Tip · Fasting

24–72 h extended fasting

Q: When does autophagy start?

This question cannot be answered with human data. Autophagic flux under fasting has so far been measured in only one randomized trial, in blood cells and without comparing different fasting durations. The hour figures in circulation come from animal experiments and cell culture. What has been measured in humans: ketosis reaches a plateau after about five days, and the systemic protein changes in the blood only become visible after three days of complete fasting.

Q: Will I lose muscle during a 72-hour fast?

In the twelve-day fast on 250 kcal per day measured with imaging, the loss of volume in the calf muscle corresponded arithmetically and completely to the loss of glycogen and bound water, and maximal strength was preserved. In pure water fasting over two weeks, by contrast, 74 percent of the weight loss at the end of the fast came from fat-free mass, which was rebuilt after refeeding. At 72 hours the effect is correspondingly smaller, but there is no separate measurement for this duration.

Q: How safe is multi-day fasting?

In the largest cohort, with 1,422 people and fasting durations of 4 to 21 days, no fatal or life-threatening event occurred, adverse events affected 0.7 percent, and only two people had to stop. The context matters: this was a medically supervised clinical program with 200 to 250 calories per day, not pure water fasting on your own. There are no safety data on unsupervised multi-day fasting.

Q: What is refeeding syndrome and does it affect me?

It is a recognized clinical entity: as soon as energy is supplied again after prolonged starvation, insulin drives phosphate, potassium and magnesium into the cells and blood levels fall – in severe cases with organ failure. According to the NICE criteria, a healthy person of normal weight after 24 to 72 hours of fasting is not a high-risk case, because they meet only a single criterion. Being underweight, prior weight loss or fasting periods of more than five to ten days change that.

Q: Does fasting raise my cholesterol?

Yes, at least temporarily. A meta-analysis of 32 studies on water fasting found a significant rise in LDL in all duration subgroups, with a progressive increase up to about day 10 followed by attenuation, as well as a fall in HDL with fasting periods of more than three days. What this means in the long term has not been studied.

Q: Who should not do extended fasting?

People who are underweight, pregnant or breastfeeding, children and adolescents, people with a history of eating disorders and people with type 1 diabetes. Anyone taking blood glucose-lowering medication or SGLT-2 inhibitors needs medical supervision – for SGLT-2 inhibitors, cases of euglycemic ketoacidosis under extended fasting have been documented. And anyone who has already lost weight unintentionally beforehand or has a BMI under 18.5 already meets refeeding risk criteria.

The well-known time markers – autophagy from 24 h, stem cells from 72 h – come from mouse experiments, not from humans.

Twenty-four, forty-eight or seventy-two hours without food – the form of fasting with the biggest promises and the best-known time markers. The tip is included here because in humans it has now been measured when something actually happens, and because the riskiest part is not the fasting but eating again afterwards.

In short

The body measurably switches over, but the major changes only begin after three days: in the proteome study of 12 volunteers over seven days of water fasting, systemic changes only became visible after three days of complete caloric restriction. Safety is well documented in the largest cohort – 1,422 people, 4 to 21 days, adverse events in 0.7%, no fatal or life-threatening event. The weight loss is largely water and glycogen; maximal strength was preserved in the imaging measurement. The time markers “autophagy from 24 h” and “stem cells from 72 h” cannot be traced back to any human measurement. And all safety data come from medically supervised programs.

What is behind it

After about twelve to sixteen hours, liver glycogen stores are largely empty and metabolism switches to fatty acids and ketone bodies. In the cohort of 1,422 people fasting, blood glucose fell into the lower normal range, ketone bodies rose significantly (p < 0.001 in each case), and ketosis reached a plateau after about five days.

What comes after that has only begun to be measured. In the methodologically strongest paper, 12 volunteers fasted for seven days on water only; more than 1,000 plasma proteins changed significantly, in nine trajectory patterns, a large share of them proteins of the extracellular matrix. These systemic changes only became visible after three days. This is the best evidence that longer fasting does more than save calories – but it is proteomics, not a functional measurement.

What happens to muscle

The muscle question has been answered with imaging. 32 participants, half of them over 50, fasted for 12 days on 250 kcal per day. The muscle volume of the right calf decreased by 271 ml (5.4%, p < 0.001) – arithmetically matching the expected loss of glycogen and bound water, estimated together at 404 to 505 ml. Maximal strength was preserved, regardless of sex and age.

With pure water fasting, the picture is different. After a median of 14 days, fat-free mass accounted for 74% of the weight loss. After 6 days of food reintroduction and a 45-day maintenance phase, the picture reversed: the fat mass share was then 67%, and bone mineral content remained unchanged. The difference between 250 kcal and zero calories is the most important practical finding on this.

The time markers

Autophagy has barely been measured in fasting humans. So far, one randomized trial has captured autophagic flux under fasting, in blood cells: in 121 people with obesity, the change after 6 months of intermittent fasting differed from the control group, though in a post hoc analysis and without a significant increase within the fasting group (Bensalem 2025). For the fasting-mimicking diet, there is a pilot study with 30 healthy people over eight days, likewise in mononuclear blood cells, with significant group differences only at individual time points and funded by the manufacturer. None of these measurements compares different fasting durations. The 72-hour marker for stem cell renewal comes from a mouse study; the authors themselves describe its human data as preliminary. The stronger human argument against the 24-hour marker comes from proteomics anyway: before day 3, little happens systemically.

