A 72-hour fast is not a longer version of skipping breakfast. By the third day without calories, the body has moved from using stored carbohydrate as its main buffer to running heavily on fat-derived ketones, with lower insulin, higher glucagon, and meaningful changes in fluid and electrolyte handling. The biology is interesting. The clinical case for doing it, especially without supervision, is much thinner.
What changes in the first three days
The early fast is mostly a carbohydrate story. Blood glucose does not simply fall until it runs out; the liver releases glucose from glycogen and, as that reserve falls, makes new glucose through gluconeogenesis. NCBI’s StatPearls chapter on fasting notes that liver glycogen has the largest role in maintaining blood glucose during roughly the first 24 hours, after which fat and protein stores contribute more heavily to fuel production.
That transition is why a 72-hour fast feels different from a 16-hour eating window. Free fatty acids rise. The liver converts some of those fatty acids into ketone bodies, particularly beta-hydroxybutyrate and acetoacetate. These are not mystical longevity signals; they are alternative fuels, and they also interact with cellular signalling pathways.
By 48 to 72 hours, many people will be in clear nutritional ketosis. Hydration status, baseline diet, activity, body size, and medication use all alter the curve. A person who starts after several days of low-carbohydrate eating will not follow the same trajectory as someone who begins after a high-carbohydrate diet and a hard training week. That variability is one reason fixed online timelines are less useful than they look.
Mechanism is not the same as outcome
The fasting internet tends to move too quickly from mechanism to promise. Autophagy is the cleanest example. It is a real, well-described cellular housekeeping process, and fasting plausibly changes autophagy-related signalling in humans. What we do not have is a direct clinical translation from a three-day fast to slower ageing, fewer chronic diseases, or longer life.
A 2018 peer-reviewed review on metabolic switching during fasting describes the shift from glucose to fatty-acid-derived ketones and the activation of signalling pathways involved in autophagy, mitochondrial biogenesis, and cellular stress resistance. That paper is useful for mechanism, but it is not proof that a self-directed 72-hour water fast improves longevity outcomes in otherwise healthy adults.
The distinction matters. Lower insulin during a fast is physiology. A temporary fall in body weight is expected. A durable improvement in cardiometabolic risk after normal eating resumes is a clinical outcome, and that requires harder evidence.
What human studies actually show
For prolonged fasting specifically, the human evidence base is small and uneven. A 2024 narrative review in Nutrition Reviews summarised human trials of water-only and modified prolonged fasts. Across studies lasting five to 20 days, body weight fell by about 2 to 10 per cent, circulating ketones rose substantially, and blood pressure often decreased.
The same review also makes the caution clear. Roughly two-thirds of the lost weight in those trials was lean mass rather than fat mass. Plasma lipid findings were variable. Glycaemic markers improved in some people without diabetes but not consistently in people with diabetes. Several benefits were no longer present three to four months after the fast in studies that tracked follow-up.
That does not make prolonged fasting useless. It means the evidence supports short-term physiological change more strongly than lasting disease modification. It also means the setting matters. Many studies involved screening, monitoring, structured refeeding, or residential clinics. That is not the same exposure as a person deciding on a Sunday night to drink only water until Thursday.
The risks are not theoretical
For a healthy adult, the common short-term problems are usually mundane: headache, dizziness, hunger, sleep disruption, dry mouth, reflux, constipation, irritability, and reduced exercise tolerance. Those are not trivial if someone needs to drive, work with machinery, train hard, or care for others.
The higher-risk problems sit around glucose, blood pressure, dehydration, electrolytes, uric acid, and medication timing. Diabetes is the clearest example. Diabetes UK’s fasting guidance warns that people using insulin or certain glucose-lowering tablets can develop hypoglycaemia during a fast, and that fasting without medical advice can also raise the risk of diabetic ketoacidosis in some circumstances.
Endotext’s chapter on diabetes management during Ramadan reports that severe hypoglycaemia requiring hospitalisation increased during Ramadan in studies of people with type 1 and type 2 diabetes, and that hypoglycaemia was significantly associated with insulin and sulphonylurea use. Ramadan fasting is not identical to a 72-hour fast, but it is a clinically useful warning: food restriction plus glucose-lowering medication needs individual risk assessment, not generic advice.
