What this page establishes
- Alternate-day fasting was the only fasting strategy in the 99-trial BMJ network meta-analysis to significantly outperform continuous energy restriction, with a mean difference of -1.29 kg (95% CI -1.99 to -0.59).
- The reported ADF advantage amounted to under 3 kg over a full year in a trial with notable dropout, so the difference is small relative to the adherence cost.
- Controlled evidence for fasts beyond roughly 48 hours as a weight-management strategy is essentially absent; the fasting literature summarised in the major meta-analyses covers ADF, whole-day (5:2) and time-restricted protocols.
Quick answers
Is alternate-day fasting the best fasting protocol?
It has the strongest comparative signal in the 99-trial network meta-analysis, and the edge over continuous calorie restriction was 1.29 kg.
Read the full answer
Dropout is high, so most people trade that advantage back.Should I do a 3-day fast?
Not for weight loss. There is no controlled evidence it beats a sustained deficit, much of the scale drop is glycogen and water that returns, and it carries electrolyte, refeeding and medication risks.
Is extended fasting dangerous on diabetes medication or a GLP-1?
Potentially, yes. Hypoglycaemia and dehydration are the immediate concerns on insulin or a sulfonylurea, and appetite suppression already reduces intake on a GLP-1.
Read the full answer
This needs a prescriber, not an internet protocol.
Two different things share this page because people search for them together. Alternate-day fasting is a legitimate protocol with the best comparative result in the fasting literature, and that result is 1.29 kg. Extended fasting beyond about 48 hours is not a weight-loss technique with a weak evidence base, it is a practice with essentially no controlled weight-management evidence and a genuine risk profile. The two deserve very different treatment.
Alternate-day fasting: the protocol
Strict ADF alternates a full fast day with a normal eating day. Modified ADF, which is what most trials actually used, allows roughly 500 kcal on fast days. The modified version is the one the evidence describes, so read published ADF results as modified unless a trial says otherwise.
The mechanism is calorie restriction. Roughly half your days at a very low intake produces a large weekly deficit, and the eating days in most trials do not fully compensate.
The one result where fasting beat continuous restriction
In the network meta-analysis of 99 randomised trials and 6,582 adults, alternate-day fasting was the only fasting strategy to significantly outperform continuous energy restriction. The mean difference was 1.29 kg (95% CI 0.59 to 1.99).
That is a genuine finding and it should be read at its actual size. The advantage amounted to under 3 kg over a full year in a trial with notable dropout. A randomised trial in adults without obesity also found ADF elicited larger changes in fat mass than time-restricted eating, which is consistent with the deficit being larger rather than with a different mechanism.
So ADF is the highest-yield fasting protocol in the literature, by a margin that most people will trade back through the difficulty of sustaining it.
| Protocol | Typical weekly loss, % body weight | Dropout risk | Evidence |
|---|---|---|---|
| 16:8 time-restricted eating | 0.05 / 0.15 / 0.35 | Low | Strong |
| 5:2 | 0.20 / 0.40 / 0.70 | Medium | Strong |
| Alternate-day fasting | 0.25 / 0.50 / 0.85 | High | Strong |
| OMAD | 0.30 / 0.60 / 1.00 | High | Weak |
| Extended fast, 48h+ | Not applicable | High | No controlled weight-management evidence |
Dropout is the real cost
Adherence dropoff on ADF is high, and the pattern is worth understanding before you start. Every other day is a hard day, and there is no week in which the difficulty pauses. That is a different psychological load from 5:2, where five days out of seven are ordinary.
A protocol you abandon in week five has produced a smaller result than a gentler protocol you keep for a year. The 1.29 kg edge assumes you are still doing it, and the dropout figures suggest many people are not.
Extended fasts, 48 hours and beyond
The evidence here is close to absent, and that is the first thing to say. Controlled evidence for fasts beyond roughly 48 hours as a weight-management strategy does not exist in the meta-analytic literature. The major reviews cover ADF, whole-day protocols and time-restricted eating. Multi-day fasting is outside the scope of what has been tested.
The scale does move during a long fast, and most of that early movement is not fat. Glycogen stores bind water at roughly three grams of water per gram of glycogen, so depleting them drops several kilograms that return within days of eating normally. Reading that number as fat loss is the most common misinterpretation in this whole area.
This site's stack builder deliberately assigns extended fasting a contribution of zero and shows a safety note instead. That is not an oversight in the model.
Who should not attempt either
The list below applies to alternate-day fasting. For extended fasting, treat it as a list of absolute exclusions rather than cautions.
- Pregnancy and breastfeeding.
- A current or past eating disorder.
- Under 18.
- Type 1 diabetes, or treatment with insulin or a sulfonylurea.
- Any GLP-1 agonist or other appetite-suppressing medication, without prescriber agreement.
- A history of electrolyte disturbance, kidney disease, or a low starting body weight.
- Any medication that requires food or stable intake for safe dosing.
If you want the ADF result without the ADF load
5:2 is the same mechanism at lower frequency, with better adherence and a smaller measured effect. Most people who cannot sustain ADF do better on it than they would on repeated failed ADF attempts.
The other route is to accept a lower weekly rate and protect the composition of the loss instead. Protein at 1.2-1.6 g/kg/day with two to four resistance sessions a week barely registers on the scale and substantially changes what comes off it. Over a year that matters more than 1.29 kg.
The evidence, one row per claim
| Claim | Tier | Source |
|---|---|---|
| Alternate-day fasting was the only fasting strategy in the 99-trial BMJ network meta-analysis to significantly outperform continuous energy restriction, with a mean difference of -1.29 kg (95% CI -1.99 to -0.59). | established | BMJ Group |
| A randomised clinical trial in adults without obesity found alternate-day fasting elicited larger changes in fat mass than time-restricted eating. | emerging | Clinical Nutrition |
| The reported ADF advantage amounted to under 3 kg over a full year in a trial with notable dropout, so the difference is small relative to the adherence cost. | established | BMJ Group |
| Controlled evidence for fasts beyond roughly 48 hours as a weight-management strategy is essentially absent; the fasting literature summarised in the major meta-analyses covers ADF, whole-day (5:2) and time-restricted protocols. | established | PMC — comparing caloric restriction regimens (scope of included protocols) |
Questions people ask
Why does the stack builder give extended fasting a zero?
Sources
- BMJ Group
- Clinical Nutrition
- PMC — comparing caloric restriction regimens (scope of included protocols)
Keep reading
- Intermittent Fasting for Weight Loss: What 99 Trials SayHub for the fasting cluster. Fasting is a scheduling tool for calorie restriction, not a separate metabolic le
- The 5:2 Diet: Two Low-Calorie Days a Week, and What the Trials FoundThe real mechanism is spontaneous under-eating on the five unrestricted days, not the two hard days.
- How Fast Should You Actually Lose Weight?Numeric and decision-shaped. A floor, a ceiling, and the reason both exist.
- Protein and Lifting: How to Lose Fat Without Losing MuscleThe most actionable page here. A deficit decides how much you lose. Protein and lifting decide what.