Intermittent fasting is a family of protocols, not one thing. They differ in how much deficit they create, how hard they are to sustain, and how much medical supervision they require. For most adults the answer is 16:8, and the reason is adherence rather than any metabolic advantage.
The choice that matters more than the protocol is whether the eating window actually produces a calorie deficit and still delivers enough protein and micronutrients. Most fasting failures are one of those two problems, not a failure of the fasting window itself.
The Verdict
The protocols compared
| Protocol | How it works | Primary benefit | Adherence | Best for |
|---|---|---|---|---|
| 16:8 (time-restricted eating) | 16-hour fast, 8-hour window, daily | Weight loss, insulin sensitivity, blood pressure | High — most people adapt in 7–10 days | The default. Best evidence-to-effort ratio. |
| 14:10 | 14-hour fast, 10-hour window, daily | Modest metabolic benefit, circadian alignment | Very high | Starting point for anyone who finds 16:8 hard, and adequate for many |
| 18:6 / 20:4 | Longer daily fast | Larger calorie deficit than 16:8 | Moderate — protein targets get difficult | Short-term fat-loss phases, not indefinite practice |
| OMAD (one meal a day) | Single meal, roughly 1-hour window | Aggressive deficit, maximum simplicity | Low over months | Rarely compatible with adequate protein or micronutrients |
| 5:2 | 5 normal days, 2 days at 500–600 kcal | Weight loss without daily restriction | Moderate–high | People who would rather have two hard days than 7 restricted ones |
| Alternate-day fasting | Alternating fast and feed days | Largest weight and blood pressure effects in trials | Low — highest dropout of any protocol | Short supervised interventions |
| Extended fast (3–5+ days) | Water, electrolytes, no calories | Deep ketosis, autophagy signaling, large acute metabolic shift | Not applicable — episodic | Physician-supervised only, 1–2 times per year at most |
Note the adherence column, because it is where most of the real difference lives. Alternate-day fasting produces the largest weight and blood pressure changes in 12-week trials and also the highest dropout rate. A protocol you abandon at week 10 loses to one you keep for three years.
One correction to the popular framing: 16:8 without any change to what you eat does not reliably create a deficit. The 2020 TREAT trial randomized adults to 16:8 or three structured meals with no calorie instruction, and found about 1% weight loss in the fasting group — not significantly different from the control, and with a large share of the loss coming from lean mass. The window creates an opportunity to eat less. It does not enforce it.
16:8 — how to actually run it
Pick a window and hold it constant. Moving the window day to day removes the circadian benefit, which is a meaningful part of the effect.
- Window placement. Early beats late. An 8 AM–4 PM window outperforms noon–8 PM on postprandial glucose and blood pressure at identical fast length, because insulin sensitivity is highest in the morning and evening melatonin blunts insulin secretion.
- Adaptation. Hunger peaks around days 3–5 and largely resolves by day 10 as ghrelin secretion re-entrains to the new meal schedule. Headaches in the first week are usually sodium loss, not caffeine.
- Sodium. Lower insulin means more sodium excretion. 1–2 g of additional sodium per day during the first two weeks prevents most of the lightheadedness and fatigue people attribute to the fast itself.
- Training. Resistance training in a fasted state is fine for strength work. If a session lasts over 90 minutes or you are training for performance rather than health, place it inside the window.
What breaks a fast
The answer depends on which outcome you care about. For weight loss, calories break the fast. For insulin and glucose, carbohydrate and protein break it. For autophagy signaling, even small amounts of leucine break it. This table sorts the common cases.
| Item | Breaks the fast? | Why |
|---|---|---|
| Black coffee, plain tea, water, sparkling water | No | Zero calories, no insulin response |
| Electrolytes with no sugar (sodium, potassium, magnesium) | No | Recommended on fasts over 24 hours |
| A splash of cream or whole milk (under ~20 kcal) | Negligible | Will not meaningfully raise insulin; may blunt an autophagy-focused fast |
| MCT oil or butter in coffee (100–200 kcal) | For weight-loss purposes, yes | Calories are calories; insulin response is small but the deficit shrinks |
| BCAA or EAA powders | Yes | Leucine directly activates mTOR — this ends a fast in the sense that matters |
| Diet soda / sucralose / aspartame | Usually no measurable insulin response | Cephalic-phase effects vary between individuals and can drive appetite |
| Sugar-free gum, mints, cough drops | Trivially, but repeatedly | Polyols add up and keep chewing-driven appetite active |
| Most medications and prenatal vitamins | No, but many need food | Levothyroxine requires an empty stomach; NSAIDs, metformin, and iron require food |
Extended fasts — the supervised category
Water-only fasts of three days and longer produce deep ketosis, a large drop in insulin and IGF-1, and the autophagy signaling that drives much of the interest. They also carry the risks that make them a medical procedure rather than a wellness practice.
- Refeeding syndrome. Past roughly five days, reintroducing carbohydrate drives phosphate, potassium, and magnesium intracellularly. Serum phosphate can fall far enough to cause arrhythmia, respiratory failure, or seizure. Refeeding starts with small low-carbohydrate volumes over 24–48 hours with electrolytes monitored.