NIDDK guidance for clinicians makes the medication issue practical. Sulphonylureas, short-acting meglitinides, and insulin are associated with hypoglycaemia, whilst SGLT2 inhibitors and diuretics raise dehydration considerations. That is before accounting for blood-pressure drugs, lithium, anticoagulants, seizure medicines, or any drug that must be taken with food.
Who should not attempt this casually
The safest answer is conservative: a 72-hour fast is not appropriate as a casual experiment for anyone with diabetes using insulin, sulphonylureas, meglitinides, or complex glucose-lowering medication; anyone pregnant or breastfeeding; anyone with a current or past eating disorder; adolescents; people who are underweight, frail, or recovering from illness; and anyone with kidney disease, gout, significant heart disease, recent heart attack or stroke, active infection, fever, or complex medication schedules.
People with a history of fainting, low blood pressure, arrhythmia, migraine triggered by missed meals, heavy training loads, or physically risky work also need more caution than social media usually gives them. A fast that is merely uncomfortable for one person may be unsafe for another because the underlying risk is not visible from the outside.
There is also a behavioural risk. Prolonged fasting can look like discipline whilst functioning as restriction. For anyone with binge-restrict cycles, compulsive weighing, fear of eating, or a history of disordered eating, the metabolic discussion is secondary. The intervention can worsen the pattern even if the physiology is textbook.
Refeeding deserves more respect
The end of a fast is not a footnote. After several days without food, large meals, alcohol, very intense exercise, or abrupt high-carbohydrate refeeding can produce symptoms even in people without recognised disease. In longer medically supervised fasts, refeeding is treated as part of the protocol because fluid shifts, gastrointestinal tolerance, and electrolyte handling matter.
A three-day fast is shorter than the classic high-risk scenarios for refeeding syndrome, but that should not turn refeeding into a challenge meal. The cautious approach is to treat the first meals as a transition, not a reward. That means smaller portions, adequate fluids, some salt if appropriate for the individual, and attention to symptoms such as confusion, severe weakness, chest pain, persistent vomiting, fainting, or palpitations. Those are medical warning signs, not normal detox sensations.
What this means in practice
- Treat a 72-hour fast as a medical-risk intervention, not a wellness reset.
- Do not attempt one without clinical guidance if you use glucose-lowering drugs, diuretics, blood-pressure medication, lithium, anticoagulants, seizure medicines, or any medicine that must be taken with food.
- Avoid prolonged fasting entirely during pregnancy, breastfeeding, adolescence, active illness, underweight, frailty, eating-disorder history, kidney disease, gout, or significant cardiovascular disease unless a clinician has specifically advised otherwise.
- Separate the goals: weight loss, glucose control, blood pressure reduction, and curiosity about ketosis are different questions with different risk-benefit profiles.
- Plan refeeding before starting, including what would make you stop and seek medical help.
- Be sceptical of claims that 72 hours is a proven longevity threshold. The mechanism is plausible; the outcome evidence is not there.
What we don’t know
We do not know whether occasional 72-hour fasts improve long-term health outcomes in generally healthy adults compared with safer, more sustainable approaches such as protein-adequate calorie reduction, resistance training, better sleep, or time-restricted eating.
We also do not know who benefits most, how often the intervention would need to be repeated, or how much of any benefit comes from the fast itself rather than the weight loss, reduced ultra-processed food intake, alcohol avoidance, or behavioural reset around it. Most studies are small, short, and selected. Many exclude precisely the people most likely to be harmed.
The honest position is narrow. A 72-hour fast produces a clear metabolic switch. It may temporarily improve some markers in some supervised settings. It is not a proven longevity treatment, and it is a poor candidate for casual experimentation in anyone with medical complexity.
The body can adapt impressively to three days without food. That adaptation is not the same thing as a recommendation.
Photo: Sasun Bughdaryan on Unsplash.