits of intermittent fasting are simply less impressive in humans than they have been in animals. Why?
Partial blame goes to fundamental biological differences. For example, a mouse’s metabolic rate is far higher than a human’s, which may be misleading. “A 16-hour fast in a rodent is really equivalent to a several-day fast in a human because their metabolic rate is so much higher,” says Courtney Peterson, an associate professor of nutrition at Harvard School of Public Health.
Then there’s the fact that the mice in these trials are usually genetic copies of each other, says Varady, which fails to reflect the variability you see in humans. “They’re just a really bad model,” she says
Perhaps the largest impediment, however, is not biology. It’s information. “It’s just so hard to know what people are actually eating,” says Peterson
Flawed memories
Part of the reason the link between diet and effects was so dramatic in animals was that there was no question about what they were eating. Mice and other lab animals have no choice but to “stick to the diet”. They eat exactly and only when and how much food a researcher provides. This makes it easy to draw direct links between diet and biological or medical effects
In the vast majority of trials involving human participants, however, you can’t control their nutrient intake. You can instruct them on the diet you want them to follow, and give them a food diary, either paper or sometimes virtual. Then they go back to living their lives and making their own choices. “In nutrition research, we have no way of measuring what people are actually eating,” says Varady. “So we’re completely reliant on them just being honest.”
But even honest people don’t have perfect memories. They may forget to enter a meal into their food diary (which looks like time restricted eating). Or, owing to feelings of shame around not adhering, they might leave out one or two things that they may have eaten during a period in which they were meant to be fasting. Varady estimates that only 50% of her research subjects even return the food records. The rest sometimes fudge it
Varady says she inadvertently received this education in her earliest days as a young nutrition professor. As she passed the room where her study subjects waited to hand in their food records, she saw one participant hastily scribbling an entire week’s worth of meals. “They were just BSing the whole thing,” she says. “And I was like, welp, that’s just bogus data. And then I was like, ‘how many other people are doing this?'”
In 2025, an international team of researchers developed an equation that was able to hint at the scale of the problem, suggesting that about 30% underreport their food intake, says Thomas Wilson, a nutrition researcher at Aberystwyth University. “It’s very rare that people overreport,” he adds
Getty Images
Doubly labelled water studies are the gold standard of tracking exactly how much energy a person takes in. In these tests, participants drink water enriched with safe stable isotopes of hydrogen and oxygen. By measuring how quickly these leave the body in sweat and urine, scientists can calculate what their bodies have used over that period.
The 2025 equation suggested that on average people underreport between 400 and more than 800 calories. But mileage may vary. When Varady used these to estimate how much a person’s recollections of their food intake diverged from reality, the results were shocking. “People are reporting eating between 1,600 to 2,000 calories per day,” she says. “And then we get back the actual data from the doubly labelled water, and it’s like 3,500 calories.”
Even nutritionists have a hard time reporting exact nutrition values, says Peterson. Portion sizes vary widely. To make matters worse, the nutrition composition tables that scientists use to estimate a food’s contents – such as how many macronutrients a burger contains – only contain averages, not product-specific detail
These limitation are so well known and so vexing that there have been calls for journals to stop publishing studies based on self-reported dietary intake. Practically speaking, however, nutritionists can’t abandon records and surveys. “It’s kind of all we got,” says Varady.
But perhaps not for much longer
Smart forks, necklaces and tooth sensors
Not all human studies rely on self-reporting. Inpatient feeding trials – a rigorous type of study where researchers control exactly what participants eat – can offer the same objective data as mouse studies. Every calorie is carefully controlled, food is weighed, and if you feel like a cheeky snack during the hours you are meant to be fasting, well, too bad. There’s no fridge
Peterson has done these trials. In her studies, intermittent fasting has shown stronger effects than in the wider literature, showing lower blood sugar, blood pressure, a reduction in a type of molecular damage called oxidative stress, and an increase in autophagy.
However, costs meant that they had to be very small and very short, which somewhat limits the conclusions we can draw from them. (Few participants have the time to disappear into a food trial for six months.)
But what if you could mimic these highly surveilled conditions with technology? Increasingly, nutritionists are looking toinnovative approaches that rely less – or not at all – on self-reporting. A number of new technologies can now provide a more objective record of what people eat.
Some studies ask participants to take photos of everything that they eat, for instance. But this also relies on the participant actively recording that data, so others use cameras mounted on glasses or necklaces, with names like FitNibble and Autographer, which take a photo every seven seconds or whenever they sense chewing. However, cameras can raise privacy concerns
Getty Images
Other sensors are in development to monitor meals and snacks without cameras. These include necklaces that identify chewing, smart forks that detect when and how much food someone is transporting to their mouths, and tooth-mounted nutrient sensors. Most of these ideas are still in the prototype stage. For now, the most practical solution is to take regular blood and urine tests that monitor the macronutrients being processed by the body. “Urine and blood analysis is going to give you much more composition information,” says Wilson.
Many new and old technologies are available, says Wilson, but no single one will solve the problem by itself. However, they could be combined into a take-home panopticon. In astudy published earlier this year, Wilson and his colleagues reviewed myriad different methods and technologies and suggested that in combination, they might do the trick
• Why weight-loss drugs aren’t a quick fix
• Why eating fibre is good for the brain
• How the nutritional benefits of foods change as you age
The team is now in the final stages of a study testing whether these combinations – including subjective food diaries, objective biomarkers derived from blood and urine tests, and quasi-subjective wearables like cameras – could help nutrition researchers get inpatient-quality results even as patients live their normal lives outside. “I don’t want to be having to ask you to interpret what you ate yesterday,” says Wilson
The need for knowledge
Mattson thinks clinical trials of intermittent fasting should go further still, to include data not just on compliance, food composition, and calories, but also physical activity and sleep, activities that are also “precisely controlled for in animal studies”
While researchers and companies try to thread a precarious needle between privacy invasion and understanding exact eating patterns, social media influencers continue to conflate the mouse results with human results, and overhype some of the benefits of intermittent fasting. “I’m thinking of starting a YouTube channel because I see so much bogus information,” says Varady.
The problem extends beyond intermittent fasting. “That’s why people get so frustrated with nutrition research,” she explains. She points to cholesterol advice, which has for years oscillated between being branded a killer to being fine. “It’s because we have no idea what people are eating.”
—
For trusted insights on health and wellbeing, sign up to theLive Well For LongerandSix Steps to Calmcourses
For more science, technology, environment and health stories from the BBC, follow us onFacebookandInstagram
Health
Diet
Food
Science
Features
