Nutrition, Diet & Fasting - Fasting
Prolonged water fasting (48-72 hours and beyond)
Two to three days or more without food. Real research exists, some of it recent and less flattering than the online conversation suggests - and the most serious risk is not the fast itself but how it ends.
A prolonged fast means water only, for 48 hours, 72 hours or longer. It is a genuinely different proposition from skipping breakfast, and it is the point in this classroom where the safety section stops being a formality. Extended fasting has been studied in medically supervised settings, and there is a reason the studies are medically supervised.
What happens across three days
The first day empties glycogen. By the second, fat is the dominant fuel and ketone levels climb well above anything reached in a daily fasting window - this is where "deep ketosis" comes from. Insulin is very low. The body also increases protein breakdown to supply glucose for tissues that need it, which is why muscle preservation becomes a real consideration rather than a theoretical one as fasts extend.
Electrolytes shift. Sodium, potassium and magnesium are lost, and this is behind much of what people experience: dizziness on standing, headache, cramps, palpitations. In supervised fasting studies these are among the commonly reported adverse events, alongside fatigue, insomnia and dry mouth - and in the less common but more serious column, hypokalaemia and arrhythmias.
What the research shows - including the awkward part
Supervised water-only fasting studies, including work following people through fasts averaging around ten days with guided refeeding, have reported improvements in cardiometabolic markers such as blood pressure and some lipid measures, with a portion of the effect persisting weeks afterwards. That is real published work and it is the strongest thing in favour.
The awkward part is more recent. A 2025 medically supervised water-only fasting and refeeding study reported that prolonged fasting promoted systemic inflammation and platelet activation in humans - the opposite direction from the anti-inflammatory effect prolonged fasting is usually sold on. One study does not overturn a field, and the clinical significance of those markers in healthy people is not established. But it is exactly the kind of finding that gets left out of enthusiastic summaries, and anyone deciding whether to do this deserves to know it exists.
A randomised crossover trial has also examined three-day water-only fasting, with and without glycogen-depleting exercise, specifically to measure autophagy markers in humans - which is a more honest approach than the usual practice of citing mouse data. What it found is discussed in the autophagy lesson, and the short version is that the human picture is messier than the claim.
Refeeding is the dangerous part
This is the section to read if you read nothing else. Refeeding syndrome is a real, potentially fatal clinical condition that occurs when someone who has been starved eats normally again too quickly. Reintroducing carbohydrate triggers an insulin surge that drives phosphate, potassium and magnesium rapidly into cells, and the resulting falls in blood levels can cause cardiac arrhythmias, respiratory failure, seizures and death.
It is well described in clinical medicine, and NICE has published criteria for identifying who is at high risk. Risk rises with the length of the fast, with low starting body weight, with alcohol use, and with any pre-existing deficiency. This is the single strongest reason that fasts beyond about 48 hours are done under supervision in the research literature - not the fast, the meal after it. Guided protocols reintroduce food gradually, typically over a period related to the length of the fast, rather than ending with a large meal.
Where the evidence stops
Prolonged fasting studies are small, and their participants are screened, monitored and refed under supervision - none of which describes someone doing this at home from a video. There is no long-term outcome data. And the balance of benefit and harm is genuinely unsettled: two credible bodies of published work now point in different directions on inflammation, which is the state of the science rather than a failure to look it up.
Who should not do this
Anyone taking insulin or other glucose-lowering medication, anyone underweight, anyone pregnant or breastfeeding, anyone with a history of eating disorders, anyone with type 1 diabetes, significant heart, kidney or liver disease, or anyone taking medication that requires food or careful blood level control. Those are not cautions to weigh - in the research literature they are exclusion criteria. The fasting safety lesson sets out the full picture and how supervised protocols handle it.
Educational content only and not medical advice, and specifically not a protocol to follow. Prolonged fasting is studied under medical supervision because it carries real risks, including refeeding syndrome, which can be fatal. Anyone considering a fast beyond 24 hours should discuss it with a qualified healthcare professional first.
Related in this classroom
Sources for this classroom
Educational content only. Not medical advice, diagnosis or treatment. Always consult a qualified healthcare professional before changing your health regimen.
