Intermittent fasting for longevity

Intermittent fasting (IF) means switching between stretches of time when you eat and stretches when you do not. Most diets change what is on your plate, while IF changes when the plate appears. Under that one label sit methods that differ far more than the name suggests. Some only stretch the overnight fast, others cut calories hard on certain days, and a few are multi-day programs that belong in a clinic.

16:8 and other daily windows

The most common form is 16:8, also called time-restricted eating (TRE): you fast for 16 hours and eat within an 8-hour window, every day. In practice that usually means skipping breakfast or eating dinner early, and most studies on daily fasting use exactly this window. A gentler 14:10 is a good place to start, while 18:6 is the stricter step up. Even shorter windows have been tested, and in a small trial in adults with obesity, 4-hour and 6-hour windows led to similar weight loss over 8 weeks.^[11]

5:2 and OMAD

With 5:2 you eat normally on five days a week and drop to 500 to 600 calories on two days that do not follow each other. That makes it closer to a mild weekly calorie cut than to a fast in the strict sense, which can suit people who dislike daily rules. OMAD (one meal a day) sits at the strict end: one large meal a day, usually eaten within one to two hours, and nothing else. With one meal it is hard to fit in enough protein and nutrients, and there is almost no long-term data, so it is not a starter protocol.

Alternate-day fasting

Alternate-day fasting (ADF) swaps a fast day (nothing, or about 500 kcal) with a normal day. Its best-known test is a 4-week randomized controlled trial (RCT, where a lottery decides who gets which diet) in 60 healthy, non-obese adults in Graz. In that trial, alternate-day fasting cut calories by about a third and lowered weight, trunk fat and blood pressure.^[10]

The fasting-mimicking diet

The fasting-mimicking diet (FMD) is a 5-day plan of very few calories, about 1,100 kcal on day one and about 720 kcal a day after that. It is designed to trigger fasting effects while you still eat a little.^[15] Its trials are small and come mostly from the developer's own lab.^{[7, 15]}

Periodic and clinic fasts

Longer fasts of 24 to 72 hours, done once a month or once a quarter, are a different intervention again. So are the multi-day fasts of the German clinic tradition. At the Buchinger Wilhelmi clinic in Überlingen, for example, supervised fasts last 4 to 21 days on 200 to 250 kcal a day. The clinic's own observational study of 1,422 people, in which the team watched what happened without a control group, is a large safety dataset.^[14]

That safety record reads well at first glance, since doctors recorded adverse events in under 1 percent of people. Milder complaints were common, though, mostly in the first days: about 15 percent reported sleep problems and about 14 percent fatigue. Two people were admitted to hospital.^[14] The study says nothing about lifespan, and fasts of three days or longer belong under medical supervision.

How the methods compare

The table puts the methods side by side, with the evidence for each in one line.

Sources: network meta-analysis^[9]; fasting-mimicking diet trials^{[7, 15]}

Of all these, 16:8 is the most studied daily pattern and practical for many people, which is why it is the usual target. It is best reached step by step from a 12-hour overnight fast, and which plan you stick with depends on your schedule more than on the label.

Nobody knows yet whether intermittent fasting slows aging in humans. Fasting extends life in many lab animals, but in people the trials only show better risk markers, and most of that tracks the weight lost.^[9] No trial has shown that people who fast live longer, so the anti-aging case rests on animal studies, proposed mechanisms and a handful of small human studies.

What the mouse studies actually show

Calorie restriction extends healthy lifespan in many lab species, from yeast to mice.^[2] The popular line that "mice on intermittent fasting live 30 percent longer" mixes up two things, though: fasting on its own, and eating less at the right time. Fasting on its own helps, but less. When male mice ate the same calories compressed into one daily meal, they lived about 11 percent longer.^[1] In genetically diverse female mice, one or two fasting days a week worked less well than calorie restriction, and the effect varied a lot with the animals' genes.^[2]

The headline numbers come from eating less at the right time. In male mice, cutting calories by 30 percent alone added about 10 percent to lifespan. The gain reached about 35 percent when the calorie cut was combined with feeding during the mice's active phase.^[3] So the largest gains came from eating less and in step with the body clock at once, a rather more demanding regime than skipping breakfast.

Which mechanisms hold up in people

Those mouse results come with a set of proposed mechanisms, and this is where animal and human data most often get blended. In people, one change is well documented. Somewhere between 12 and 36 hours after the last meal, the liver's glycogen (stored carbohydrate) runs low. The body then shifts from burning mostly glucose to burning fat and making ketones, a fuel the liver makes from fat.^[20] When exactly that happens depends on you, your activity and what you ate last.

