Intermittent fasting is the most popular eating pattern with the least official backing. That is not a criticism of it — it is the actual position of the National Institute on Aging, which states plainly: “There’s insufficient evidence to recommend any type of calorie-restriction or fasting diet.”
What follows is what the trials actually found, including the large head-to-head comparisons that almost never make it into articles on this subject.
The short version
- It works, and so does eating less. Across 99 randomised trials, only alternate-day fasting beat ordinary calorie restriction — by about 1.3 kg.
- Time-restricted eating did not beat calorie restriction in the largest comparisons, including a 12-month trial.
- Expect 3–8% of body weight over 8–12 weeks, which is the same range as any other calorie-reduction approach.
- Adherence is the real variable. The longest alternate-day fasting trial lost 38% of that group over a year.
- Side effects are mild and no more common than in control groups — but fasting is genuinely unsafe for some people.
- No US agency recommends it. None warns against it for healthy adults either.
The four protocols, defined properly
| Protocol | What it actually means |
|---|---|
| Time-restricted eating | NIA: “Meals are consumed within a limited number of hours (such as 6-8 hours) each day, with nothing consumed during the other hours” |
| Alternate-day fasting | NIA: “Eating is unrestricted every other day, and no or minimal calories can be consumed on the days in between” |
| 5:2 | NIA: “Eating is unrestricted for 5 straight days each week, followed by 2 days of restricted caloric intake” |
| Periodic / fasting-mimicking | A low-calorie, low-protein, plant-based five-day cycle, repeated every few weeks |
The American Diabetes Association adds the numbers most descriptions leave out: alternate-day and 5:2 protocols typically mean “energy restriction of 500–600 calories” on the restricted days, and time-restricted eating means a window of 8 to 15 hours.
What the head-to-head trials found
The biggest analysis to date is a 2025 network meta-analysis in The BMJ covering 99 randomised trials and 6,582 adults. Against unrestricted eating, every fasting approach produced weight loss:
| Compared with | Approach | Weight difference | Certainty |
|---|---|---|---|
| Eating freely | Alternate-day fasting | −3.40 kg | High |
| Eating freely | Whole-day fasting | −2.36 kg | High |
| Eating freely | Time-restricted eating | −1.72 kg | Moderate |
| Ordinary calorie restriction | Alternate-day fasting | −1.29 kg | Moderate |
Read the last row carefully, because it is the whole story. Compared with simply eating less every day, only alternate-day fasting came out ahead — and by about 1.3 kg. The authors’ own conclusion is measured: “Intermittent fasting diets have similar benefits to continuous energy restriction for weight loss and cardiometabolic risk factors.”
Two further results point the same way. A 2022 overview of twelve systematic reviews found a pooled difference between intermittent and continuous energy restriction of −0.33 kg, with a confidence interval from −1.17 to 0.51 — statistically indistinguishable from zero. And a 12-month randomised trial published in the New England Journal of Medicine compared an eight-hour eating window plus calorie restriction against calorie restriction alone in 139 people with obesity: −8.0 kg versus −6.3 kg, a net difference of −1.8 kg that did not reach significance. Its conclusion was that time-restricted eating “was not more beneficial with regard to reduction in body weight, body fat, or metabolic risk factors than daily calorie restriction.”
How much weight, realistically
The ADA gives the honest range in one sentence: these approaches produce “mild to moderate weight loss (3–8% loss from baseline) over short durations (8–12 weeks) with no significant differences in weight loss when compared with continuous calorie restriction.”
So the question is not which method is superior. It is which one you will still be doing in six months.
Adherence is the variable that actually decides it
This is where the trials get interesting, because the numbers differ sharply between protocols.
- Alternate-day fasting, 12 months: 38% of that group dropped out, against 29% for daily calorie restriction and 26% for controls. The trial’s verdict was that alternate-day fasting “did not produce superior adherence, weight loss, weight maintenance, or cardioprotection versus daily calorie restriction.”
- Time-restricted eating, 12 weeks: 83.5% adherence, self-reported, against 92.1% for the control condition.
- Eight-hour window plus calorie restriction, 12 months: 84.9% completed the study.
The pattern is consistent: the more aggressive the protocol, the better it performs on paper and the worse people stick to it. Alternate-day fasting wins the meta-analysis and loses the year.
