The fasting-mimicking diet: what its human trials measured, and what they didn't

The fasting-mimicking diet — a low-calorie, low-protein, high-fat regimen eaten for five consecutive days a month — improves several cardiometabolic risk markers in small randomised trials, and one exploratory analysis links three cycles to a 2.5-year drop in an estimated "biological age" [s1][s2]. What no human trial has shown is the thing the concept is famous for: that cycles of the diet prevent an age-related disease or extend life [s1][s2]. The human evidence is real but early, built on surrogate markers rather than outcomes.

The diet is designed to trigger the fasting state — lowering glucose and the growth factor IGF-1 — while still providing food, on the theory that this makes prolonged fasting tolerable enough to repeat. Its selling point over continuous calorie restriction is adherence: a person eats normally for most of the month, which is easier to sustain than a permanent cut of the kind that even motivated volunteers struggle to maintain. In mice, periodic cycles reduce inflammation and extend lifespan; the human question is whether the biomarker changes seen so far translate into anything durable [s2].

The first randomised trial

The foundational human study randomised 100 generally healthy US adults to three months of their usual diet or to the fasting-mimicking diet for five consecutive days each month over three months [s1]. Three cycles reduced body weight, trunk fat and total body fat, lowered blood pressure and decreased IGF-1, with no serious adverse effects reported [s1]. After the control group crossed over to the diet, 71 subjects had completed three cycles; a post-hoc analysis found that body-mass index, blood pressure, fasting glucose, IGF-1, triglycerides, total and LDL cholesterol and C-reactive protein were more beneficially affected in participants who started out at higher risk than in those who did not [s1]. The authors were explicit about the limits — the study measured risk factors, not disease, and they called for larger trials in people with diagnosed conditions before any claim about prevention or treatment [s1].

The "biological age" result, read carefully

The most quoted recent finding needs its qualifiers attached. A 2024 paper reported a secondary and exploratory analysis of blood samples from a randomised trial (registered NCT02158897): three cycles of the diet were associated with reduced insulin resistance and other pre-diabetes markers, lower liver fat on MRI, and a higher lymphoid-to-myeloid ratio, an indicator of a younger immune profile [s2]. Using a validated biological-age measure, three cycles were associated with a decrease of 2.5 years in median biological age, independent of weight loss, and the authors reported nearly identical findings in a second study (NCT04150159) [s2].

Two words in that description do most of the work: "secondary and exploratory." These were not the trials' pre-specified primary endpoints, which makes the biological-age figure a hypothesis-generating result rather than a confirmed effect. A biological-age estimate is itself a surrogate — a prediction of risk, not a recorded health outcome — so a 2.5-year drop is a promising signal awaiting a trial powered to test whether it changes what actually happens to people [s2].

What the evidence does not cover

There is also a signal worth keeping in view: in the first trial, the benefits concentrated in people who began at higher risk, and were smaller in those who were already healthy [s1]. That pattern — a bigger effect where there is more to correct — is common to many lifestyle interventions and means a low-risk person may see little from the same cycles, a caveat the marketing rarely carries.

Both trials are small, short and centred on markers. Neither tracked participants long enough to record fewer heart attacks, cancers or deaths, and neither compared the branded five-day regimen against simpler, cheaper approaches such as ordinary calorie restriction or other eating-window schedules [s1][s2]. The diet is also commercialised — sold as a packaged five-day kit — which is a reason to hold its marketing to the same evidentiary line as any supplement: the risk-marker data are genuine, but "reduces biological age" is a claim about a surrogate, not a demonstrated extension of healthy life [s2].

For the wider fasting literature, the site has assessed the network meta-analysis of intermittent fasting and a six-month multi-omics fasting trial; for sustained rather than periodic restriction, see the human evidence on caloric restriction and healthspan and its muscle sub-study.