What Happens to Your Body During a Supervised Multi-Day Fast, Day by Day

What is a therapeutic fast — and how is it different from 16:8?

Time-restricted eating such as 16:8 never takes you far past the point where the liver has used up its stored glucose. A therapeutic (or periodic) fast does: it runs for days, usually with water, tea and a token 200–250 kcal of vegetable broth or juice, so that the body spends most of the time in deep ketosis. The two best-documented supervised protocols are the German Buchinger method (4–21 days, ~250 kcal/day, 2–3 litres of water, daily clinical monitoring) and medically supervised water-only fasting (≥2 days, water alone). Both are done in residential clinics with physicians on site. That supervision is not a formality — it is what makes the safety numbers below apply.

The timeline below stitches together the classic physiology work on starvation (Cahill, Owen, and the Kerndt review) with the largest modern safety dataset: 1,422 people fasting 5, 10, 15 or 20 days at the Buchinger Wilhelmi clinic, monitored daily.

Day 1: glycogen runs out and the metabolic switch flips

For the first several hours after your last meal the body runs on the glucose it absorbed and on liver glycogen. Anton and colleagues define the "metabolic switch" as the point at which liver glycogen is depleted and fatty acids are mobilised, "typically beyond 12 hours after cessation of food intake"; by the end of the first day the glycogen store is essentially spent. Two things happen next. The liver starts making new glucose from amino acids (gluconeogenesis) to feed the brain, which is why early fasting is the phase of highest protein breakdown. And fat stores begin releasing fatty acids, which the liver converts into ketone bodies.

The first day is also when the kidneys dump sodium. Kerndt's review notes that the rapid weight loss of early fasting — averaging 0.9 kg a day in the first week — is "primarily due to negative sodium balance", i.e. water. This matters for two reasons: the scale flatters you early, and it is the reason the first days can bring light-headedness on standing.

Days 2–3: ketosis, hunger fades, and the rough patch

Ketone production rises steeply through days two and three. This is the least comfortable window in every dataset. In the Buchinger cohort the mild symptoms — headache, sleep disturbance, muscle pain, hunger — "occurred mainly in the first days of the fast". Over the whole fast the most frequent self-reported symptoms were sleep disturbance (14.9%), fatigue (13.7%), dry mouth (8.8%), back pain (7.4%), hunger (6.8%), bad breath and headache (5.4% each), muscle pain (4.3%) and bloating (4.2%). Bad breath is the acetone; dry mouth and headache are usually fluid and salt.

The remarkable finding is what happens to appetite: 93.2% of participants reported an absence of hunger once the fast was under way. Rising ketones are the usual explanation, though the mechanism is not fully settled. Hormonally, insulin and the active thyroid hormone T3 fall while glucagon rises, which is the endocrine signature of a body switching from storing to mobilising fuel.

Days 4–5: ketones plateau and the brain changes fuel

In the Buchinger study, urinary ketones rose to a plateau "reached after 5 days", with men reaching higher levels than women. This is the point where the fast becomes metabolically stable. Owen and Cahill's catheter studies of the 1960s showed why it matters: as ketone levels rise they replace glucose as the brain's predominant fuel, which cuts the demand for gluconeogenesis and spares muscle protein. Early fasting burns protein at a high rate; established fasting burns much less. This protein-sparing adaptation is the reason a five-day fast and a five-day crash diet are not the same thing physiologically — and also why breaking a fast early, around day two or three, delivers most of the discomfort and little of the adaptation.

By day five the Buchinger cohort had lost a mean of 3.2 kg and 4.6 cm of waist circumference. Blood glucose had settled into the low-normal range in every group.

Days 5–10: blood pressure, lipids, well-being

This is the phase where the cardiometabolic changes become large. Across the whole 1,422-person cohort, systolic blood pressure fell from 131.6 to 120.7 mmHg and diastolic from 83.7 to 77.9, stabilising around 120/78; the longer groups fell further because they started higher. Heart rate did not change. Triglycerides fell by 0.44 mmol/L on average and total cholesterol fell more the longer the fast lasted, while HDL also decreased. Physical and emotional well-being, self-rated daily on a 0–10 scale, rose significantly in every group — and interestingly the people who chose longer fasts started with lower well-being scores and ended at the same level as everyone else. Whether these changes persist after refeeding is a different question, and this study did not follow people home.

Days 10–20: what keeps changing, and what to watch

Weight loss slows — by week three Kerndt puts it at about 0.3 kg a day — because the early water loss is over and the body is now burning fat at the rate its energy needs dictate. The 20-day Buchinger group lost 8.6 kg and 8.8 cm of waist. Ketosis stays high. Erythrocytes and haemoglobin drifted up slightly (haemoconcentration), platelets drifted down, and liver enzymes rose within the normal range.