Growth signals are the next step, and here the type of fast matters. Three cycles of the 5-day fasting-mimicking diet lowered IGF-1 (insulin-like growth factor 1, a hormone that tells tissues to grow).^[15] Daily time-restricted eating does not seem to, since an 8-week trial of 4- and 6-hour windows found no change in IGF-1.^[27] In human muscle, a 72-hour fast cut signaling through mTOR, a molecular switch that tells cells to grow when food is around, by about half.^[13] Whether lower IGF-1 or mTOR adds years to a human life is not known, because that link comes from animals.

Two other popular claims rest on even less human data. Sirtuins, proteins involved in DNA repair, metabolism and stress resistance, switch on during fasting in animal and cell studies.^[20] For inflammation the human results are mixed. An 8-week window trial saw no change in the marker hsCRP,^[27] while long-term alternate-day fasters in the Graz study had lower levels of another one.^[10]

Rather than track one marker at a time, the fasting-mimicking diet's developers combined many routine blood and clinical markers into a biological-age score. In two pooled trials, that score fell by a median of about 2.6 years after three monthly 5-day cycles.^[7] The result comes from a secondary, exploratory before-and-after analysis, and in one of the two trials about a third of participants saw no drop at all. It shows a change in a calculated score rather than slower aging.

Autophagy: real, but oversold

Autophagy is your cells' built-in recycling program for damaged proteins and organelles (the small working parts inside a cell), and in mice fasting boosts it strongly.^{[12, 36]} Fasting may nudge autophagy up in humans too, but the human data are few and small, and nobody knows when it starts or how it differs between tissues. The "autophagy switches on at 16 hours" line from social media is stretched from mouse data.

The human studies show why. In 8 healthy men on a 72-hour fast, autophagy was only a side question of a study on mTOR. Their muscle markers moved a little and in different directions, with no on-off switch in sight.^[13] A larger study in adults with obesity measured autophagic flux in blood cells, which is how much material the cells actually break down. After 6 months of intermittent fasting plus time-restricted eating, flux was higher than in the control group.^[36] That comparison was post hoc, meaning it was not planned in advance, and flux did not rise within the fasting group itself.

Neither study points to a switch that flips at a set hour, so pinning autophagy to a magic hour is marketing rather than a study result. Our autophagy guide goes through the human data in more detail. For fasting itself, the firmer human evidence lies elsewhere, in weight and blood markers measured over weeks and months.

Intermittent fasting can help with weight loss, but it works about as well as ordinary calorie cutting, not better. It also brings modest improvements in blood pressure, blood fats and insulin sensitivity (how well your cells respond to insulin). Those markers are what the human trials measured, and most of the improvement tracks the weight people lost.^[9]

Weight: as good as calorie cutting, not better

IF works by creating a calorie deficit, so the fair comparison is plain daily calorie cutting. A 2025 network meta-analysis, which compares many diets across trials at once, pooled 99 trials and found that fasting styles matched calorie cutting for weight loss.^[9] Only alternate-day fasting did a little better, by about 1.29 kg, and only in trials shorter than 24 weeks. Most of the trials were that short, which leaves long-term effects open. In those lasting 24 weeks or longer, no fasting pattern beat plain calorie cutting, and the authors call for longer trials.^[9]

The 2026 Cochrane review compared intermittent fasting with regular diet advice instead and reached a similar result. Across 22 trials in adults with overweight or obesity, fasting made little to no difference to weight loss, on low-certainty evidence.^[16] A fasting window is therefore one way to eat less, not a way around the calorie balance.

Blood sugar, blood pressure and blood fats

Daily eating windows can also improve how the body handles insulin. In an 8-week trial of 4- and 6-hour windows, insulin levels and insulin resistance (cells responding poorly to insulin) both improved.^[27] The changes tracked how well people kept to the window, but whether they lower diabetes risk in the long run is not known. With existing type 2 diabetes, fasting belongs under medical supervision.

Blood pressure and blood fats tend to improve modestly, mostly in line with the weight lost. In the same 99-trial analysis, daily windows and whole fasting days brought no extra benefit over plain calorie cutting. Alternate-day fasting gave small additional improvements in some blood fats and in systolic blood pressure (the upper number). For HDL ("good" cholesterol) and HbA1c (a blood sugar average over about three months), no fasting style even beat eating freely.^[9]

The 4-week Graz trial of alternate-day fasting fits that picture only in part, because blood pressure and heart rate fell but blood lipids did not change. Lower LDL and triglycerides showed up only in people who had already fasted this way for over six months, a non-randomized comparison.^[10]

Brain and cells

Beyond blood markers, brain and cell benefits are the least proven part. In animal studies fasting raises BDNF, a protein that supports nerve cells, but whether that matters for human brains is not known.^[20] The same goes for autophagy, the cells' recycling program: whether fasting-driven autophagy lowers cancer risk in humans is unknown.