- Hypotension and syncope. Combined volume depletion and vasodilation cause fainting, particularly on standing. This is the most common reason extended fasts end early.
- Gout flare. Ketones compete with uric acid for renal excretion, so serum uric acid rises during fasting. Anyone with a gout history should expect a flare risk.
- Gallstones. Rapid weight loss and prolonged gallbladder stasis raise gallstone formation risk substantially.
- Medication conflicts. SGLT2 inhibitors can produce euglycemic ketoacidosis during a fast. Insulin and sulfonylureas cause hypoglycemia. Lithium levels rise with fluid and sodium shifts. Each of these needs a prescriber-directed plan.
Who should not fast
Fasting is contraindicated in pregnancy and breastfeeding, in anyone with a current or past eating disorder, in type 1 diabetes without specialist supervision, and in adults who are underweight or frail. It requires medical direction for anyone on insulin, sulfonylureas, SGLT2 inhibitors, lithium, or multiple antihypertensives.
A subgroup where the standard advice inverts: lean, highly active women. Trials in this group report menstrual irregularity and reduced luteinizing hormone pulsatility under aggressive restriction, while the metabolic upside is small because there is little metabolic dysfunction to correct. For this group, 12:12 or 14:10 is the appropriate ceiling, and cycle changes are a stop signal rather than a side effect to tolerate.
A second inversion: adults over 75. The dominant risk shifts from metabolic disease to sarcopenia and falls. Any protocol that reduces total protein intake in this group works against the outcome that matters most.
What to measure
Set a baseline before you start, then re-test at 12 weeks. Fasting is easy to feel good about and hard to evaluate without numbers.
- Fasting insulin and HOMA-IR — the most responsive markers, and the mechanism behind most of fasting’s benefits.
- HbA1c — reflects about 90 days of glucose exposure, so do not re-test before three months.
- ApoB — the cardiovascular risk marker that matters. Weight loss usually lowers it; a high-saturated-fat eating window sometimes raises it despite weight loss.
- hsCRP — inflammation, which typically falls with fat loss.
- Blood pressure — seven-day home average, morning and evening. See the blood pressure guide for the measurement protocol.
- Body composition, not just weight. A DEXA or a consistent bioimpedance scale distinguishes fat loss from lean mass loss. Weight alone cannot.
Frequently Asked Questions
What is the best fasting protocol for beginners?
Start at 14:10 for two weeks, then move to 16:8 if it feels easy. Most of the metabolic benefit shows up by 16 hours, and adherence over months matters far more than the length of any single fast. Set the window early — roughly 8 AM to 4 PM outperforms noon to 8 PM on glucose and blood pressure at identical fasting duration.
Can I drink coffee while fasting?
Black coffee, plain tea, water, and unsweetened sparkling water do not break a fast. A splash of cream under about 20 calories is functionally irrelevant for weight loss, though it likely blunts an autophagy-focused fast. Butter-and-MCT coffee at 150–200 calories does break the fast in the way that matters for a calorie deficit. BCAA and EAA powders definitely break it — leucine activates mTOR directly.
How fast will I see results?
Weight starts moving in 2–4 weeks, and the first 1–2 kg is largely glycogen and water. Fasting insulin and blood pressure shift over 8–12 weeks. ApoB and HbA1c need 3–6 months, because HbA1c reflects roughly 90 days of red-cell glucose exposure and does not respond faster no matter what you do.
Should I fast if I am trying to build muscle?
You can, but the arithmetic gets tight. A 180 lb person with 20% body fat needs roughly 100–145 g of protein daily. In an 8-hour window that is two 50 g servings or three 35 g servings, which is workable. In a 4-hour window or on OMAD it usually is not — 100 g of protein in one sitting exceeds what most people will eat, and single-bolus absorption favors distributing it. If you cannot hit protein, widen the window rather than accept the shortfall.
Is fasting different for women?
The evidence is genuinely contested. Some trials show no sex difference in metabolic outcomes; others report menstrual irregularity and reduced luteinizing hormone pulsatility in lean, active women under aggressive restriction. The practical position: women with low body fat, high training loads, or a history of amenorrhea should use 12:12 or 14:10 rather than 16:8 or longer, and should treat a missed or irregular cycle as a signal to stop.
What is refeeding syndrome and when does it matter?
After a fast of roughly five days or more, reintroducing carbohydrate triggers an insulin surge that drives phosphate, potassium, and magnesium into cells. Serum levels can crash, producing arrhythmia, respiratory failure, or seizure. It is the specific reason extended fasts require supervision. Breaking a long fast starts with small volumes of low-carbohydrate food — broth, eggs, a little fat — over 24–48 hours, with electrolytes checked.
Will fasting slow my metabolism?
Adaptive thermogenesis happens with any sustained calorie deficit, fasting or not, and typically runs 5–15% below the rate predicted by the new body weight. Fasting does not make it worse. What does make it worse is losing lean mass, which is why resistance training twice a week and adequate protein matter more inside a fasting protocol than outside one.