What is well supported

Three findings hold up. First, the time threshold: 12 volunteers, seven days of pure water fasting, mean weight loss 5.7 ± 0.8 kg, more than 1,000 significantly altered plasma proteins – but systemic changes only after three days. Second, safety: in an observational study with 1,422 participants and 4 to 21 days of fasting, adverse events occurred in 0.7%, none was fatal or life-threatening, and only two had to stop; weight, waist circumference and blood pressure fell significantly (p < 0.001 in each case), and of 404 participants with pre-existing complaints, 341 (84.4%) reported an improvement. Third, muscle: 271 ml less calf muscle volume after 12 days on 250 kcal, fully explained by glycogen and bound water, with strength preserved.

What the studies show

Seven days of water fasting, 3,000 plasma proteins

12 volunteers (5 women, 7 men), seven days of pure water fasting, measured before, daily during and afterwards. The authors separate effects of fasting duration from effects of weight loss; systemic changes only became visible after three days. Using proteogenomic methods, they estimated the health consequences for 212 proteins and found both presumably favorable and unfavorable associations – a model calculation, not an outcome measurement. Limitation: twelve people, no control group.

1,422 people fasting, 4 to 21 days

Buchinger protocol with 200 to 250 kcal per day from juices and broths, 3 liters of water daily, exercise program. Of 1,311 questionnaires, the most common mild complaint was sleep disturbance with 169 mentions (14.94%), followed by fatigue with 155 (13.70%). Documented adverse events: cardiac arrhythmia 3 (0.21%), hyponatremia 3 (0.21%), hypoglycemia 2 (0.14%). Without a control group, the paper cannot demonstrate efficacy – for the safety question it is the best data base.

Blood lipids across 32 studies

Systematic review with meta-analysis and meta-regression across 32 human studies on pure water fasting. LDL cholesterol rose significantly (g = 0.489) in all duration subgroups, with a progressive increase up to about day 10 followed by attenuation. HDL fell overall (g = −0.233), with no change up to three days but a clear decrease beyond that.

Where the data stop

There is no randomized controlled trial of 24-, 48- or 72-hour fasting with hard endpoints. The human evidence consists of an uncontrolled proteome study with 12 people, an observational study without a control group and single-arm intervention studies. Blood lipids are also an open question: whether the rise in LDL is clinically relevant or a temporary side effect of lipolysis is unknown. A structured review lists transient hypogonadism, relative hypothyroidism, hypercortisolemia, orthostatic intolerance and electrolyte disturbances as potential risks – and notes that the evidence is limited by small samples and surrogate endpoints. On bone health there is an open signal: a critical review highlights bone fragility as a potential risk, while bone mineral content remained unchanged after 14 days of water fasting.

How to do it

What the studies actually did: 7 days of pure water fasting, supervised as inpatients, in 12 people. 4 to 21 days of Buchinger fasting with 200 to 250 kcal per day plus 3 liters of water daily and a moderate exercise program, followed by food reintroduction over an average of 4 days, increasing from 800 to 1,600 kcal per day – the best-documented implementation of all. And 12 days on 250 kcal per day in the muscle MRI study. There is no study on the frequency “24 h once a month, 72 h once a quarter”. The most reliable statement on duration remains: under three days, little happens systemically beyond the calorie deficit. And food reintroduction is a matter of days, not of one meal.

Safety

The most serious documented risk is refeeding syndrome – the shift of phosphate, potassium and magnesium into the cells as soon as energy is supplied again. The ASPEN consensus recommendations define it as a fall in serum phosphate, potassium and/or magnesium of 10 to 20% (mild), 20 to 30% (moderate) or more than 30% (severe) within 5 days of resuming or substantially increasing energy intake; severe forms additionally include organ failure resulting from these shifts or from thiamine deficiency. According to NICE, anyone with one of these characteristics is considered high risk: BMI under 16 kg/m², unintentional weight loss of more than 15% in 3 to 6 months, little or no food intake for more than 10 days, or low potassium, phosphate or magnesium levels before starting. A healthy person of normal weight after 72 hours meets only one criterion and is not a high-risk case; anyone who is underweight, has already lost weight beforehand or fasts for longer than five to ten days is. The second risk pathway is thiamine deficiency: Wernicke’s encephalopathy has been documented after 40 days of pure water fasting in a 36-year-old man with no history of alcohol use. Extended fasting is not an option for people who are underweight, pregnant or breastfeeding, for children and adolescents, for people with a history of eating disorders and for people with type 1 diabetes; in type 2 diabetes with medication it belongs in a doctor’s hands, and under SGLT-2 inhibitors cases of euglycemic diabetic ketoacidosis have been documented. Two points tend to get lost: in the 1,422-person cohort, hyponatremia occurred, not sodium excess, even though participants drank 3 liters of water daily and were given broths. And all safety data come from medically supervised programs; there are none on multi-day fasting on your own.

BK-Score Long used, barely studied

Human evidence		3
Mechanism		4
Safety data		5
Hype gap		2
Track record of use		7

The time markers – autophagy from 24 hours, stem cell renewal from 72 – come from mouse experiments. In humans, systemic changes in the plasma proteome only became visible after three days of complete fasting, and autophagy has barely been measured during fasting, with no comparison of different fasting durations.

The score rates the state of knowledge, not the effect. “Safety data 9” means well studied – not harmless.
Subjective assessment by Biohacking Kompakt based on published scoring rules – not a scientific rating and not a medical recommendation. Rules and all ratings (German)

Frequently asked questions about 24–72 h extended fasting

When does autophagy start?