Two lab values need naming because they are the ones that turn into complications:

• Uric acid rose from 338 to 495 µmol/L — a near-50% increase — because ketones compete with urate for excretion by the kidney. Kerndt lists gout and urate kidney stones among the documented complications of prolonged fasting, and the Buchinger series recorded one case of gout. Anyone with a history of gout or kidney stones needs to discuss that before a fast of any length.
• Sodium and potassium fell (sodium from 140.1 to 138.7 mmol/L; potassium slightly), staying within the normal range in this monitored cohort — three cases of hyponatraemia and one of hypokalaemia were recorded as adverse effects. Unmonitored, low potassium is the pathway to the cardiac arrhythmias Kerndt describes.

How safe is a multi-day fast under supervision?

Two datasets answer this, and they agree. In the Buchinger cohort (250 kcal/day, daily nurse and physician review), adverse effects beyond mild symptoms were reported in less than 1% of the 1,422 participants; there were two hospitalisations, three cases of cardiac arrhythmia, two of hypoglycaemia, and no deaths or permanent harm. In Finnell and colleagues' review of 768 supervised water-only fasts of two days or more at a U.S. residential centre, 75% of the recorded adverse events were mild and known fasting reactions; 212 visits had a grade-3 event as their worst, one had a grade-4 event, two visits out of the 768 involved a serious adverse event, and there were no deaths.

Read those numbers with their setting attached. Both cohorts were screened before starting, examined daily, had bloods taken, and were refed under supervision. None of that is available on a self-directed five-day water fast at home, which is why "less than 1%" is a statement about supervised fasting, not about fasting.

Breaking the fast: why refeeding is the dangerous part

The body that has spent a week adapting to fat and ketones is not ready for a normal meal. When carbohydrate returns, insulin surges and drives phosphate, potassium and magnesium into cells; the hallmark of refeeding syndrome is hypophosphataemia, and in the series reviewed by Mehanna it appeared a mean of 1.9 days after feeding restarted. The UK's NICE criteria flag anyone with "little or no nutritional intake for more than 5 days" as at risk when they have one other factor (BMI under 18.5, recent weight loss over 10%, or use of alcohol, insulin, diuretics or antacids), and anyone with no intake for more than 10 days as high-risk on that basis alone. This is why the supervised protocols refeed gradually — Finnell's centre required a refeeding period of half the fast's length — starting with small amounts of easily digested food and building up over days.

Who should not attempt a therapeutic fast

The complications above translate directly into who should not fast without medical sign-off: anyone with type 1 diabetes or anyone on insulin or sulfonylureas (hypoglycaemia), pregnancy and breastfeeding, a history of eating disorders, low body weight, kidney disease, gout or urate stones, cardiac arrhythmias, and anyone on medication whose dose depends on food or hydration — which includes many blood-pressure drugs, because the pressure fall above is real. If you take any regular medication, a fast longer than a day is a prescriber conversation before it is anything else.

What to track during a fast — and why

The clinics track a short list every day, and the reason is that fasting complications announce themselves early in the numbers before they are obvious in how you feel. At home, on a supervised programme or in the shorter fasts most people actually do, the same list applies:

1. Water intake — the Buchinger protocol specifies 2–3 litres a day. Headache, dry mouth and dizziness are mostly a hydration and salt story.
2. Weight, read as a trend — expect the fast early drop; it is water. What matters is the slope after day three.
3. Sleep — the most common complaint at 14.9%. If your fasting window trades sleep for fat loss, that is worth knowing; we cover the mechanism in does fasting affect sleep.
4. Mood and energy — well-being rose in the clinic data, but the first three days often dip; a daily 0–10 shows whether you are on the curve or off it.
5. Symptoms with a time stamp — palpitations, fainting, persistent vomiting or confusion are stop-and-call signs, not things to push through.

tr8ck's fasting module logs the fasting timer alongside water, weight, sleep and mood in one timeline, so the pattern of a fast — the day-two headache, the day-four energy return, the sleep that got shorter — is visible rather than remembered. It is a record to bring to the person supervising you, not a substitute for them.

FAQ

Medical disclaimer: This article is for informational purposes only and does not constitute medical advice. Fasting is not appropriate for everyone, and fasts longer than 24 hours should only be undertaken with medical supervision. If you are pregnant or breastfeeding, have a history of disordered eating, have type 1 diabetes, kidney disease, gout or a heart rhythm disorder, or take insulin, sulfonylureas or any medication that must be taken with food, speak to a qualified healthcare professional before changing your eating pattern.