What is a fasting mimicking diet? Four or five days of eating, not fasting

A fasting mimicking diet is a four to five day eating plan, low in calories, carbohydrates and protein and relatively high in fat, taken in repeated cycles rather than followed every day [3]. Pooling 11 randomised controlled trials, a 2025 review found the fasting mimicking diet moved four blood measurements and left the rest where they were [6].

Average glycated haemoglobin - HbA1c, a blood test that reflects average blood sugar over the previous two to three months - fell by about 8.6 units on the mmol/mol scale that test is reported in [6]. The top number in a blood pressure reading fell by about 4 mmHg (millimetres of mercury, the unit blood pressure is measured in) and the bottom number by about 2, though the pooled range ran from a drop of nearly 8 points down to almost none [6]. Insulin-like growth factor 1, a signal in the blood that tells cells to grow and divide, fell by about 19 nanograms per millilitre [6]. None of the other cardiovascular risk factors the review examined changed measurably [6].

The calorie level is where it parts company with a true fast. One hospital programme built from ordinary recipes supplied 34 to 54% of a normal day's calories across four days plus the day of treatment [2]. In the 27 breast cancer patients who completed three cycles of it, body mass fell by 2.04 kg and fat mass by 1.88 kg, while muscle mass barely moved, by 0.05 kg [2]. Visceral fat area, the fat packed around the organs inside the abdomen, fell by about 15%, and waist measurement fell too [2]. Severe side effects were uncommon: five of 84 diet cycles produced one [2].

The clearest randomised human evidence sits in cancer care rather than in healthy ageing. In the DIRECT trial, 131 women with HER2-negative stage II or III breast cancer were randomly assigned to the diet or their usual food for three days before and during each round of chemotherapy, and the odds of a tumour shrinking on scans were about three times higher in the diet group, with no increase in side effects [1]. A separate comparison set 14 patients with advanced triple-negative breast cancer who added the diet to carboplatin-based chemotherapy against 76 who did not, and found a median survival of 30.3 months against 17.2 months [4].

Against intermittent fasting, where eating is squeezed into part of the day or week, the only direct comparison is in mice. Both patterns switched on the same main fat-burning pathway in the liver, but the fasting mimicking diet pushed blood sugar lower and ketone levels higher, and repeated cycles blunted the fat-burning gene response that a single cycle produced [5].

The body reads several days of very low calories, protein and carbohydrate as a fast and changes fuel. In mice given either diet pattern, blood sugar dropped while ketone bodies - the fuel the liver makes out of fat when sugar runs short - rose, along with free fatty acids and FGF21, a hormone the liver releases during fasting that shifts the body towards burning fat [5].

Both patterns activated the same top pathway in the liver, PPAR-alpha, a master switch for the genes that burn fat, and the fasting mimicking diet moved those genes further than intermittent fasting did [5]. Repeating the cycles blunted that response, so the fourth cycle was not doing what the first one did [5].

The second strand is growth signalling. Insulin-like growth factor 1 falls during these cycles [6][2], and in cancer biology the proposed route is that lower blood sugar plus weaker growth signalling leaves tumour cells more exposed to chemotherapy while normal cells drop into a low-growth, protected state [4]. In the DIRECT trial the diet also reduced chemotherapy-related DNA damage in T-lymphocytes, a type of white blood cell [1].

That signal is not durable. In the single-arm hospital trial the drop in insulin-like growth factor 1 was clear after the first cycle and the second, and no longer clear after the third [2].

The evidence is narrower than the diet's reputation. Most human data comes from people with cancer, on chemotherapy, under hospital supervision [1][2][4] - not from healthy middle-aged adults hoping to slow ageing.

The studies here do not cover:

• Healthy people followed for years. No study here measured lifespan, heart attacks or any disease outcome in healthy adults; the outcomes are blood markers, plus tumour response and survival in cancer patients [1][4][6].
• People with diabetes and people who are underweight, both excluded from the DIRECT trial [1].
• Anyone comparing it with intermittent fasting on human outcomes: that comparison exists only in mice, which lost up to 20% of their body weight before regaining it on refeeding days [5].

The survival figure looks stronger than its design. It compares 14 patients who used the diet against 76 who did not, at a single centre, without randomisation, so the two groups may have differed in ways the adjustment could not remove [4].

The pooled review did not report what the participants' blood sugar or blood pressure was to begin with, so a fall of 8.6 units on the blood sugar scale cannot be turned into an everyday number, and there is no basis for assuming it means the same thing in someone whose readings were already normal [6].

Criticism of the underlying trials is on the record. A 2020 review in a nutrition journal argued the human studies behind the diet were small enough, and carried enough design problems, that the claims about heart and metabolic markers were not established [3]. The 2025 pooled analysis still described findings in its own field as inconsistent [6]. Cancer treatment is decided with an oncologist, and multi-day calorie restriction is not a neutral act for everyone.

What the trials tested was a defined menu for four or five days, repeated after a gap of weeks, with ordinary eating in between [3][5]. Someone weighing it against a water fast is weighing two different things: the trial diets handed over between about a third and a half of a normal day's calories every day, with protein and carbohydrate held down and fat proportionally higher [2][3]. The hospital version was delivered as recipes, with the cycles timed around chemotherapy and side effects monitored [2]. People with diabetes, a low body weight, a history of eating disorders or a cancer diagnosis fall into the group the trials either excluded outright or handled under medical supervision [1][2].