Side effects: mild, and no more common than in control groups
A 2024 systematic review pooled 15 randomised trials covering 1,365 adults with overweight or obesity and compared adverse events directly against control groups:
| Effect | Fasting groups | Control groups |
|---|---|---|
| Fatigue | 14.5% | 16.2% |
| Headache | 13.5% | 15.9% |
| Dizziness | 9.8% | 9.4% |
| Serious adverse events | 6 people in total (0.4%), none judged related to the intervention | |
The reviewers concluded that “IF was not associated with a greater risk of AEs in adult patients affected by overweight or obesity.” Hunger and headaches show up in the first weeks of most trials and tend to fade.
Who should not do this
The NIH’s own list. Fasting “can be dangerous for some people.” Talk to a healthcare provider first if you: “Are under the age of 25. Are pregnant or breastfeeding. Take insulin or other medications to control diabetes. Have been prescribed any medication that must be taken with food. Have a seizure disorder. Work the night shift. Operate heavy machinery at your job.”
Two additions worth knowing. The same NIH source notes that people who regularly fast more than 16 or 18 hours a day have a higher risk of gallstones. And anyone with a history of disordered eating should avoid fasting protocols — while trials have not shown that fasting causes eating disorders, researchers in the field are explicit that a prior history is a reason not to start.
For people with diabetes, the ADA’s position is nuanced rather than prohibitive. It states that shortening the eating window “has been shown to be safe for adults with type 1 or type 2 diabetes” — but also that anyone taking insulin or insulin secretagogues “should be medically monitored during the fasting period,” because hypoglycaemia risk is higher with those medications and roughly half of hypoglycaemic events happen in the last few hours before breaking a fast.
The metabolic switch, without the mythology
The physiological argument for fasting is real. The NIA describes it accurately: fasting “triggers the body to switch its source of energy from glucose stored in the liver to ketones, which are stored in fat”, and ketone bodies “are not just fuel … They are potent signaling molecules with major effects on cell and organ functions.”
The timing matters for how you interpret popular claims. Circulating ketones begin rising after roughly 8 to 12 hours without food, and the metabolic switch itself typically occurs somewhere between 12 and 36 hours. A 16-hour window sits at the early edge of that range, not past it.
The NIA is equally clear about the limits: hundreds of animal studies and scores of human trials show improvements in various conditions, but “the evidence is less clear for lifespan effects” and human trials “have mainly involved relatively short-term interventions.” On longevity specifically: “There are no data in humans on the relationship between calorie restriction and longevity.”
FAQ
Is intermittent fasting better than counting calories?
Not for time-restricted eating — the 12-month trials and the pooled reviews found no significant advantage. Alternate-day fasting did beat calorie restriction in the 99-trial analysis, by about 1.3 kg, and had the worst long-term dropout rate of any protocol tested.
Which protocol should a beginner pick?
The one you will actually keep doing. Alternate-day fasting produces the biggest numbers in trials and the highest dropout; time-restricted eating produces smaller numbers and better adherence.
Does fasting slow your metabolism?
In the trials that measured it, no. The TREAT trial measured total and resting energy expenditure and found no significant change in either. A separate controlled-feeding study found no difference in energy expenditure from meal timing at all.
Does intermittent fasting extend lifespan?
Unknown in humans. The NIA states there are no human data on calorie restriction and longevity, and that lifespan evidence for fasting is less clear than the short-term health-marker evidence.
Is it safe long term?
Nobody knows yet. The NIA notes that more research is needed to determine whether the benefits hold, or whether the approach is even feasible, over years rather than weeks.
Where this fits
The most popular protocol has its own guide: 16/8 intermittent fasting — what an eight-hour window actually does, which covers the meal-timing evidence, what you can drink during the fast, and the lean-mass question in detail. For the eating pattern with the strongest long-term evidence behind it, see what the evidence actually supports.
Every figure here is linked to its source in the text: the National Institute on Aging, NIH News in Health, the American Diabetes Association Standards of Care, and peer-reviewed trials and reviews in The BMJ, the New England Journal of Medicine, JAMA Internal Medicine, Nutrition Journal and Nutrients. Checked against those sources on 1 September 2026. Educational information, not medical advice — see our medical disclaimer.
This content is for educational purposes only and is not medical advice. Always consult a qualified healthcare professional for personalised guidance.