Muscle: the TREAT trial

Muscle is the possible cost when the window is all you change, and the TREAT trial of 16:8 shows how. After 12 weeks, 16:8 with no exercise program gave essentially no extra weight loss over three regular meals a day, so the trial missed its main goal.^[4] What it did show was where the weight came from. In a small subgroup measured in person, the 16:8 group lost about 1.7 kg, and about 65 percent of that was lean mass (everything in the body except fat). Total lean mass did not differ significantly from the control group, but the 16:8 group lost more lean mass from arms and legs.

Pooled trials show the same cost and a hint of how to avoid it. Time-restricted eating trims about 0.6 kg of fat-free mass (lean mass by another name), on its own or added to calorie cutting.^[22] In small trials where people also exercised, it did not significantly change fat-free mass, though the authors say more studies are needed.^[23]

That makes protein and training the guardrail. About 1.2 to 1.6 g protein per kg body weight a day, ideally 25 to 30 g per meal, is linked to better appetite control and weight loss.^[28] Resistance training 2 to 3 times a week is a sensible partner to it. Both matter most if you are over 50 or at any risk of sarcopenia (age-related muscle loss).

Older adults have a target of their own. From 65, the PROT-AGE group advises at least 1.0 to 1.2 g/kg a day, and 1.2 to 1.5 g/kg with acute or chronic illness.^[8] The exception is severe kidney disease not on dialysis, where protein may need to be limited.^[8] If you have kidney disease, agree your protein target with your doctor.

Mortality: three observational studies, three directions

The data on deaths come from observational studies, which watch what people do rather than assign a diet, and they do not agree. The table sets the three side by side.

Sources:^[6],^[17].^[24]

The Shanghai result first made headlines as a 2024 conference abstract.^[25] The peer-reviewed analysis that followed compared windows under 8 hours with 12 to 14 hours.^[24] The authors say they cannot tell whether the short window itself or other factors behind it explain the higher cardiovascular risk.^[24] The Utah result, in turn, may reflect religion and lifestyle more than fasting.^[6] None of the three studies can show cause and effect, so they prove neither that fasting protects you nor that it harms you.

Skip intermittent fasting if you are pregnant or breastfeeding, under 18, underweight or have a history of eating disorders. Do not fast on diabetes medication without medical supervision either. For most healthy adults the side effects are mild and short-lived, but some groups face real risks.

Common side effects (usually short-term)

• Hunger, especially in the first 1 to 2 weeks
• Irritability and trouble concentrating
• Headaches, whose cause is unclear: they are more common in people who already get headaches and in longer fasts, and they usually ease after eating.^{[19, 30]}
• Low energy at first

Who should not fast, or only with a doctor

• Pregnant or breastfeeding women
• Children and teenagers, because they are still growing
• People who are underweight
• People with a history of eating disorders, since IF can trigger or worsen anorexia, bulimia or binge eating
• People with type 1 diabetes, or on insulin, sulfonylureas or SGLT2 inhibitors, without medical supervision
• People with kidney disease: check the protein advice with your doctor first^[8]

Medication: low blood sugar and ketoacidosis

Diabetes medication carries the most serious risks, so with the drugs below, fast only under medical supervision. If you take insulin or sulfonylureas, fasting can drop your blood sugar dangerously low. If you take an SGLT2 inhibitor (empagliflozin, dapagliflozin and others), prolonged fasting and eating little raise the risk of ketoacidosis. Ketoacidosis is a dangerous acid build-up that can happen even with normal blood sugar^[32] If you take any regular medication, especially for blood sugar or blood pressure, talk to your doctor before you start.

Nutrients and hormones

Shorter eating windows make it harder to get enough protein and nutrients, above all with one meal a day. Food quality therefore matters more on a fasting schedule, not less.

Some women also report cycle changes on strict schedules. If that happens to you, a gentler 14:10 is a cautious choice, though trial data on it are thin.

Longer fasts need the most care. Fasts of three days or longer belong in a supervised setting such as a fasting clinic. If you take medication or have a medical condition, talk to a doctor before any fast over 24 hours.

Start with a 12-hour overnight fast, stretch to 14:10 by week 3 to 4, and only move to 16:8 from week 5 if it still feels easy. Along the way, drink enough and eat about 1.2 to 1.6 g protein per kg body weight a day plus plenty of fiber.^[28] If your sleep, performance or cycle suffers, shorten the fast again, and skip OMAD unless you are supervised.

Week 1 to 2: build up slowly

Start with a 12-hour overnight fast, say 7 pm to 7 am. Most of it happens while you sleep, so this step mainly gets you used to a fixed last meal.

Week 3 to 4: stretch to 14:10

Push breakfast later or dinner earlier to reach a 14-hour fast. For many people this is a setup they can keep long term.

Week 5 and beyond: try 16:8 if you want

If 14:10 feels easy, go to 16:8. Most people skip breakfast and eat from noon to 8 pm, while others prefer an early window from 8 am to 4 pm. Skip OMAD unless a clinician is supervising you.

Early or late window?

Timing matters a little, and adherence matters more. In a 12-week trial of 197 adults, early, late and self-chosen 8-hour windows did no better than each other at shrinking visceral fat, the fat around the organs. None of them beat Mediterranean-diet advice alone either.^[18] For body weight and fasting insulin, though, pooled trials gave earlier windows a small edge. People whose last meal came before 5 pm lost about 1 kg more than those eating past 7 pm, and their fasting insulin fell more.^[26] Pick the window you can keep, and lean early if it fits your day.

The table shows three common 16:8 windows and who they suit.

Drinks, salt and fasting headaches

Water, black coffee and plain tea are fine during the fast, and you should drink enough of them. Plain food covers salt and potassium for most people. Skip added salt if you have high blood pressure, heart or kidney disease. WHO advises adults to stay under 2 g sodium (5 g salt) a day in total.^[33]

If you feel dizzy during a fast, lie down until it passes, get up slowly and drink water. If you have diabetes, check your blood sugar. Anyone whose dizziness will not go away or keeps coming back should see a doctor.^[37]

Headaches are the other common complaint in the first weeks, and why fasting causes them is not settled. Low blood sugar and caffeine withdrawal are the usual suspects, but the international headache classification says fasting headache does not appear to be caused by either.^{[19, 30]} Coffee is a separate issue. If you regularly have more than 200 mg caffeine a day, skipping it can trigger a caffeine-withdrawal headache within 24 hours, so keep your usual coffee.^[31]

What to eat in your window

Whatever your window, weight change still comes down to your weekly calorie balance.^[9] You can estimate your maintenance calories with our free calorie calculator, and these habits help you stay within them:

• Do not compensate by overeating during the window.
• Lean on protein and fiber to stay full.
• Quality still matters, because IF is not a free pass for junk food.
• Stay flexible, since social plans may mean shifting your schedule.
• Track your energy, sleep and mood. If you feel unwell, eat, and if IF is not working after a fair trial, try something else.

Break-fast meal template (DACH edition). It delivers about 25 to 40 g protein depending on the combination, or roughly 20 g more with a scoop of protein powder. It also brings about 10 g fiber or more, at prices you can manage at Rewe, Edeka or Aldi:

• 3 eggs or 150 g Skyr / Magerquark
• 50 g rolled oats or 2 Vollkorn-Brötchen (wholegrain rolls)
• 100 g berries: fresh, or frozen raspberries heated through first, as Germany's food-safety institute BfR advises for frozen berries,^[35] which works well as warm berries over oats
• 1 to 2 tbsp olive oil or 30 g walnuts
• Optional: 1 scoop whey or plant protein

When to ease off

Stop signals (women especially). These are rules of thumb we use, not validated cut-offs. Pause or shorten the fast if:

• HRV (heart rate variability, the beat-to-beat rhythm your watch tracks) drops more than 10 percent from your baseline for 5+ consecutive days
• Resting heart rate climbs 5+ bpm
• Your menstrual cycle shortens or lengthens by more than 3 days
• You feel wired and exhausted at the same time most mornings
• Sleep efficiency drops below 80 percent

The cycle deserves its own word, because there is little trial data on cycle-timed fasting. Some practitioners suggest avoiding strict 18:6 or longer fasts in the luteal phase, the roughly two weeks between ovulation and your next period. A 12:12 or 14:10 window through that phase is a cautious default rather than a tested rule, and our women's longevity guide goes deeper. Beyond that, let symptoms guide you: if your cycle changes, ease off to 14:10 or 12:12.